(Now a bi-weekly program)
Feb 5, 2013
Join the "CWTD Yahoo Group" for
email discussion in between our bi-weekly sessions by clicking
What Time Is It?!!!!!
Precision Time Display & Synchronization Methods ... and the
Time … you can’t beat it, make it, save it, store it, replay it or out run
it. But we amateurs can measure time using a number of techniques for our
benefit in the shack, in our rigs, in our QSOs, and more. In this session we
overview the sometimes-esoteric world of timekeeping and time synchronization
available to us (WWV, CHU, GPS, Rubidium, Cesium, etc.) and then spend a good
amount of time on a unique Arduino-based project that can grow and be useful to
each of us.
73, George N2APB & Joe N2CX
to the MP3 podcast)
Notes: (Lots of discussion this time ...
scroll down for the start of the actual whiteboard.)
<20:10:15> "Joe N2CX": We need to have a 3-week
gap (instead of 2) for the next show, to occur on Feb 26th
<20:12:58> "Jim N2GXJ": Yes, Gloucester County ARC tomorrow with N2APB & N2CX
presenting. Visitors welcome: details on web site
<20:17:22> "Alan W2AEW": A broken analog watch is exactly right twice a day!
<20:17:35> "Pete - WB2QLL": Telephone time services have been around for years.
<20:17:55> "George - N2APB": Great point about the telephone time service ... I
recall doing that when I was a kid!
<20:18:00> "John - NG0R (Mac)": Funny, something before NTP (network time
protocol) time sync for our phones & computers
<20:18:28> "Pete - WB2QLL": WU had leased phone lines for their special clocks
<20:18:58> "Ray K2ULR": The WU clocks were common at broadcast stations.
<20:19:27> "Pete - WB2QLL": That's what I was thinking of, but all sorts of
other businesses used them.
<20:19:42> "Rick K3IND": Great Britain offered a large prize for an accurate
clock for use on board ship in the 18th century.
<20:19:52> "Joe N2CX": My mistake CCW used 100 ms time windows
<20:25:36> "George - N2APB": And we have some big "time measurement" events even
today. The "Time Nuts" group (on Yahoo Groups, I think) is very intense about
<20:27:21> "Howie - K3HW": CHU no longer on 7335 ?
<20:27:25> "Bill - KD5TFD": TimeNuts info:
<20:28:45> "Bill - KD5TFD": CHU hearable in Tx w/ a wire antenna pretty easily
<20:30:09> "Howie - K3HW": Rick K3IND - Yes - clock used to determine longitude
at sea - man's name was Harrison
<20:30:51> "Pete - WB2QLL": CHU hasn't been on 7335 for years and years.
<20:31:43> "Pete - WB2QLL": Changed in 2009 because of band reallocations.
<20:32:01> "Bill - KD5TFD": chu moved in 2009 -
<20:32:02> "Howie - K3HW": Tnx Pete - shows how long ago I have been above 40
Meter CW band
<20:32:24> "Alan W2AEW": CHU on 3330kHz very strong in NJ now
<20:32:44> "Howie - K3HW": used to use 7335 as an approximate marker for the top
of the 40 meter phone band
<20:33:33> "Howie - K3HW": cesium beam standard used on many Navy Warships since
<20:35:44> "Chris KD4PBJ": I saw an article in an electronics magazine a while
back about building a device to extract time pulses from TV as their network
shows were locked to a precise clock. I assume this is no longer possible now
that NTSC has gone away and everything is DTV now?
<20:37:04> "John - NG0R (Mac)": how do systems sync to WWV/WWVB? Are they using
the leading (rising) edge for the 1 second sync?
<20:37:53> "Terry WB4JFI": Not all TV stations had precise clocks in the old
<20:38:20> "Clint-ka7oei": WWVB RX's sync to the beginning of the bit, which is
signified by a 17dB power increase.
<20:38:34> "Joe N2CX": NG0R the time codes on WWVB have particular edges that
are the active sync locations. Check out the WWVB link at the end of the
<20:38:34> "John - NG0R (Mac)": I remember working at TV stations that pulled
down our network time via the network sat feed.
