A rubidium clock is a secondary frequency standard. It is only as accurate as what it was adjusted to when compared to a higher order standard. How did you set up your rubidium standard? If there is a noticable difference in the frequency of the GPSDO & the rubidium after a few days, I would suspect that it was the rubidium that was out, not the GPSDO. Also, it would have been intersting to have seen the configuration screen for the GPSDO, to see the number & quality of the satellites received.
I'm going to swap out the fan in my 53131 as you did but will still continue to keep it on a switched AC power strip since I'm not depending on the internal clock but the Rb standard instead.
It won't be accurate to 15 digits even if could display them. There's loads of stuff about the accuracy of frequency counters on the time-nuts mailing list. Have a look, the time lords go deep into this very fun stuff.
I have a GPSDO but no rubidium or cesium standard. For that reason I'd be very curious how they compare. I suspect the GPSDO would have better phase noise, while the rubidium has lower drift. Is this true? I'd be very interested to see how the two look together on a scope!
I have done some testing on this. GPSDO’s have an inherent issue where they get yanked around by the self-correcting nature of the clock. The MOST stable is a GPSDO feeding a 1pps in to a rubidium unit. The RB units have a better phase noise and “short term” accuracy and stability while the GPSDO’s have a much better long term stability but a worse short term stability due to self-correcting nature of the GPS lock. With a LONG integration time 48 hours for the RB unit the GPSDO can steer the RB but not yank it around. HOWEVER this is really getting down in to the weeds as I have found a standard Deviation of 115 uHz on 10 Mhz with a GPSDO only. And with the GPSDO coupled rubidium this dropped to 27uHz on the same 10Mhz carrier. So as long as frequencies of interest stay under a few Ghz both are REALLY down in the noise. The analysis was done on a 53131A with a new RB unit from Stanford Research, and a Samsung GPSDO module.
@Zenwizard Studios Very interesting. Thanks for the detailed response! I'll have to conjure a way to capture and identify the infrequent minor phase shift that would inevitably occur. I would be curious to see how often there's an adjustment. Though I could also likely look into the software I use to monitor it, it may report these adjustments
Why leave the Racal 1992 resolution set to 0.1Hz, when you could have compared it to the HP at the same resolution? A single press of the resolution UP arrow would have given resolution to 0.01Hz. Then another long press of the button would have given you 0.001Hz resolution.
Great video. The slight difference in counts could also be that the clocks are likely slightly out of phase. .007Hz is nothing to complain about. 😂
Hi mr imsai putting ones rubidium up the rear came as a surprise on a sunday
The Agilent counter with the rubidium reference is awesome at (.0001Hz) resolution.
Great video, makes me want a frequency counter.
A rubidium clock is a secondary frequency standard. It is only as accurate as what it was adjusted to when compared to a higher order standard. How did you set up your rubidium standard? If there is a noticable difference in the frequency of the GPSDO & the rubidium after a few days, I would suspect that it was the rubidium that was out, not the GPSDO.
Also, it would have been intersting to have seen the configuration screen for the GPSDO, to see the number & quality of the satellites received.
I'm going to swap out the fan in my 53131 as you did but will still continue to keep it on a switched AC power strip since I'm not depending on the internal clock but the Rb standard instead.
It won't be accurate to 15 digits even if could display them. There's loads of stuff about the accuracy of frequency counters on the time-nuts mailing list. Have a look, the time lords go deep into this very fun stuff.
Over a longer timer scale, the GPS is more accurate, it's phase locking onto signal from many atomic clocks.
I have a GPSDO but no rubidium or cesium standard. For that reason I'd be very curious how they compare. I suspect the GPSDO would have better phase noise, while the rubidium has lower drift. Is this true? I'd be very interested to see how the two look together on a scope!
I have done some testing on this. GPSDO’s have an inherent issue where they get yanked around by the self-correcting nature of the clock. The MOST stable is a GPSDO feeding a 1pps in to a rubidium unit. The RB units have a better phase noise and “short term” accuracy and stability while the GPSDO’s have a much better long term stability but a worse short term stability due to self-correcting nature of the GPS lock.
With a LONG integration time 48 hours for the RB unit the GPSDO can steer the RB but not yank it around. HOWEVER this is really getting down in to the weeds as I have found a standard Deviation of 115 uHz on 10 Mhz with a GPSDO only. And with the GPSDO coupled rubidium this dropped to 27uHz on the same 10Mhz carrier. So as long as frequencies of interest stay under a few Ghz both are REALLY down in the noise.
The analysis was done on a 53131A with a new RB unit from Stanford Research, and a Samsung GPSDO module.
@Zenwizard Studios Very interesting. Thanks for the detailed response! I'll have to conjure a way to capture and identify the infrequent minor phase shift that would inevitably occur. I would be curious to see how often there's an adjustment. Though I could also likely look into the software I use to monitor it, it may report these adjustments
Another way to get more digits is to supply higher frequencies. Try a PLL with the 10 MHz base.
If all else fails, read the manual ;-)
Yup. Reading the manual is like looking at the map. We only do it when we're lost :-)
Why leave the Racal 1992 resolution set to 0.1Hz, when you could have compared it to the HP at the same resolution? A single press of the resolution UP arrow would have given resolution to 0.01Hz. Then another long press of the button would have given you 0.001Hz resolution.
How does it produce a fraction of a Hz resolution output every second ?
filtering and statistics
and interpolation resource.npl.co.uk/docs/networks/time/meeting5/johansson.pdf
How does your 5316A compare?
When you have to check the accuracy of a frequency counter over 10 MHZ, example: 15 GHZ, do you use your rubidium 10 MHZ reference?
yes
@@IMSAIGuy If you use a diy rubidium 10MHZ reference, please share it with us. It's important to know how to build diy divices for calibration.
Merci Guy.
2:17 2:17 2:17 2:17 2:17 😮s 2:17 se 2:17 s
2:17 2:17 eee
😮 2:17 2:17 2:17 2:17 eese 2:17 ß 2:27 2:27
Nice
Manuals are boring.