Author Topic: Economical option for precision frequency reference?  (Read 155294 times)

0 Members and 1 Guest are viewing this topic.

Offline Bryan

  • Frequent Contributor
  • **
  • Posts: 618
  • Country: ca
Re: Economical option for precision frequency reference?
« Reply #325 on: December 07, 2015, 11:49:52 am »
Getting a bit frustrated, I'm trying to make Lady Heather talk to my NTBW50AA.  I bought a USB-RS232 cable and it installs on my Windows 10 system as COM4.  I'm pretty sure the cable connections are good but all that happens when I start LH is it says 'NO COM1 SERIAL PORT DATA SEEN' - I'm now trying Windows 8 compatibility mode and am running LH as Admin.

I added a heather.cfg file with nothing but '/4' in it but I still get the message about COM1 when LH starts up, I also tried typing in '/4' at the command interface.  I messed with various setting for the port - current settings are 9600, 8 bit, no parity, 1 stop bit

Any suggestions anyone?

[EDIT1] It seems it's a Lady Heather problem.  I have found the serial port setting and it's 19,200,7,Odd,1, and I think XOn/XOff, with these settings I can connect to Port 4 and then when I type '1' it responds with "E-113>

Have my Nortel working fine with LH on Windows 10 on Com 1. You may want to try a download of tboltmon.exe and try a hard or cold reset. I assume your ports in device manager are set properly. Mine are 9600,8,none,1. no flow control.
-=Bryan=-
 

Offline Gandalf_Sr

  • Super Contributor
  • ***
  • Posts: 1729
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #326 on: December 28, 2015, 07:40:43 pm »
Getting a bit frustrated, I'm trying to make Lady Heather talk to my NTBW50AA.  I bought a USB-RS232 cable and it installs on my Windows 10 system as COM4.  I'm pretty sure the cable connections are good but all that happens when I start LH is it says 'NO COM1 SERIAL PORT DATA SEEN' - I'm now trying Windows 8 compatibility mode and am running LH as Admin.

I added a heather.cfg file with nothing but '/4' in it but I still get the message about COM1 when LH starts up, I also tried typing in '/4' at the command interface.  I messed with various setting for the port - current settings are 9600, 8 bit, no parity, 1 stop bit

Any suggestions anyone?

[EDIT1] It seems it's a Lady Heather problem.  I have found the serial port setting and it's 19,200,7,Odd,1, and I think XOn/XOff, with these settings I can connect to Port 4 and then when I type '1' it responds with "E-113>

Have my Nortel working fine with LH on Windows 10 on Com 1. You may want to try a download of tboltmon.exe and try a hard or cold reset. I assume your ports in device manager are set properly. Mine are 9600,8,none,1. no flow control.
Bryan, thanks for the input.  I did get Lady Heather working on Windows 10, the issue is that the OS protects itself by not allowing unknown software to read from the actual location it's installed in so the instructions to place the Heather.cfg file in the same folder as the .exe file are wrong but, in fact, the '?' function tells you (all the way at the end), that the heather.cfg file needs to be in c:\user\<yourname>\OneDrive\Documents and when the file is placed there, it works. There are some issues if you try to change screen resolution but that's OK.
« Last Edit: December 28, 2015, 09:29:58 pm by Gandalf_Sr »
If at first you don't succeed, get a bigger hammer
 

Offline N8AUM

  • Regular Contributor
  • *
  • Posts: 132
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #327 on: December 28, 2015, 09:50:00 pm »
I also had problems with Lady Heather. Tried different com port settings with her AND windblows 7. Ended up installing an old serial I/O board in one of the slots so that it would be com1.

 
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #328 on: December 31, 2015, 12:49:43 am »
Due to a GPSDO sickness and too much time spent trolling ebay I bought another GPSDO. This is an HP 58540A - likely built to compete with the Trimble Thunderbolt.
It has an SCPI interface, it doesn't provide a ton of functions and is pretty basic overall. It uses a single 24 VDC supply and has a switching supply under the timing board which provides the needed voltages. I need to run it for a month or two before really seeing how it compares with my Trimble units.

« Last Edit: April 17, 2016, 07:59:59 am by TheSteve »
VE7FM
 

Online Vgkid

  • Super Contributor
  • ***
  • Posts: 2710
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #329 on: December 31, 2015, 12:55:24 am »
I wonder what the trimpot does?
If you own any North Hills Electronics gear, message me. L&N Fan
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #330 on: December 31, 2015, 01:05:45 am »
I wonder what the trimpot does?
Probably something boring like setting the 10 MHz output level.
« Last Edit: December 31, 2015, 01:07:37 am by TheSteve »
VE7FM
 

Online Vgkid

  • Super Contributor
  • ***
  • Posts: 2710
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #331 on: December 31, 2015, 01:57:26 am »
That makes sense, I believe that they do that in the lucent boxes.
If you own any North Hills Electronics gear, message me. L&N Fan
 

Offline N8AUM

  • Regular Contributor
  • *
  • Posts: 132
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #332 on: December 31, 2015, 02:02:25 am »
cool serial # the osc. !
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #333 on: December 31, 2015, 02:19:09 am »
cool serial # the osc. !

