Author Topic: FS: high accuracy Rubidium Frequency Standard, with 10M, 60M, 0~20M and 1pps  (Read 1546 times)

0 Members and 1 Guest are viewing this topic.

Offline testpoint1

  • Regular Contributor
  • *
  • Posts: 194
  • Country: us
by gps disciplined (more than 24 hours result, less than 24 hours the value will NOT been stored to EEPROM);

price is 750 USD, https://www.ebay.com/itm/184036643084

if you found a complete atomic clock's price is less than 1K USD and the accuracy can reach mine, just tell me.

 

Offline testpoint1

  • Regular Contributor
  • *
  • Posts: 194
  • Country: us
if you need the Spectratime made, just check here:

https://www.ebay.com/itm/184050224670

price is 1600 USD.
 

Offline testpoint1

  • Regular Contributor
  • *
  • Posts: 194
  • Country: us
« Last Edit: January 27, 2020, 11:20:21 pm by testpoint1 »
 

Offline testpoint1

  • Regular Contributor
  • *
  • Posts: 194
  • Country: us
I also list one SRS PRS10 for sell, calibrated, with all designed accessories, 599 USD:

https://www.ebay.com/itm/Used-SRS-PRS10-Rubidium-Oscillator-with-all-interface-GPS-board-etc/184121607183
 

Offline jpb

  • Super Contributor
  • ***
  • Posts: 1664
  • Country: gb
Hi testpoint1,

I'm not a customer yet as I already have 2 working Rubidiums and 4 GPSDOs but I follow your measurements with interest.

One question, what do you use as a reference in your measurements? You may have already stated this somewhere but I didn't see it.
 

Offline testpoint1

  • Regular Contributor
  • *
  • Posts: 194
  • Country: us
Hi testpoint1,

I'm not a customer yet as I already have 2 working Rubidiums and 4 GPSDOs but I follow your measurements with interest.

One question, what do you use as a reference in your measurements? You may have already stated this somewhere but I didn't see it.

verified by this:
« Last Edit: January 12, 2020, 04:02:52 pm by testpoint1 »
 
The following users thanked this post: jpb

Offline testpoint1

  • Regular Contributor
  • *
  • Posts: 194
  • Country: us
 

Offline Leo Bodnar

  • Frequent Contributor
  • **
  • Posts: 758
  • Country: gb
Quote
all the phase noise testing about OCXO is better than Rubidium in the web is a rumor
...
It's a well known governmental conspiracy.
Leo

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 18900
  • Country: nl
    • NCT Developments
Quote
all the phase noise testing about OCXO is better than Rubidium in the web is a rumor
...
It's a well known governmental conspiracy.
Leo
From own experience: I 've seen a Rubidium being worse in terms of phase noise compared to a good oscillator. Putting a clean-up oscillator behind a Rubidium clock is a good solution to get a very clean clock with reasonable long term stability.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline thinkfat

  • Supporter
  • ****
  • Posts: 919
  • Country: de
    • Matthias' Hackerstübchen
Quote
all the phase noise testing about OCXO is better than Rubidium in the web is a rumor
...
It's a well known governmental conspiracy.
Leo
From own experience: I 've seen a Rubidium being worse in terms of phase noise compared to a good oscillator. Putting a clean-up oscillator behind a Rubidium clock is a good solution to get a very clean clock with reasonable long term stability.

I don't quite see how a Rubidium-locked OCXO can possibly have less phase noise than the same, undisciplined OCXO. The servo loop will always introduce some disturbance.
 

Offline jpb

  • Super Contributor
  • ***
  • Posts: 1664
  • Country: gb
Quote
all the phase noise testing about OCXO is better than Rubidium in the web is a rumor
...
It's a well known governmental conspiracy.
Leo
From own experience: I 've seen a Rubidium being worse in terms of phase noise compared to a good oscillator. Putting a clean-up oscillator behind a Rubidium clock is a good solution to get a very clean clock with reasonable long term stability.

I don't quite see how a Rubidium-locked OCXO can possibly have less phase noise than the same, undisciplined OCXO. The servo loop will always introduce some disturbance.
The point, I think, is that the OCXO has less phase noise than the rubidium while tracking its long term accuracy, the OCXO will not improve its own phase noise - just the long term stability or drift. Quartzlock sell a commercial version of this as do other companies I think.
 

Offline Leo Bodnar

  • Frequent Contributor
  • **
  • Posts: 758
  • Country: gb
It's a well known governmental conspiracy.
Leo
From own experience: I 've seen a Rubidium being worse in terms of phase noise compared to a good oscillator. Putting a clean-up oscillator behind a Rubidium clock is a good solution to get a very clean clock with reasonable long term stability.
It was just irony. As per tradition.
Leo

Offline testpoint1

  • Regular Contributor
  • *
  • Posts: 194
  • Country: us
Quote
all the phase noise testing about OCXO is better than Rubidium in the web is a rumor
...
It's a well known governmental conspiracy.
Leo
From own experience: I 've seen a Rubidium being worse in terms of phase noise compared to a good oscillator. Putting a clean-up oscillator behind a Rubidium clock is a good solution to get a very clean clock with reasonable long term stability.

