Ah, I was unclear.  I'm adjusting the 10MHz output.  It was pretty far
out when I got it.

--
Paul

Hi

Ok, you are fine with the 10 MHz reading on the 5335 to get things "sort of close". Once you get past about 1.0x10^-10 you may want to switch over to a pps based approach. More or less:

Fire up 5335 with pps from gps / gpsdo and from the PRS-10. (I'm assuming it's a GPSDO)
Watch the reading for a few seconds, record the average and the time.
Go have lunch / nap / beer / what ever.
Check the time and the reading.
Do the math
Bump the PRS-10
Step / repeat at the couple hour time interval for a day or so.
Drop back to a couple days, check and (maybe) adjust.
Drop back to once a week, check and (maybe) adjust.

Note that the 5335 only has to be on while you are actually observing the pps. No need for the heat / wear and tear the rest of the time. Your pps needs to run all the time, but that's only a problem if there is a power outage.

The advantage of doing it this way is you get both high resolution and averaging at the same time. Since there's pretty much nothing to set up (no computers / no GPIB) it's reasonably foolproof. Assuming that the 5335 is doing it's thing and that the GPSDO is good to roughly the same level:

3 hours is 10800 seconds
your reading is good to 3 ns (2 on the GPSDO and 1 on the counter)
you  get +/- 3.0x10^-9 at 1 second
you get about +/- 3x10^-13 at 3 hours.

Since the ADEV of the Rb is much better at the longer Tau's, the data you are using is probably a better bet than anything you can do in less time.

Bob

On Sep 2, 2013, at 9:20 AM, Paul tic-toc@bodosom.net wrote:

This is a dumb question but why not let the PRS-10 set itself using the
1PPS input and an appropriate time constant?

Thanks,
Bob Darby

On 9/2/2013 12:09 PM, Bob Camp wrote:

Hi

1. Because it's way more fun to do it this way
2. Because some / most PRS-10's don't have the 1 pps input enabled in the firmware they have loaded into them
3. Because you have more control over when you do / don't bump the frequency (you can see what's going on).

Bob

On Sep 2, 2013, at 12:21 PM, Robert Darby bobdarby@triad.rr.com wrote:

On 09/02/2013 06:27 PM, Bob Camp wrote:

You don't need to use the PRS-10 internal algorithms, you can implement
it all yourself with a little program communicating with the PRS-10,
using it's time-stamping hardware and remote steering.

Cheers,
Magnus

Hi

That of course assumes the firmware is there to do the time stamp in the first place. There are a lot of them out there that simply have nothing at all on the pps input.

Bob

On Sep 2, 2013, at 8:03 PM, Magnus Danielson magnus@rubidium.dyndns.org wrote:

On 09/03/2013 02:14 AM, Bob Camp wrote:

Naturally. If you have that option, then it is a neat and low-complexity
way of getting started.

Cheers,
Magnus

Hi Bill
I guess technology moved along. I think of the mix type of syn like genrad
made 1164 maybe. Son of a guns to keep working and I think they are noisey.
Granted in the day they were not bad for what they were doing.
You could emulate much in todays technology. But I think cost becomes an
issue.
Regards
Paul.
WB8TSL

On Mon, Sep 2, 2013 at 10:50 PM, Bill Hawkins bill@iaxs.net wrote:

On 9/2/13 7:50 PM, Bill Hawkins wrote:

We do.. there's a fair amount of talk about boxes like the PTS
synthesizers.  And there's been talk about 866x series synths, and
perhaps the 3325, although I'd have to go search to be sure.