<20:39:20> "Ray K2ULR": Time of day is usually embedded into the video signals
that originate at the various TV network facilities.
<20:40:01> "Ray K2ULR": Even when a network is "in black"... it's still sending
<20:40:05> "Clint-ka7oei": Several of the local DTV stations' clocks aren't much
more accurate than a typical computer's clock - terrible for those who have DVRs
synced to them! They are usually tied to time from the internet, though...
<20:40:12> "Howie - K3HW": Tnx for link on CHU article Bill KD5TFD
<20:41:27> "Todd K7TFC": Just came in . . . where are we on the white board?
<20:41:45> "Ray K2ULR": And you cannot depend on accurate timing, once the TV
Network signal is distributed by cable networks. Over-the-air signals should be
<20:41:57> "Alan W2AEW": Todd - down at the CWTD Project, showing the tone
<20:42:40> "Todd K7TFC": Thanks, Alan!
<20:43:03> "Terry WB4JFI": Most Network TV is now distributed via satellite,
destroying accurate time, unless corrected for. Many TV stations use "GPS
Ageless Master Oscillators" if they don't have separate Cesium or Rubidium
devices. DTV usually references a GPS-refrenced oscillator.
<20:44:56> "Ray K2ULR": And then there's the 8 second delay needed for "HD"
radio... throwing off any "time tone" by 8 seconds... as with KYW in Philly.
<20:45:20> "Joe N2CX": I missed the code since I did not realize that the text
was actually a hyperlink... My bad!
<20:52:29> "Todd K7TFC": Somewhere in the various Yahoo groups there's been a
project for a dedicated WWV receiver. I'll look for it and get back.
<20:52:49> "Joe N2CX": Software always takes time go get working right...
<20:53:39> "Joe N2CX": QRP-tech yahoo group had a WWV receiver deisgn by Chuck
<20:53:51> "Terry WB4JFI": That "CAL" option is a great idea!
<20:54:59> "Todd K7TFC": Okay, it's a KD1JV design: http://kd1jv.qrpradio.com/WWVRX/WWV%20receiver.htm
<20:55:39> "John - NG0R (Mac)": nice link (KD1JV)
<20:59:33> "Joe N2CX": RTC stands for Real Time Clock - gotta love acronyms...
<21:00:40> "Bill - KD5TFD": wwv's not bell103
<21:06:10> "Todd K7TFC": Phase "7": multiple LCD displays, each with selectable
<21:06:37> "Bill - KD5TFD": phase 8: Big LED display board
<21:07:08> "Todd K7TFC": Yes!! Something that glows in the dark!
<21:07:08> "Terry WB4JFI": Phase 9: fits on your wrist...
<21:07:10> "Bill - KD5TFD": Phase 9 .... broadcast WWVB signal to other atomic
clock in house
<21:07:53> "Howie - K3HW": Phase "10" NJQRP Predator Drone
<21:08:47> "John - NG0R (Mac)": phase 10, nice, very timely
<21:10:41> "Todd K7TFC": Isn't it a regular Hitachi HD44780 compatible?
<21:11:25> "Joe N2CX": Yes the LCD is compatible.
<21:11:49> "John - NG0R (Mac)": The LCD is being used in the Parallel mode
<21:12:14> "Joe N2CX": Yes parallel mode
<21:12:41> "Todd K7TFC": 8 bit or 4 Bit?
<21:12:44> "John - NG0R (Mac)": the I2C backpack is a nice addition to it for a
4 wire (2 power, 2 data) interface
<21:15:23> "John - NG0R (Mac)": it would be neat to setup the project/shield to
compare WWV/CHU and something like NMEA GPS.
<21:16:07> "John K2ZA": Could lat/long location data be used to compensate for
speed of light delay from CHU/WWV?
<21:16:38> "Bill - KD5TFD": think there'd be a bit of variability based on
<21:17:31> "Joe N2CX": Yes you can approximate propagation time and add that to
the time offset.
<21:17:54> "Todd K7TFC": Isn't *precision* more important than *accuracy*?
<21:18:23> "Bill - KD5TFD": but would you not expect the propagation delay to
vary based on conditions ?
