PRIMARY FREQUENCY AND TIME STANDARDS IN THE CONTEXT OF COMPUTER NETWORKS
While there are few NTP primary servers outside national laboratories that derive synchronization from
primary frequency and time standards, it is useful to assess the accuracy achievable using these means.
A primary frequency standard is an oscillator that can maintain extremely precise frequency relative to a
physical phenomenon, such as a transition in the orbital states of an electron or the rotational period of
an astronomical body. Existing atomic oscillators are based on the transitions of hydrogen, cesium,
rubidium, and mercury atoms, although other means using active and passive masers and lasers of
various kinds and even pulsars are available [14].
In addition, there is a wide variety of oscillator types, including oven-stabilized, temperature-
compensated, and uncompensated quartz crystal oscillators; rubidium gas cells; and cesium beam
oscillators. The best of these may be the hydrogen ion trap oscillator developed by NASA/Jet Propulsion
Laboratory (JPL) for use in the Deep Space Network, which has a stability of 1 × 10−15 [15]. Pulsars can
be better than this over long averaging times and may be the ultimate cosmic stabilizer since they are
self-powered and visible with only a telescope. However, only one of them has been studied so far [16].
For reasons of cost and robustness, cesium oscillators are used worldwide for national primary
frequency standards. The characteristics of cesium oscillators have been extensively studied and
accurate parametric models developed [17]. The current TAI timescale is maintained by a worldwide
ensemble of some 250 cesium oscillators in laboratories throughout the world. Among the best cesium
oscillators today is the NIST-F1 Cesium Fountain, which boasts a stability of 2 × 10−15 per day, although
future developments are expected to yield stabilities on the order of 1 × 10−18 per day.
Achieving this goal requires cryogenic devices and places extreme demands on oscillator and counter
technology. The frequency of crystal oscillators gradually changes over their lifetime, a phenomenon
called aging. Even if a crystal oscillator is temperature compensated by some means, it must be
periodically compared to a primary standard to maintain the highest accuracy. Various means have been
developed to discipline precision quartz crystal oscillators using GPS to calibrate parameters specific to
each individual crystal, but in general, aging is not a factor in computer clock oscillators. The
telecommunication industry has agreed on a classification of clock oscillators as a function of minimum
accuracy, minimum stability, and other factors [18].
There are three factors that determine the stability of a clock: drift, jitter, and wander. Drift refers to
long-term systematic variations of frequency with time and is synonymous with aging, trends, and the
like. Jitter (also called timing jitter) refers to short-term variations in frequency with components greater
than 10 Hz, while wander refers to intermediate-term variations 244 Computer Network Time
Synchronization in frequency with components less than 10 Hz. The classification determines the
oscillator stratum (not to be confused with the NTP stratum), with the more accurate oscillators
assigned the lower strata and less-accurate oscillators the higher strata, as shown in Table 13.5. The
construction, operation, and maintenance of stratum 1 oscillators is assumed to be consistent with
national standards and often includes cesium oscillators and sometimes precision crystal oscillators
synchronized via LORAN-C (Long Range Navigation System C) or GPS to national standards. Stratum 2
oscillators represent the stability required for interexchange toll switches such as the AT&T 4ESS and
interexchange digital cross-connect systems, while stratum 3 oscillators represent the stability required
for exchange switches such as the AT&T 5ESS and local cross-connect systems. Stratum 4 oscillators
represent the stability required for digital channel banks and PBX (private branch exchange) systems.
13.11 Time and Frequency Coordination A network of clocks that are synchronized (syntonic) at some
multiple or submultiple of a common frequency is called isochronous. A network of clocks that are not
syntonic but operate very close to a multiple or submultiple of a common frequency is called
plesiochronous.
