[time-nuts] Re: Flutter Issues with the timebase in Philips PM6654C

[Magnus Danielson]

Hi Walter,

On 2021-04-02 15:24, Walter Shawlee 2 wrote:
> I recently got 3 of these excellent counters, all ex-NASA in excellent
> physical condition.
> They all have the PM9696 OCXO option installed, but they also all have
> the same odd problem:
>
> Measuring my very stable GPSDO (which is monitored with another PM6654C),
> the new ones constantly bounce around, with a pronounced rapid
> 0.1-0.3Hz flutter.
> The reference counter is rock solid, as is the scope monitor, down to
> 0.01Hz with no flutter.
> No adjustment or setting on the new counters will eliminate it.
>
> I am very puzzled by this effect, as I have never seen this kind of
> short term fault
> manifest in multiple units before. Is there anyone with experience
> with these high end
> counters that has seen this and knows the correction?
>
> my only tack at this point is to rip one apart, and go through
> everything from power
> supplies to the OCXO looking for clues.
>
> any advice appreciated. I must say I do love these counters,their
> performance is quite amazing
> when working properly. I also wish I had a source for the 1.5GHz C
> channel input options, as these
> had it missing.

OK, so the 10 MHz OCXO is the PLLed to a 500 MHz which is used for
counting, so the ECL counting runs at that rate, giving the 2 ns
resolution. I'll try to see if I can measure the same behavior in the
one I have access to.

It could be that the PLL is the source of the flutter/oscillation.
Looking at the phase-detector would be a very good start.

I've used the PM6654C for one measurement just to illustrate that you
can use that old counters and performance to track the issue at hand
that we had, or rather, a customer had with third party equipment. I was
able to run with TimeLab nicely, it was very quick to set that up.

I need to pick it up from work, as it is currently idling on my
work-bench there.

Cheers,
Magnus