<21:19:14> "John K2ZA": I would think for skywave propagation that would be the
case but for ground wave is should be more uniform.
<21:19:46> "Todd K7TFC": Repeatability is the essence of precision.
<21:19:50> "Howie - K3HW": John K2ZA - we used lat and long posit to correct for
time delay when syncing our crypto gear at sea
<21:19:53> "John - NG0R (Mac)": I think that prop conditions are are going to
have more impact on the stability (fading) of the signal, but very little on the
<21:20:55> "Howie - K3HW": the north atlantic is a long way from colorado
<21:22:47> "Bill - KD5TFD": Does Radioshack still make time cube radios?
<21:22:57> "Howie - K3HW": cesium/rubidium standards used by USN were
correctable/adjustable in nanoseconds
<21:23:08> "Joe N2CX": I think they dropped them several years ago.
<21:23:34> "Joe N2CX": Howie all fine and good but what do you use as a time
reference to set them?
<21:23:35> "Bill - KD5TFD": Bummer .. output of one of those into the gizzmo to
decode the time code would be a neat hack
<21:24:16> "John K2ZA": George are you aware of Fritzing?
<21:24:32> "John - NG0R (Mac)": The DDS on a shield would have some sales demand
<21:24:44> "Howie - K3HW": Joe - we set them in port at new london ct with lab
standards and corrected at sea - as soon as the crypto locked we were good...
<21:24:47> "Joe N2CX": BTW there are published studies on the time variation
over day/night for WWVB
<21:25:19> "Joe N2CX": Howie - good I'll schlep up to New London to set my
<21:25:34> "Howie - K3HW": Oy Vay
<21:25:44> "N8WQ - Al": great session George and Joe
<21:25:49> "John - NG0R (Mac)": a field trip to visit Howie?
<21:26:40> "Howie - K3HW": John NG0R - I was out of New London CT 30 years ago -
I am in PA now...
<21:27:43> "Rick K3IND": What is the 'special' price for the Rookey?? ... $10
<21:32:03> "Mark KC4GIA": They've been having a series of articles about 'fritzing'
in nuts and volts magazines for a few months now.
<21:32:42> "John - NG0R (Mac)": Fritizing is an ok package. Makes for nice
pictures showing the physical wiring diagram for blog postings
<21:33:14> "Eldon - WA0UWH": I have used Fritizing, I like it.
<21:34:01> "John - NG0R (Mac)": Fritizing is a multi (os) platform so that is a
big Pro for it
<21:34:21> "Bill - KD5TFD": just run the win 7 time sync stuff more often?
<21:34:38> "John - NG0R (Mac)": Google NTP
<21:35:29> "Ray K2ULR": Macs update time continually.
<21:35:33> "John - NG0R (Mac)": most of the current OS have a NTP client, you
should be able to set the duration of the updates and the sources
<21:35:59> "Bill - KD5TFD":
<21:36:10> "John - NG0R (Mac)": NTP = network time protocol ... Google will be
<21:41:43> "Todd K7TFC": Has anyone tried the beta release of the TeamSpeak
client for Android that's in the scrolling news at the bottom.
<21:41:45> "Don NJ2E": nj2e
<21:44:09> "Terry WB4JFI": Digikey has the DS1307 8-pin DIP for $4.02, 8-pin
SOIC for 3.05 in single quantities
<21:45:56> "Pete - WB2QLL": GPS time has not kept up with the leap seconds added
to UTC; however a steering signal is broadcast to GPC receivers. See explanation
<21:48:12> "Howie - K3HW": Tnx Joe N2CX and George N2APB - Great topic... BTW -
my son David KB2LFN received an IC-718 for Hanukah - He is interested in PSK-31
and I will be setting him up with a NUE-PSK modem !!!
Precision Time Display & Synchronization Methods
... and the Arduino
Background of time keeping and measurement … “In the old days
Sunrise/sunset in agrarian times
Factory and school alarms
Radio and TV broadcasts approximately correct
HF/LF government time stations
Why is “reasonably accurate” time important to us hams?
Keeping track of local time DST, UTC
Scheduling operating events – contests, “skeds”, etc.