In this context, the cesium clocks at the national laboratories are plesiochronous with each other and
with TAI. At NIST, USNO, and some other national laboratories, cesium clocks provide synchronization to
one or more NTP servers, which then synchronize the NTP subnet at large. In particular, to synchronize
frequency means to adjust the subnet clocks to run at the same frequency, to synchronize time means
to set the clocks so that all agree at a particular epoch, and to synchronize clocks means to synchronize
them in both frequency and time. To synchronize clocks, there must be some way to directly or
indirectly compare their readings. If two clocks can communicate directly over paths of precisely known
delay, then the time difference can be determined directly using algorithms similar to NTP. This is the
basis of the two-way satellite time and frequency transfer (TWSTFT) method described in Section Table
13.5 Clock Stratum Assignments Stratum Minimum Accuracy (per day) Minimum Stability (per day) 1 1 ×
10−11 Not specified 2 1.6 × 10−8 1 × 10−10 3 4.6 × 10−6 3.7 × 10−7 4 3.2 × 10−5 Not specified Metrology
and Chronometry of the NTP Timescale 245 18.1. If they cannot communicate directly but they can
communicate with a third clock over paths of precisely known delay, their differences can be
determined relative to the third clock and the difference of each clock communicated to the other.
Techniques based on this method use the GPS and LORAN-C navigation systems. Some timescales,
including TAI, are generated by an algorithm that combines the relative time differences measured
between contributing national and international laboratories.
The laboratories themselves usually use an algorithm, not necessarily that used for intralaboratory
coordination, to generate a laboratory timescale from an ensemble of cesium clocks. Not all laboratories
have a common view of these algorithms, however. In the United States, the national timescale is
officially coordinated by both NIST and USNO [14], although both laboratories cling to their own
timescales as well. Coordination methods incorporate both Kalman filter and parameter estimation
(autoregressive, integrated, moving-average, ARIMA) models [4]. The NIST algorithm that generates
NBS(AT1) is described in Weiss, Allan, and Peppler [19], while the USNO algorithm that generates
UTC(USNO) is described in Percival [20]. It is important to realize that it is not possible at the present
state of the art to establish a permanent time and frequency standard that operates continuously and is
completely reliable. A physically realizable standard is an active device that requires power and
environmental resources, must occasionally be repaired, and has only a flicker of life compared to the
age of the universe. While the TAI timescale in use today is based on a mathematical average of a large
ensemble of atomic clocks, which improves the stability and reliability of its institutional memory, it is
assumed that there are no subtle atomic conspiracies not yet discovered and that all the clocks in the
global ensemble do not burn out at the same instant. The recent discovery of millisecond pulsars may
provide a useful sanity check for the timescale as well as a means to detect gravitational waves [16].
13.12
Time and Frequency Dissemination To improve accuracy and minimize the effects of individual clock
variations, it is the practice in national standards laboratories to construct a synthetic timescale based
on an ensemble of at least three and possibly very many contributing primary clocks. The timescale is
produced by an algorithm using periodic measurements of the time offsets between the various clocks
of the ensemble. The algorithm combines the offsets using computed weights to produce an ensemble
timescale more accurate than the timescale of any clock in the ensemble. The algorithm used by USNO
is based on ARIMA models [20], while the algorithm used by NIST is evolved from Kalman filter 246
Computer Network Time Synchronization models [17,19,21]. These algorithms result in long-term
fractional frequency stabilities on the order of 1.5 × 10−14. So that atomic and civil time can be
coordinated throughout the world, national administrations operate primary time and frequency
standards and coordinate them cooperatively using GPS and common-view satellite methods described
in Section 18.1. Many seafaring nations of the world operate a broadcast time service for the purpose of
calibrating chronometers used in conjunction with ephemeris data to determine navigational position. In
many countries, the service is primitive and limited to seconds-pips broadcast by marine communication
stations at certain hours. For instance, a chronometer error of one second represents a longitudinal
position error of about 0.23 nautical mile at the equator.
13.12.1 Shortwave Radio Services
NIST operates shortwave broadcast services for the dissemination of standard time and frequency
information on frequencies of 2.5, 5, 10, 15, and 20 MHz from station WWVH in Fort Collins, Colorado,
and on frequencies of 2.5, 5, 10, and 15 MHz from station WWVH in Kauai, Hawaii. The timecode is
transmitted over a 60-s interval at a data rate of 1 b/s using pulse-width modulation on a 100-Hz
subcarrier. The National Research Council (NRC) of Canada operates a shortwave broadcast service for
the dissemination of time and frequency information on frequencies of 3.33, 7.85, and 14.67 MHz from
station CHU in Ottawa, Ontario. The timecode is transmitted during seconds 31–39 at 300 b/s using Bell
103-compatible FSK (frequency shift keying) modulation. For all three stations, the timecode format
includes UTC time of year in seconds, together with leap second warning, standard/daylight indicator,
and DUT1 adjustment.