Yeah, not sure how many of these they made. Google doesn't give too much info but the others I have seen use a different OCXO. The PCB layout fits several different OCXO models. It would be nice to upgrade it to something newer with lower phase noise but the odds of finding one that is drop in compatible with the hardware and firmware seems pretty slim.
VE7FM
 

Offline dannyf

  • Super Contributor
  • ***
  • Posts: 8221
  • Country: 00
Re: Economical option for precision frequency reference?
« Reply #334 on: December 31, 2015, 03:10:40 pm »
Quote
But what I'd really like is if someone can give me some practical suggestions on what is the most practical way to set up a good frequency reference - what is needed, and what's the best place to source one.

A few approaches, depending on what you want to do:

1) (VC)TCXO: usually good for 2.5ppm max - and typically < 0.5ppm at room temperature. Cheap and lots of frequency options;
2) GPS: slightly more expensive, but many are programmable so you can generate whatever frequencies you so desire.

I used a hybrid approach: I built a platform that adjusts a VCTCXO's output to a GPS. Once caliberated, that VCTCXO can operate without the GPS. Done on a PIC16F684 but should be easily adopted onto other platforms, as it is essentially a phase comparator.
================================
https://dannyelectronics.wordpress.com/
 

Offline Bryan

  • Frequent Contributor
  • **
  • Posts: 618
  • Country: ca
Re: Economical option for precision frequency reference?
« Reply #335 on: January 01, 2016, 09:40:53 pm »

Have my Nortel working fine with LH on Windows 10 on Com 1. You may want to try a download of tboltmon.exe and try a hard or cold reset. I assume your ports in device manager are set properly. Mine are 9600,8,none,1. no flow control.
Bryan, thanks for the input.  I did get Lady Heather working on Windows 10, the issue is that the OS protects itself by not allowing unknown software to read from the actual location it's installed in so the instructions to place the Heather.cfg file in the same folder as the .exe file are wrong but, in fact, the '?' function tells you (all the way at the end), that the heather.cfg file needs to be in c:\user\<yourname>\OneDrive\Documents and when the file is placed there, it works. There are some issues if you try to change screen resolution but that's OK.

Yes, I noticed the same thing when trying to adjust the screen resolution , it will crash. Too bad it has not been updated since 2013, I think it is a great program. Would be great if it could be upgraded to incorporate some of the other GPSDO boxes out there. 

Another tip if I could share is to do a Autotune of the oscillator after it has been running for a number of months "&" then "A"
« Last Edit: January 02, 2016, 10:17:22 pm by Bryan »
-=Bryan=-
 

Offline _Andrew_

  • Regular Contributor
  • *
  • Posts: 85
  • Country: gb
Re: Economical option for precision frequency reference?
« Reply #336 on: January 02, 2016, 03:25:37 am »
Stumbled across these OCXO & GPS TCXO reference oscillators from Total Frequency Control

http://www.tfc.co.uk/products.php?CategoryID=3&gclid=CKmYp_-TisoCFRUTGwodmu4Bbw

http://www.tfc.co.uk/products.php?CategoryID=6&SubID=3



 
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #337 on: January 04, 2016, 05:15:21 am »
Thought I would compare the 1 PPS outputs from the three GPSDO's I have here.

The blue trace is the Nortel NTGS50AA which is the trigger channel
The yellow trace is the cheapie Trimble 57964-15 with 63090 OCXO
The purple trace is the HP 58540A

The scope has been running on infinite persistence for a few hours.
You can see the jitter on the HP is probably double what it is on either of the others(I have triggered on the Trimble with similar results).
I have changed the antenna delay to get them to be reasonably close on the scope so it is easier to see. By default the Nortel leads the Trimble by roughly 500ns which leads the HP by 100ns. So even with 3 GPSDO's I have no idea when the pulse really occurs :)
VE7FM
 

Online edpalmer42

  • Super Contributor
  • ***
  • Posts: 2260
  • Country: ca
Re: Economical option for precision frequency reference?
« Reply #338 on: January 04, 2016, 06:54:33 am »
Now do the same test with 3 of your Rb oscillators.  You'll be shocked by the difference!  :o  Over a period of a few hours, the Rb oscillators are more stable than the GPSDOs.  It's only over longer time frames that the long-term frequency stability of the GPSDO starts to show its advantage over the Rb standard.

Of course, we have to remember that these GPSDOs are not optomized to give the lowest possible jitter.  You could make the time constants longer, but there are tradeoffs there too.  The same applies to the Rb units.  Low jitter just wasn't one of the design requirements.

Ed
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #339 on: January 04, 2016, 07:22:41 am »
Now do the same test with 3 of your Rb oscillators.  You'll be shocked by the difference!  :o  Over a period of a few hours, the Rb oscillators are more stable than the GPSDOs.  It's only over longer time frames that the long-term frequency stability of the GPSDO starts to show its advantage over the Rb standard.

Of course, we have to remember that these GPSDOs are not optomized to give the lowest possible jitter.  You could make the time constants longer, but there are tradeoffs there too.  The same applies to the Rb units.  Low jitter just wasn't one of the design requirements.