Needed for accurate syncing of modes like WSPR
Older mode CCW synced to 10ms windows
When is very accurate time used by us?
Traffic (message) handling
When is extreme accuracy needed?
Special weak signal operating modes: JT65A
Timing, not TOD
What are the various sources for time synchronization (NA)
WWV, WWVH, WWVB
Atomic standards (Rubidium, Cesium)
GPS derived from them
Internet NIST and Naval Observatory
Specifics of WWV (we’ve got a separate section)
Specifics of CHU (we’ve got a separate section)
“Time Products” available to us
GPS receivers with time/1 pps output
“Atomic” clocks (actually synced to WWVB
MFJ 132 RC and more
La Crosse Technology (Fry's)
Old Heathkit GC-1000 “World's Most Accurate Clock”
GPS modules for computer interface
“Time Projects” we can build
The CWTD Arduino Clock … A simple demonstration and
training accessory that can evolve into an ongoing useful addition to the ham
Modular – start with simple manual clock add software and hardware
Interface RTC ic for more accuracy stability
Other modules to sync to time standards sources
BROADCAST TIME STANDARDS
WWV, WWVB, CHU
[From Wikipedia] ...
WWV is the
National Institute of Standards and Technology's (NIST)
radio station in
Fort Collins, Colorado. WWV continuously transmits official U.S. Government
signals on 2.5, 5, 10, 15 and 20
frequencies and time signals are controlled by local
clocks traceable to NIST's primary standard in
Boulder, Colorado by
common view observations and other
transfer methods. NIST also operates the very similar radio station
Hawaii. WWV and
WWVH make recorded announcements; since they share frequencies, WWV uses a male
voice to distinguish itself from WWVH, which uses a female voice. They also make
other recorded announcements of general interest, e.g., the
satellite constellation status and severe oceanic weather warnings. WWV shares
its Fort Collins site with radio station
WWVB that transmits
carrier and time code (no voice) on 60
kHz in the
frequency (LF) band.
Standard Frequencies ....
WWV and WWVH transmit 44 seconds
of audio tone in most minutes
. It begins after the 1-second
minute mark and continues until the beginning of the WWVH time
announcement 45 seconds after the minute. Even minutes (except for
minute 2) transmit
is heard during odd minutes. The tone is interrupted for
40 ms each second by the second ticks. WWVH is similar, but
exchanges the two tones: 600 Hz during even minutes and 500 Hz
during odd. WWV also transmits a
440 Hz tone
commonly used in music (A440
) during minute :02 of each hour, except for the first
hour of the UTC day. Since the 440 Hz tone is only transmitted once
per hour, many
may use this tone to mark off each hour of the
day, and likewise, the omission of the 440 Hz tone once per day can
be used to mark off each twenty-four hour period. WWVH transmits the
same tone during minute :01 of each hour. No tone is
transmitted during voice announcements from either WWV or WWVH; the
latter causes WWV to transmit no tone during minutes :43–:51
(inclusive) and minutes :29 and :59 of each hour.
Likewise, WWVH transmits no tone during minutes :00, :30, :08–:10
and :14–:19. Audio tones
and other voice announcements
are sent at 50% modulation.
Digital Time Codes ... Time of day is also continuously transmitted
using a digital
, interpretable by
. The time code uses a 100 Hz
of the main signal. That is, it is an additional
low-level 100 Hz tone added to the other AM audio signals.
This code is similar to, and has the same framework as, the
IRIG H time code
and the time code that
transmits, except the individual fields of the code are rearranged
and are transmitted with the least significant bit sent first. Like
the IRIG timecode, the time transmitted is the time of the start of
the minute. Also like the IRIG timecode, numeric data (minute, hour,
day of year, and last two digits of year) are sent in
(BCD) format rather than as simple binary
integers: Each decimal digit is sent as two, three, or four bits
(depending on its possible range of values).
[Excerpts from Wikipedia] ...
CHU transmits 3
kW signals on 3330 and 14670
and a 10 kW signal on 7850 kHz.