Additional details of the signal design are in Section 7.6. Signal propagation from shortwave stations is
usually by reflection from the upper ionospheric layers. While these transmissions can be received over
large areas in North America, reliable frequency comparisons can be made only to the order of 10−7,
and time accuracies are limited to the order of a millisecond [4]. As far as is known, only one
manufacturer is still producing WWV/H receivers, and these would not ordinarily be considered
precision time sources. The current NTPv4 software distribution includes audio drivers for WWV,
WWVH, and CHU, as described in Section 7.6. The drivers demodulate and decode the audio signal from
a conventional shortwave receiver with accuracies generally to the millisecond or better. Additional
details of the signal demodulation and decoding algorithms are in Section 7.6. 13.12.2 Long-Wave Radio
Services NIST also operates a long-wave broadcast service for time and frequency dissemination on a
frequency of 60 kHz from station WWVB in Boulder, Metrology and Chronometry of the NTP Timescale
247 Colorado. The timecode is transmitted over a 60-s interval at a rate of 1 b/s using periodic
reductions in carrier power. The format is similar to that used by the shortwave stations.
The station can be received over the continental United States and adjacent coastal areas. Signal
propagation is via the lower ionospheric layers, which are relatively stable and have predictable diurnal
variations in height. With appropriate receiving and averaging techniques and corrections for diurnal
and seasonal propagation effects, frequency comparisons to within 10−11 are possible and accuracies of
from a few to 50 µs can be obtained [4]. Table 13.6 lists several other services similar to WWVB and
operated by national government agencies in Europe and Japan. A typical long-wave transmitter
antenna uses a network of wires connected between two or four towers 100 to 250 m tall and spaced
several hundred meters apart. The transmitter powers are in the range 20–50 kW, but the antenna
efficiency is low, on the order of 30 percent. These stations can be received at distances of 1,000– 3,000
km. In addition to these services, medium-wave stations are operated by BBC Radio 2 on 401 kHz from
Droitwich, United Kingdom, and France Inter on 162 kHz from Allouis, France. The transmitter powers
are 400–2,000 kW, and signals are usable throughout western Europe. The primary purpose of these
stations is for the radio broadcasting service, but the broadcast carrier can be used as a precision
frequency reference as well. 13.12.3 Geosynchronous Operational Environmental Satellite Service NIST
also provides a time and frequency dissemination service on about 468 MHz from the Geosynchronous
Operational Environmental Satellites (GOES), three of which cover the Western Hemisphere. The
timecode is interleaved with messages used to interrogate remote sensors. It consists of 60 four-bit BCD
nibbles transmitted over an interval of 30 s. The timecode information is similar to the terrestrial
services. There are only a few receivers for this service, which may not be supported in the future. A
distinguishing feature of all the services described so far is that they are one way; that is, they provide
no means to measure the propagation delay. The delay must be determined by geographic positions and
ray-path Table 13.6 Low-Frequency Standard Time Stations Call Sign and Location Frequency (kHz)
Power (kW) WWVB Fort Collins, Colorado, USA 60 50 DCF77 Mainflingen, Germany 77.5 30 MSF Rugby,
United Kingdom 60 50 HBG Prangins, Switzerland 75 20 JJY Fukushima, Japan 40 50 JJY Saga, Japan 60 50
248 Computer Network Time Synchronization geometry. This is complicated by movements of the
ionospheric layers during the night and day, changing propagation modes, and orbit station-keeping in
the satellite system. Generally, this limits the accuracy with these services to a millisecond. 13.12.4
Telephone Modem Services NIST also operates the Automated Computer Time Service (ACTS) over the
public switched telephone network from Boulder, Colorado. A call to the ACTS modem pool returns
about 1 min of timecode data, including time of year, leap warning, and DUT1 value. Calls to ACTS can
travel different routes for each call, which can result in significant errors, so ACTS measures the round-
trip propagation time if the caller modem echoes received characters.
