[time-nuts] FMT on October 13

John Ackermann N8UR jra at febo.com
Tue Sep 25 08:01:10 EDT 2007

Tom Van Baak said the following on 09/24/2007 08:57 PM:

> I would think this is especially true for non-local frequencies,
> such as one received over the air. I'll leave it to you FMT
> guys to comment on the magnitude of degradation due to
> transmission and reception noise.

Absolutely.  Probably the best real-world performance you can get with a
skywave signal is on the order of 0.01 Hz.  Propagation effects play
havoc, but the longer the averaging period, the more short-term effects
will average away.  One of the reasons for the fairly long transmission
periods is to both allow longer averaging, but also provide the
opportunity to observe the atmospheric conditions at work.

> While were at it, in the case mentioned above I'm a curious
> about their FMT frequency standard -- if it's really accurate
> to parts in 10^12, as they imply, over 10 minutes. I could
> believe this if it were an Rb or Cs-based GPSDO.

We're using an Austron 1250A OXCO that's been measured as better than
9x10e-13 for averaging times of 1 second out to 1000 seconds; over a
broader range, it's better than 3x10e-12 from 0.1 seconds to 40,000 seconds.

Now, an important point -- we're not trying to trim the Austron to be
precisely on frequency.  We're going to let it run at whatever offset it
happens to be.  That will help make sure that the signal doesn't have
lots of zero's at the end, even though the resolution of the
synthesizers driving the transmitters is limited to 0.1 Hz.

We'll be comparing the Austron against a Z3801a (via my TSC-5120A
analyzer) and logging the frequency difference for at least several
hours prior to the test until several hours following.  The TSC gives 16
digits over 1000 seconds; depending on how much jitter we see, we'll
probably throw away the last two or three.  Even though the Z3801A may
be wandering around a bit, with successive 1000 second measurements we
should have confidence in the actual frequency over 1000 second periods
to at least parts in the 12s, ultimately limited by the Austron's
stability.  But since that's known to be in the 13s over the averaging
period of interest, we think we're safe in claiming accuracy and
stability of parts in the 12s.

Tom, if I'm missing something in this analysis, I'm seriously open to

By the way -- the synthesizers used to drive the transmitter amplifiers
will be PTS 250 SX-51 low noise units, so hopefully the transmitted
signals will have a better-than-the-average-ham-rig phase noise.  The
synthesizers will directly feed the driver and final amplifier stages of
some vintage Kenwood TS-520 ham transceivers with no other mixing --
it'll purely be the synthesizer and a transistor buffer amp driving two
vacuum tube stages to get up to about 75 watts (the rigs can run 100
watts, but we're derating -- and adding fans -- to support the long
key-down times).


More information about the time-nuts mailing list