These are nonstandard time signal channels; however, due to the
prospect of interference with WWV and WWVH, the nonstandard
frequencies are necessary. The signal is
, with the lower
H3E). The same information is carried on all three
frequencies simultaneously including announcements every minute,
alternating between English and French. The CHU transmitter is
, 15 km southwest of
central business district.
The primary time signal is a series of 300 ms-long
tones, transmitted once per second, on the second. The
following exceptions to the pattern provide additional information:
- The top of the minute is marked by a half-second-long beep.
- The top of the hour is marked with a one second-long beep,
followed by nine seconds of silence.
- The 29th second of a minute is always omitted (no beep).
- Between one and sixteen seconds past the minute (except at
the top of the hour), CHU transmits the difference between
Coordinated Universal Time (UTC) by using split tones. For
positive DUT1 values from +0.1 to +0.8 s, seconds 1 through 8
are split. For negative DUT1 values from −0.1 to −0.8 s, seconds
9 through 16 are split.
- Between 31 and 39 seconds past the minute inclusive, the
once-per-second tones are reduced to 10-millisecond "ticks"
while a digital time code is transmitted. The digital time code
is formatted so that a
Bell 103-compatible 300-baud modem can decode it,
and CHU is the only time signal station that uses this format
for its time code transmissions.
- For the last 10 seconds of each minute (seconds 50 to 59),
the once-per-second tones are again cut to 10 milliseconds each,
while CHU transmits a brief voice station identification,
followed by voice announcements of the next minute in UTC,
English. French announcements, using the voice of
Radio-Canada news anchor
Simon Durivage, are transmitted first on the odd minutes,
while English announcements, voiced by former
CBC Radio announcer
Harry Mannis, come first on the even minutes.
The digital time code sends 10 characters at 300 bits per second
asynchronous serial communication. This follows the
Bell 103 standard, a 2225 Hz tone to represent a mark (1 bit)
and 2025 Hz tone for a space (0 bit). Immediately after the 10 ms
tick, a mark tone is sent until 133.3̅ ms, then 110 data bits,
ending at precisely 500 ms. The final stop bit is extended by 10 ms
of mark tone to ensure it is detected reliably, and the final 490 ms
of the second are silent.
The time of day (day of year through second) is transmitted twice
during each second from 32 to 39. During second 31, additional
information (year, DUT1, daylight saving time, and leap second
warning bits) is transmitted.
Experience with CHU Reception
experience with CHU is that the 7.850 MHz signal is pretty good within 600 miles
of Ontario during most of the day and great at night. The 3.33 MHz signal is
good out to 200 miles or so during the day and up to 1000 miles at night. The
14.67 signal is fairly good for daytime use from 500 miles out, depending on
The Arduino requires some sort of synchronization/updates to maintain minimal
accuracy. For example, you need a 10 MHz clock to be accurate/stable
within 10 Hz to hold 2.5 seconds per month, 43 Hz for 2.5 seconds per week or
300 Hz for 2.5 seconds per day. Most simple computer clocks are not in that
league. The software doing the synchronization is important too since errors
are cumulative a little bobble in an interrupt driven program can drive a clock
(and the designer) bonkers!
ATOMIC TIME STANDARDS
Rubidium, Cesium, GPS
From Wikipedia ...
A rubidium standard or rubidium atomic clock is
frequency standard in which a
hyperfine transition of
rubidium-87 atoms is used to
control the output frequency. It is the most inexpensive,
compact, and widely used type of
atomic clock, used to control the
cell phone base stations, in test
global navigation satellite systems
GPS. Commercial rubidium clocks
are less accurate than
cesium atomic clocks which serve
primary frequency standards, so
the rubidium clock is a
secondary frequency standard.
However, rubidium fountains are currently being developed that
are even more stable than caesium fountain clocks.
All commercial rubidium frequency standards operate by
crystal oscillator to the rubidium
hyperfine transition of 6 834 682 610.904 324 Hz. The amount of
light from a rubidium
discharge lamp that reaches a
photodetector through a resonance
cell will drop by about 0.1% when the rubidium vapor in the
resonance cell is exposed to
microwave power near the
transition frequency. The crystal
oscillator is stabilized to the rubidium transition by detecting
the light dip while sweeping an
synthesizer (referenced to the
crystal) through the transition frequency.
