[time-nuts] Favorite counters (current production)?

Attila Kinali attila at kinali.ch
Sun Nov 12 16:08:59 UTC 2017


On Sat, 11 Nov 2017 04:44:53 -0800
"Tom Van Baak" <tvb at LeapSecond.com> wrote:

> That said, I want to point out that the latest GPSDO / counter from Stanford Research continues their tradition of relatively open design. If you have an hour, go through the very detailed user manual, which includes theory of operation and BOM and schematics, just like the old days:
> 
> http://www.thinksrs.com/downloads/PDFs/Manuals/FS740m.pdf
> http://www.thinksrs.com/downloads/PDFs/Catalog/FS740c.pdf

Oh.. very nice. I wasn't aware of this beauty!

I only had a quick look at the schematics, but a few things stick out:

1) They use "half" Nutt interpolators for time stamping instead of
A-B measurements as is usual with the Nutt interpolator. I was wondering
why nobody did that, as it is kind of the obvious thing to do, today.

2) They use simple, time-to-amplitude converter for the interpolation.
The only interesting bit there, is the, in my opinion, weird way of
starting the charge using an opamp, that at the same time compensates
for the diode forward voltage.

3) This is the first time I have ever seen someone even mentioning
the possibility of metastability in counters or the Nutt interpolator.
And they solve it in a quite interesting way too.


4) Even though they call it TCXO in the manual, the normal oscillator
in the FS740 is an ovenized AT cut crystal. As it is built manually
on the main PCB itself, it's nothing fancy, but apparently does the
job quite well.

5) Every other clock source in the instrument is derived from the 20MHz
that this central oscillator produces. Even the 12MHz of the frontpanel
controller.

6) The OCXO and Rb options only phase lock the central oscillator with
a small PLL at 300Hz BW. That explains that the phase noise performance
is constant over all 3 options. Though I would have expected a higher
close in phase noise for the rubidium, as it does in the ADEV plot.

7) The Rb only performs slightly better between taus of 100 to 10k
than the OCXO.


I probably have to spend some time reading the schematics, which by the
way are superb. So much love to details, even small ones. You can almost
feel what kind of trade-offs the engineers made and for what reason.

			Attila Kinali
-- 
You know, the very powerful and the very stupid have one thing in common.
They don't alters their views to fit the facts, they alter the facts to
fit the views, which can be uncomfortable if you happen to be one of the
facts that needs altering.  -- The Doctor



More information about the Time-nuts_lists.febo.com mailing list