[time-nuts] Tbolt issues
Charles Steinmetz
csteinmetz at yandex.com
Sat Sep 10 06:40:43 EDT 2016
Bert wrote:
> would you please share your settings, this is exactly what we are looking
> for. We are doing it by trial and error but your expertise will help greatly.
Well, I spent the holiday weekend looking for the "safe place" where I
recorded my final Tbolt tuning parameters -- without success. After I
tuned them and qualified them as fit for long-term duty, I disconnected
the com ports and stashed them away in a very quiet and hard to access
location with just the 10MHz coax connected, so I can't use the comms to
extract the parameters without disturbing the Tbolts (which have now
settled nicely, undisturbed for ~10 years).
But what I did wasn't rocket science -- I just read up on the tuning
parameters, determined which ones would likely affect the stability of
the 10MHz output, made some educated guesses about the likely best
settings, and started playing. It took me several weeks of
experimentation (on and off), and the parameters I settled on were
somewhat different between the two units I kept (primarily, the loop
time constant and damping, which ideally should be set to complement the
particular OCXO in each unit).
I recommend extreme caution when you hear suggestions to use low loop
damping, or to monkey very much with the oscillator scale factor. I
found that high damping (far above the 1.2 default value) worked best
for my units. (Like you, I care most about the stability and accuracy
of the 10MHz output. I don't even have the PPS turned on.)
Before you do anything else, I strongly suggest a full factory reset to
put everything into a known state, and work from there.
You have received some advice to use the "autotune" routine in Lady
Heather. I seem to recall several people reporting that it worked well
for them, and there is nothing to lose by trying it. However, in my
case it screwed up the tuning of both units so badly that I had to do
factory resets and then re-enter my custom parameters. I might have
been using a version of LH that didn't have the latest autotune code, or
perhaps the autotune function needs to start from factory default
settings, or maybe the phase of the moon was wrong -- but I was sure
glad I had recorded the tuning parameters I worked out by experiment, so
I didn't have to start over again! Just be prepared to do another
factory reset and start experimenting if autotune doesn't work to your
liking.
So, I would suggest: (1) do a full factory reset, let it run for a few
weeks, and take data; (2) use the autotune routine, let it run for a few
weeks, and take data. If you are not satisfied with the results at this
point, (3) do another full factory reset and begin experimenting manually.
Best regards,
Charles
More information about the time-nuts
mailing list