Following from ....
Devices to measure time are incredibly interesting - especially
electronic devices (mechanical clocks are - ummm - too mechanical for me, I
fear). When I was a child I built some simple digital clocks either based on
counting the zero crossings of the power line signal (50 Hz in Germany) or
relying on a quartz (one day someone donated a 100 kHz quartz for my clock -
at least the quartz has survived, the clock is long since gone, taken apart
for parts and lives on in many different incarnations).
But even with a good quartz, maybe even a oven stabilized quartz, it is
difficult to built a clock with a stability of better than 10 ** -9. So I
dreamed about having my own atomic clock for many years, but getting a
cesium primary standard is highly unlikely (although there are enthusiasts
like Tom Van Baak, one of the Time Nuts,
who owns not only multiple cesium clocks but also hydrogen masers!) and even
getting a rubidium secondary frequency standard was out of the question.
Luckily this has changed - a couple of days ago I found an EFRATOM (now
Symmetricom) LPRO-101 rubidium
oscillator at a large online auctioning house for a very reasonable price.
With this oscillator as the basic building block for a really stable clock I
decided to built yet another clock (my last digital clock design was at
least 20 years earlier, so this will be a journey into my own childhood :-)
The picture above shows the clock in its current state - since this is a
work in progress this picture will change as the web page will be extended
rubidium oscillator is a completely self contained unit requiring only
minimal external circuitry to be used in a generator/counter/clock circuit.
Since the device gets quite hot it is important to mount it onto a heat sink
or metal frame to get rid of the excess heat (read the respective section of
the manual - for bench use a large heat sink is recommended).
See an entire program on this: CWTD Episode #30 ....
The CWTD Arduino Clock! ...
A simple demonstration and
training accessory that can evolve into an ongoing useful addition to the ham
CWTD Clock Software ... load using
the Arduino 1.0 development environment
The main circuit we use in our Phase 0 implementation is that from the Rookey
Phase 0: Tone
detection at top of minute from WWV or CHU
Phase 1: Addition of 1pps input signal from GPS for seconds sync
Phase 2: Addition of RTC chip to maintain date/time
Phase 3: Addition of NMEA input from GPS for date/time setting & sync
Phase 4: Addition of USB serial input from PC for sync via Internet
Phase 5: Addition of MC145442 modem chip to decode WWV digital data
- 1 pps output
- Alarm function
2) WWVB <http://www.nist.gov/pml/div688/grp40/wwvb.cfm>
3) WWVB Clocks <http://www.nist.gov/pml/div688/grp40/radioclocks.cfm>
4) CHU <http://www.nrc-cnrc.gc.ca/eng/services/time/short_wave.html>
5) Arduino Uno R3 …
6) LCD Shield Kit …
7) Prototype Shield: Maker Shed:
8) LCD …
order from the CWTD Project Room …
9) Rookey + LCXD Combo from Midnight Design .... If you don’t already have a Rookey Kit, this project will be a good
set of parts to use in this Arduino project. We reduced the price of the Rookey
Kit to make this a good deal for our Arduino project and you’ll benefit from
getting it along with the LCD (if you need either.) You can order this from the
CWTD Project Room as well, of course.
COMMERCIAL TIME PRODUCTS
1) PIC WWVB Clock ...
2) Decoding WWBV on Arduino with CMMP-6P module from Digi-Key ...
3) CMMR-6P data WWVB receiver chip
BEST ARDUINO REFERENCES
And here again are the
Arduino references from last week. These are very useful - hence the
repeat listing ...
Arduino Home Page ...
Arduino Software Download Page ...
Arduino Tutorial by ladyada.net ...
102 Projects from Make Magazine ...
Reference Books ...
The Definitive Guide to (All) Arduino Reference Books ...
The Arduino Cookbook ...
Programming Arduino - Getting Started With Sketches
30 Arduino Projects for the Evile Genius ...
Purchase the Arduino at ...
Adafruit Arduino Starter Pack ...
Maker Shed ...
Atmega AT328P chip with programmed-bootloader ...
(For a complete overview and comparisons of boards, see the
index of Arduino
Back to CWTD Home