The ACTS driver in the NTP software distribution does this and can realize a reliable error less than a
millisecond. The driver can also operate with the telephone format commonly used in Europe, the
model for which is the German PTB system, and the format used by the USNO; however, neither of
these services can measure and compensate for the propagation time, so the accuracy is degraded, in
the case of USNO to the order of 30 ms. 13.12.5 Global Positioning System The U.S. Department of
Defense operates the GPS for precision navigation [22] on land and sea and in the air. This system
provides 24-h worldwide coverage using a constellation of satellites in 12-h orbits inclined at 55°. The
original constellation of 24 satellites in six equally spaced planes of 4 satellites each has been expanded
to 31 satellites in an irregular configuration such that at least 6 satellites are always in view, and 8 or
more satellites are in view in most parts of the world. Services similar to GPS are operated or planned by
other countries. The Russian service is called GLONASS and has been operating for several years. The
European Union service is called Galileo and has completed the design and development phase and is
expected to become operational in 2013. While navigation is not the main topic of this section,
understanding the navigation function helps to clarify the time transfer function.
Each satellite vehicle (SV) transmits a navigation message that includes the time of transmission, an
almanac to determine which SVs are in view, and an ephemeris from which the precise position of an SV
in view can be determined. Each transmission from an SV appears as an expanding sphere that
eventually intercepts the receiver. The locus of intersection of two spheres is a circle, while the locus of
intersection of three spheres is two points. The point nearest Earth is the location of the receiver. As the
accuracy of this point is critically dependent on the receiver clock, the fourth SV is used to discipline the
receiver clock, which then can provide a pulse-per-second timing signal for the computer clock. If the
receiver is not moving, once the receiver position Metrology and Chronometry of the NTP Timescale 249
has been accurately determined, only a single SV is necessary to provide accurate time transfer. The
radio-frequency (RF) channel operates using a code division multiple access (CDMA) channel access
protocol. Each SV in the constellation is assigned a unique code number and PR sequence. Receivers
identify a particular SV by correlating the received signal with the assigned sequence. The sequence is
transmitted using BPSK (binary phase shift keying) modulation of the carrier, while the navigation
message is modulated on the sequence.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
Four L-band frequencies have been assigned for global navigation purposes, as shown in Table 13.7. All
but L4 are multiples of the highest chip rate 10.23 Mb/s. Only the L1 and L2 channels are considered
here; the others are used for a special purpose or for future systems. The coarse acquisition (C/A) signal
is used by military and civilian receivers to quickly acquire as many SVs in view as the receiver supports.
The PR has a chip rate of 1.023 Mb/s with period 1,023 chips, so it has a period of 1 ms. It implements
the Standard Positioning Service (SPS) with a nominal accuracy of 15 m. The precision (P) signal is
encrypted by a secret code (Y) for military receivers. It has a chip rate of 10.23 Mb/s with a period of
almost a week. It implements the Precision Positioning Service (PPS) with nominal accuracy of 1 m. The
L1C and L2C signals are for future civilian enhancements, while the M signal is for future military
enhancements. The navigation message is modulated on the PR sequence at a rate of 50 b/s in 1,500-bit
frames with five 300-bit subframes, so new complete navigation messages are repeated at 30-s
intervals. Many GPS navigation receivers can decode the GPS navigation message and produce a serial
ASCII (American Standard Code for Information Interchange) message in National Marine Electronics
Association (NMEA) formats that can be used by reference drivers in the NTP software distribution.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
The position of the receiver is the intersection of two hyperbolas, one for each station pair. Additional
stations can be used to refine the position. All LORAN-C stations use cesium clock ensembles that are
routinely calibrated against the USNO master clock. The U.S. Coast Guard publishes station time
differences for each station so that the pulse emission time can be calibrated to less than a
microsecond. While the LORAN-C system provides a highly accurate frequency and time reference
within the ground-wave area, there is no timecode modulation, so the receiver must be supplied with
UTC time from an external source. This can be done in an interesting way. A LORAN-C chain consists of a
master station and two or more slaves, all operating with the same pulse code group repetition rate
(GRI), which is some multiple of microseconds near 10,000. For instance, the eastern U.S. LORAN-C chain
operates with a GRI of 9,960 µs, so the least-common multiple of this interval and 1 s is 249 s. The U.S.
Coast Guard publishes a table showing past and future times of coincidence (TOC) when the GRI is
coincident with the UTC second. So, if the time can be set from other means within the 249-s interval,
the TOC can be used to establish the correct time. For instance, the first TOC after 17 h on 6 December
2009 was at 17:02:49 UTC and repeated every 249 s after that.
