Author Topic: Need advice on GPSDO behaviour please  (Read 1561 times)

0 Members and 1 Guest are viewing this topic.

Offline Mike99Topic starter

  • Frequent Contributor
  • **
  • Posts: 307
  • Country: gb
Need advice on GPSDO behaviour please
« on: February 10, 2025, 06:02:43 pm »
Hi everyone, I've just built my third 10MHz GPSDO. Does this make me a timenut?  :)

It's this one: https://ve2zaz.net/GPS_Std_Modernized/GPS_Std.htm

My GPS receiver is a ublox neo-M8T running in Fixed mode after surveying-in to better than 0.5 metres accuracy.

The OCXO is a Blilely NVG47A of unknown ancestry with a measured tuning slope of 3.5Hz/Volt.

The samples of frequency deviation and DAC value in the graph were taken at 10 second intervals.

Now I know that an OCXO that hasn't been used for a while can take time for the tuning voltage to settle, but my main concern is the alarming jumps in frequency.

Any ideas what might be causing this please?

Mike

 

Online edpalmer42

  • Super Contributor
  • ***
  • Posts: 2528
  • Country: ca
Re: Need advice on GPSDO behaviour please
« Reply #1 on: February 10, 2025, 06:34:31 pm »
Both frequency & phase jumps as well as rapid aging are common when an OCXO is working out the kinks after its long sleep.  Aging should slowly decrease over the coming days and weeks.  The frequency jumps should also fade away, but some OCXOs never really settle down, no matter how long you wait.

You could try isolating the OCXO from the disciplining function and see what happens.  If the frequency jumps disappear, then there must be glitches on the DAC line.  Aging, as reflected by the change in DAC will continue, but due to the isolation, the frequency actually will drift.  As long as the rate of aging is decreasing or stable, that's expected.

How are you measuring the frequency?  Good quality meter?  Oscilloscope? Have you checked your trigger level settings?  Have you considered adding a low pass filter to the input of your measurement device to reject external noise?

Welcome to the rabbit hole!
« Last Edit: February 10, 2025, 06:42:57 pm by edpalmer42 »
 

Offline Mike99Topic starter

  • Frequent Contributor
  • **
  • Posts: 307
  • Country: gb
Re: Need advice on GPSDO behaviour please
« Reply #2 on: February 10, 2025, 08:37:14 pm »
That's very helpful, thank you.

The frequency deviation is reported by the GPSDO software, I'm not measuring it independently.

I have a new Vectron I can use to confirm that it's the OCXO and not some other part of the system that's the issue.

Mike
 

Offline bingo600

  • Super Contributor
  • ***
  • Posts: 2315
  • Country: dk
Re: Need advice on GPSDO behaviour please
« Reply #3 on: February 10, 2025, 08:57:02 pm »
Note that the author is using a FLL loop , not a PLL
 

Offline rf-messkopf

  • Frequent Contributor
  • **
  • Posts: 323
  • Country: de
  • Mario H.
    • Homepage
Re: Need advice on GPSDO behaviour please
« Reply #4 on: February 10, 2025, 09:39:47 pm »
Might this also be an artifact of the way the circuit measures frequency? As already mentioned by others, you should verify if these glitches are real, and if its the OCXO or something else.

What I find odd are the rather large and infrequent jumps of the DAC value. Is that behavior of the disciplining algorithm intentional?

For comparison, see a similar test with my OCXO project: https://www.mariohellmich.de/projects/gpsdo/gpsdo.html (scroll to the bottom of the page). This is with an actual PLL, based on a high resolution phase measurement between the GPS time pulse and the OCXO output divided down to 1 Hz with a small FPGA, and an 18-bit DAC.
 

Offline Mike99Topic starter

  • Frequent Contributor
  • **
  • Posts: 307
  • Country: gb
Re: Need advice on GPSDO behaviour please
« Reply #5 on: February 10, 2025, 09:51:42 pm »
Yes I did wonder if it was the way that the deviation is calculated by the GPSDO software but I spoke to the designer and he hasn't seen this behaviour. I'm limited in what I can measure accurately so I think the first thing to do is try another OCXO.

Mike
 

Online edpalmer42

  • Super Contributor
  • ***
  • Posts: 2528
  • Country: ca
Re: Need advice on GPSDO behaviour please
« Reply #6 on: February 10, 2025, 10:06:38 pm »
Yes I did wonder if it was the way that the deviation is calculated by the GPSDO software but I spoke to the designer and he hasn't seen this behaviour. I'm limited in what I can measure accurately so I think the first thing to do is try another OCXO.

How long has it been running?  I see that the NV47 OCXOs are speced for aging after 7 days of operation.  That's when they're fresh from the factory.  I'd just watch it for at least a couple of weeks and see how things go.

If the OCXO frequency is galloping along, it could confuse the disciplining.  You'd end up chasing your tail just because you didn't let the OCXO settle.

It wouldn't hurt to power up your other OCXO now so that it has a chance to settle in case you decide to swap them later.
 

Offline Mike99Topic starter

  • Frequent Contributor
  • **
  • Posts: 307
  • Country: gb
Re: Need advice on GPSDO behaviour please
« Reply #7 on: February 10, 2025, 10:20:19 pm »
OK thanks, I will do as you advise. I also have a new Isotemp coming so I can try all three.

Mike
 

Offline lucido

  • Contributor
  • Posts: 41
  • Country: es
Re: Need advice on GPSDO behaviour please
« Reply #8 on: September 29, 2025, 11:20:13 am »
Good morning, Mike99.

I duplicated the ve2zaz gpsdo, In my case, it works correctly, without the drifts you show, in my gpsdo the oscillator is morion MV89.
I wanted to ask you something. I noticed you posted a very well-made graph. If you could explain how you did it, what program do you use, and if you could give me and everyone else some more information, please explain.
Thank you very much.
« Last Edit: September 29, 2025, 05:07:13 pm by lucido »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf

 

-->