[time-nuts] HP 10811 versions for 5065A

[Magnus Danielson]

Hi Corby,

This is where the phase-noise properties becomes relevant, as the phase
noise at 137 Hz and 274 Hz will be relevant as it will add to the noise
of the physical package to be down-converted by the synchronous
detection as it goes into the loop, and through that affect the
stability for long-term. This is documented before, so it was only when
I did phase-noise measurements that it became so apparent.

I will make tests and try a few different strategies and see how that
compare up against H-maser, rubidiums and BVA.

I was amazed how clear the connection became in this case, and since I
had multiple 5065As to test on I could try both variants relatively quickly.

Turns out that being able to measure phase-noise can be a critical tool
to handle instability for long-term. However, you do not need to go very
deep on the phase-noise part to be able to handle the problems of most
dominant noise-source. Once you removed this as a dominant noise source,
you have something else to limit you and you need to address that. Which
is the easy win will depend on which is your limiting factor where you
are with that device.

Anyway, if and when I get progress in the lab on this, I will report. I
think the things I learned was useful enough at the time, but I was not
able to come to concrete modifications until my time was consumed by
other things.

I hope the observations and notes can be inspirational to some.

Cheers,
Magnus

On 2020-07-20 23:30, cdelect at juno.com wrote:
> Magnus and Tom,
>
> Thanks.
>
> I'm not (and don't want to be!) a phase noise nut!
>
> However I am an Allan Deviation nut and have documented many 5065A with
> the 10811 retrofit that drop into the -14ths between 64 and 128 Sec and
> the best show 4.5X10-14th at 1000 Sec.
>
> So I'd say the original circuit does not degrade performance that much!
>
> Unfortunately I can't play around with that 7474 stage and see if I can
> help performance as I not longer have access to an active Maser. :(
>
> Cheers,
>
> Corby
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.