This detail might seem arcane, but it will pop up again in Section 18.6. LORAN-C receivers are used to
monitor local cesium clocks and other LORAN-C stations. Commercial LORAN-C receivers, such as the
Austron 2000 shown in Figure 13.5, are specialized and extremely expensive (up to $20,000). However, a
useful LORAN-C receiver for NTP use can be built with a junkbox personal computer (PC) and a handful
of inexpensive parts. Figure 13.6 shows an example of one built in our laboratory using an
ovencontrolled crystal oscillator (OCXO). It is not likely that LORAN-C service will be continued
indefinitely as GPS receivers are more accurate and less expensive. Where the highest availability is
required, multiple reference clocks can be operated in tandem and connected to an ensemble of
servers. Perhaps one of the more extreme configurations is operated at the University of Delaware in
Metrology and Chronometry of the NTP Timescale 251 Newark and shown in Figure 13.7. It consists of
dual-redundant primary GPS receivers, dual-redundant secondary WWVB receivers, a primary cesium
frequency standard, and a secondary quartz frequency standard. The ensemble of radio and satellite
receivers is connected using serial ASCII timecode, Inter-Range Instrumentation Group (IRIG), and PPS
signals to four primary time servers for the research laboratory and public at large. Figure 13.7 shows
Figure 13.6 LORAN-C receiver and OCXO. Austron 2201A GPS receiver Austron 2000 LORAN-C receiver
Spectracom 8170 WWVB receiver Hewlett Packard 5061A cesium beam frequency standard Figure 13.5
University of Delaware laboratory test equipment. 252 Computer Network Time Synchronization
auxiliary laboratory equipment used in performance experiments and performance evaluation. 13.13
Parting Shots You may have noticed that nothing has been said in this chapter about local time zone or
about daylight or standard time. This is intentional; there is nothing about NTP, or UTC for that matter,
that has anything to do with local time or spring leaps forward and fall leaps back.
This is the same philosophy practiced by mariners, aviators, and other long-distance runners; UTC is,
well, universal. Where local time becomes important, we expect the operating system to include
provisions to apply the correct offset. But, there are large corporations running IBM mainframes that
insist on local time, at least until they open a branch in Shanghai. The problem becomes acute on the
day of changeover between standard and daylight time. Spring is okay, as the clocks are stepped
forward 1 h in each time zone, so it is not just 1 h when message timestamps are inconsistent between
time zones, it is four hours as the ripple passes over the United States. It is even worse in the fall
because the same time can occur twice. When this happens, says the advice in at least one business
computer model, the Spectracom 8170 WWVB receiver Spectracom 8170 WWVB receiver Spectracom
8183 GPS receiver Hewlett Packard 105A quartz frequency standard Hewlett Packard 5061A cesium
beam frequency standard Figure 13.7 University of Delaware master clock facility.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
However, the timing of the NMEA message precesses over the 1-ms C/A code period. For more precise
timing, NMEA receivers also output a PPS signal aligned to the internal receiver clock. 13.12.6 LORAN-C
Radio Navigation System The U.S. Coast Guard, along with agencies of other countries, has operated the
LORAN-C radio navigation system for many years [23]. It currently Table 13.7 Global Navigation
Frequency Assignments Channel Frequency (MHz) Signals L1 1575.42 C/A, P(Y), M, L1C L2 1227.60 P(Y),
M, L2C L3 1381.05 Nuclear detection (NUDET) L4 1379.913 Ionosphere correction (future) L5 1176.45
Safety of life (aeronautical) 250 Computer Network Time Synchronization provides time-transfer
accuracies of less than a microsecond within the ground-wave coverage area of a few hundred
kilometers from the stations. All stations operate at 100 kHz using a pulse code modulation. Beyond the
ground-wave area, signal propagation is via the lower ionospheric layers, which decreases accuracies to
the order of 50 µs. The current deployment of LORAN-C stations permits almost complete coverage of
the continental United States, Alaska, and western Europe. A few notes about the navigation function
will help explain how the system can be used for time transfer. LORAN-C receivers measure the time
difference of arrival (TDOA) between pulses received from three or more transmitting stations. The locus
of points of a specified TDOA between two stations is a hyperbola with its axis the line connecting the
stations and focus the nearest station.