I don't quite see how a Rubidium-locked OCXO can possibly have less phase noise than the same, undisciplined OCXO. The servo loop will always introduce some disturbance.
The point, I think, is that the OCXO has less phase noise than the rubidium while tracking its long term accuracy, the OCXO will not improve its own phase noise - just the long term stability or drift. Quartzlock sell a commercial version of this as do other companies I think.

please tear down several OCXO and  Rubidium Oscillator (such as spectratime SRO), to compare its circult and infrastructure, then test the phase noise (do not forget use a CS reference),  paste here.
 

Offline Leo Bodnar

  • Frequent Contributor
  • **
  • Posts: 758
  • Country: gb
please tear down several OCXO and  Rubidium Oscillator (such as spectratime SRO), to compare its circult and infrastructure, then test the phase noise (do not forget use a CS reference),  paste here.
Are you confusing phase noise with timing stability?
Why on earth would you use Cs source as a phase noise reference?  It has just average phase noise performance.
Compare current top of the line Cs reference with modern high quality OCXO
Leo
« Last Edit: January 29, 2020, 11:16:39 am by Leo Bodnar »
 
The following users thanked this post: thinkfat

Offline testpoint1

  • Regular Contributor
  • *
  • Posts: 194
  • Country: us
please tear down several OCXO and  Rubidium Oscillator (such as spectratime SRO), to compare its circult and infrastructure, then test the phase noise (do not forget use a CS reference),  paste here.
Are you confusing phase noise with timing stability?
Why on earth would you use Cs source as a phase noise reference?  It has just average phase noise performance.
Compare current top of the line Cs reference with modern high quality OCXO
Leo

if no frequency accuracy, hard to test the phase noise of a cesium standard, I ever check a E5505A test system to test the 5071A, they use another 5071A as frequency reference, any way, I just want to remind you, any atomic clock, is is not output from the Rubidium or Cesium cell, it is also from the crystal (temperature controlled).
 

Offline testpoint1

  • Regular Contributor
  • *
  • Posts: 194
  • Country: us
I would like to say here, if the phase noise testing using DUT as the frequency reference, I have no word; if test for 10MHz (usually is), 1Hz offest (dBc level), are you sure the 10MHz DUT is 10.000000MHz ? if the DUT frequency is not accurate (eg 10.000000 is 10.000002 MHz), any mean for test the "phase noise" of 1Hz  (10,000,001Hz) ?
I try to use a hign dynamic spectrum analyzer to test the parameter.
we used the atomic oscillator the demodulation wireless signal, we just care the frequency stability, sorry, never test the phase noise, if someone can test it, just paste here your test result and test diagram.
 

Offline thinkfat

  • Supporter
  • ****
  • Posts: 919
  • Country: de
    • Matthias' Hackerstübchen
Not so easy to measure phase noise at these levels. You'll mostly measure your test instruments synthesizer. You need a rather complex setup to measure much below -100dBc/Hz

"TheSignalPath" on YT had a pretty interesting tutorial about this just recently.



PS: The absolute frequency offset between the DUT and your instrument doesn't count, since the phase noise is always measured referenced to the signal.
« Last Edit: January 29, 2020, 04:14:01 pm by thinkfat »
 

Offline Leo Bodnar

  • Frequent Contributor
  • **
  • Posts: 758
  • Country: gb
Only if you are using the wrong tool - like spectrum analyser.
Even 50 years old cheap and cheerful phase noise test kits will happily go to -170.. -180dBc/Hz if you feed them a good [=Wenzel or suchlike] clean reference.
Leo
P.S. Assuming 10MHz carrier ofc, not 1GHz
« Last Edit: January 29, 2020, 06:34:42 pm by Leo Bodnar »
 

Offline testpoint1

  • Regular Contributor
  • *
  • Posts: 194
  • Country: us
Not so easy to measure phase noise at these levels. You'll mostly measure your test instruments synthesizer. You need a rather complex setup to measure much below -100dBc/Hz

"TheSignalPath" on YT had a pretty interesting tutorial about this just recently.



PS: The absolute frequency offset between the DUT and your instrument doesn't count, since the phase noise is always measured referenced to the signal.

yes, BB60C can not test it, I try to use a SM200B to test it (I need to buy it, 13K USD), but, as a complete product (frequency standard reference), seems no mean to me to test it, since it is NOT a element (oscillator). I just care its frequency jitter wander, my focal point is 5G demoduation, not the universal electronics.
« Last Edit: January 29, 2020, 08:03:43 pm by testpoint1 »
 

Offline thinkfat

  • Supporter
  • ****
  • Posts: 919
  • Country: de
    • Matthias' Hackerstübchen
Only if you are using the wrong tool - like spectrum analyser.
Even 50 years old cheap and cheerful phase noise test kits will happily go to -170.. -180dBc/Hz if you feed them a good [=Wenzel or suchlike] clean reference.
Leo
P.S. Assuming 10MHz carrier ofc, not 1GHz

Well yes, 1GHz seems to be Shahriars' idea of "low frequency". But the video still demonstrates the main point: if you haven't got a signal source that is way better than what you want to measure, you need to pull some pretty clever stunts. A Cs beam standard might be OK compared to a LPRO-class Rb standard, but that one is just -95dBc/Hz @10Hz. But a good OCXO should be well below that.

Random find on the net: -120dBc/Hz @10Hz
https://www.petermann-technik.de/fileadmin/documents/datasheets/Quartz-Crystal-Oszillators/ocxo/THT_LOW_POWER_OCXO-20.pdf

I guess a spectrum analyzer is what an RF lab typically has available. Phase noise test set - pretty specialized type of gear, no?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf