[time-nuts] Time syncing WiFi routers using FM radio

Bill Byrom time at radio.sent.com
Sat Nov 14 15:37:29 EST 2015

An improved technique using the 3.579(54) MHz NTSC color burst frequency distribution was described in this NBS circular a year or so later (1972, believe):

I was a General class ham (then WA5ZBJ) just entering the UT Austin EE
school in 1972. I had a WW2 surplus BC-221 precision tunable frequency
reference (which was an amazing device for it's time), and participated
in ARRL frequency measuring tests using it.

After I heard about the color burst distribution (probably through NBS
circular 1601) I took the color burst locked signal from an old tube
color television and built a circuit using TTL IC's (7490, 7492, and/or
7493 dividers, I think) to compare this signal with a 5 MHz ovenized
crystal oscillator. The NTSC color burst frequency is 3.579(54)
[repeating decimal] MHz, which is exactly 315/88 MHz. If you multiply
the color burst signal by 88/63 you get exactly 5 MHz. So I divided the
color burst signal by 63 and compared it to the 5 MHz divided by 88.
This worked well, and I could calibrate my 5 MHz oven crystal oscillator
against the network color burst signal.

This technique only worked when the local TV station was feeding a color
program directly from the network. This was before frame synchronizers,
so when the local station switched to a local video source the
television receiver might lose horizontal and color burst lock
momentarily, and my circuit was unusable due to the poor local station
frequency reference until they switched back to the network feed.

Several years later (late 1970's) I was working at a small company (Rhodes-
Groos Laboratories in Austin) which was interested in building a
commercial product to provide an alternative to WWVB frequency receivers
by using the color burst signal. NBS was publishing the monthly results
of their measurements of network color burst signal frequency error, but
remember that this was long before the Internet so you received these
reports in their monthly printed bulletin so you could determine after
the fact errors in measurements you made weeks before.

Unfortunately, some of the TV networks read the NBS bulletin and saw
that they had systematic frequency offset errors. So they started
manually adjusting their rubidium or cesium standards in an attempt to
walk them to a lower offset. Frame synchronizers using local TV station
frequency standards disrupted the direct analog distribution of the
network color burst signal to receivers, so we dropped the project and
never developed a color burst frequency receiver.

I'm also dubious about using FM broadcast radio broadcasts inside office
buildings. My experience is that these signals are very weak in interior
rooms, and RFI often makes reception impossible. As already noted,
multipath is a major difficulty at around 100 MHz in buildings, Last
week I tried to use my old Apple iPod to listen to FM broadcast band
signals while eating dinner at the Denver airport and had great
difficulty in positioning the iPod and headphone cable (which acts as
the antenna) for good audio due to multipath distortion.

Bill Byrom N5BB
On Thu, Nov 12, 2015, at 06:19 AM, Tim Shoppa wrote:
> I think the TV-network-synchronization NBS article you are talking about,
> is this: http://tf.nist.gov/general/pdf/237.pdf
> Tim N3QE
> On Thu, Nov 12, 2015 at 3:04 AM, Hal Murray <hmurray at megapathdsl.net>
> wrote:
>> kb8tq at n1k.org said:
>>> For time sync, broadcast signals are a pretty well studied topic. Sync
>>> signals from TV stations are a much better  ...
>> Many years ago, I think it was late '70s, a friend showed me a blurb from
>> NBS.  They were distributing time by piggybacking on NBC's signal.  NBC had
>> atomic clocks at their headquarters and a collection of links from the
>> phone
>> company running to all their stations.  The stations locked on to the
>> upstream signal.  So the whole system ran in lock step give or take some
>> propagation delay.  That was needed so they could switch between local and
>> network signals without trashing the picture.  That was long before frame
>> buffers.
>> I think it was HP that measured the signal in the Silicon Valley area.  NBS
>> published and distributed the offset.
>> Does anybody remember that booklet?  Did I get the story reasonably
>> accurate?
>> --
>> These are my opinions.  I hate spam.
>> _________________________________________________
>> time-nuts mailing list -- time-nuts at febo.com
>> To unsubscribe, go to
>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
> _________________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to
> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.

More information about the time-nuts mailing list