Ed

The Rb's I have here don't have a 1 PPS output. I can watch the 10 MHz out but they will all wander a little as none are at exactly 10 MHz and it isn't possible to set them perfectly I don't think. The rate of wander is likely constant though which is what counts.
VE7FM
 

Offline Bryan

  • Frequent Contributor
  • **
  • Posts: 618
  • Country: ca
Re: Economical option for precision frequency reference?
« Reply #340 on: January 04, 2016, 09:55:07 am »
Now do the same test with 3 of your Rb oscillators.  You'll be shocked by the difference!  :o  Over a period of a few hours, the Rb oscillators are more stable than the GPSDOs.  It's only over longer time frames that the long-term frequency stability of the GPSDO starts to show its advantage over the Rb standard.

Of course, we have to remember that these GPSDOs are not optomized to give the lowest possible jitter.  You could make the time constants longer, but there are tradeoffs there too.  The same applies to the Rb units.  Low jitter just wasn't one of the design requirements.

Ed


Taking this into consideration what is the best way to "adjust" a RB to a GPSDO or is it really feasible to do so considering the GPSDO is better long term, but suffers from short term stability.
-=Bryan=-
 

Offline awallin

  • Frequent Contributor
  • **
  • Posts: 694
Re: Economical option for precision frequency reference?
« Reply #341 on: January 04, 2016, 01:24:34 pm »
Taking this into consideration what is the best way to "adjust" a RB to a GPSDO or is it really feasible to do so considering the GPSDO is better long term, but suffers from short term stability.

PI-controller that looks at the difference between the GPSDO and Rb phase (either at 10MHz or 1PPS) and steers the Rb with a DAC. Time constant for a bad Rb-clock maybe hours, for a good one 1-2 days.
http://www.thinksrs.com/assets/instr/PRS10/PRS10diag2LG.gif

The measurement of GPSDO vs Rb needs to be quite good. 9-digit/s frequency counter at least.
You'd want the GPSDO vs Rb measurement noise to reach an ADEV of 1e-12 (the limit for a typical Rb clock) a bit before the PI-loop time-constant, maybe at 1000s, so a 1PPS measurement good to 1ns that averages down as about 3e-9/tau should work.
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #342 on: January 04, 2016, 04:56:49 pm »
Now do the same test with 3 of your Rb oscillators.  You'll be shocked by the difference!  :o  Over a period of a few hours, the Rb oscillators are more stable than the GPSDOs.  It's only over longer time frames that the long-term frequency stability of the GPSDO starts to show its advantage over the Rb standard.

Of course, we have to remember that these GPSDOs are not optomized to give the lowest possible jitter.  You could make the time constants longer, but there are tradeoffs there too.  The same applies to the Rb units.  Low jitter just wasn't one of the design requirements.

Ed


Taking this into consideration what is the best way to "adjust" a RB to a GPSDO or is it really feasible to do so considering the GPSDO is better long term, but suffers from short term stability.

In the Temex LPFRS user manual it suggests using a GPSDO as a reference for an Agilent 53131A frequency counter. Being they have an analog voltage adjust this might be as stable as you can reasonably expect.
VE7FM
 

Online edpalmer42

  • Super Contributor
  • ***
  • Posts: 2260
  • Country: ca
Re: Economical option for precision frequency reference?
« Reply #343 on: January 04, 2016, 10:04:56 pm »
Now do the same test with 3 of your Rb oscillators.  You'll be shocked by the difference!  :o  Over a period of a few hours, the Rb oscillators are more stable than the GPSDOs.  It's only over longer time frames that the long-term frequency stability of the GPSDO starts to show its advantage over the Rb standard.

Of course, we have to remember that these GPSDOs are not optomized to give the lowest possible jitter.  You could make the time constants longer, but there are tradeoffs there too.  The same applies to the Rb units.  Low jitter just wasn't one of the design requirements.

Ed


Taking this into consideration what is the best way to "adjust" a RB to a GPSDO or is it really feasible to do so considering the GPSDO is better long term, but suffers from short term stability.

To do a one-time adjustment, the best way that I know is to use a time interval counter or DSO to measure the time from the rising edge of the GPSDO to the rising edge of the Rb.  It doesn't matter if it's 1 PPS or 10 MHz, but a square wave signal is best.  Note the time interval and then watch it over many minutes.  If the time interval increases, the Rb is falling behind the GPSDO, so increase its frequency a bit.  If the time interval decreases, the Rb is overtaking the GPSDO so slow it down.  If you log the time interval and the length of time between measurements, you can calculate the frequency difference.  The resolution of your measurement will be determined by the resolution of your time interval counter and the time between measurements.  A one nanosecond counter and measurements one hour apart gives you a resolution of < 3e-13.  The long interval between measurements helps average out the GPSDO's jitter.  This type of measurement can be made manually or you can collect the data electronically.

If you want to discipline the Rb to the GPS, that's a totally different situation.  I've never really understood the advantages to that (other than the technical challenge).  The only situation where I can see an advantage is if you have really poor GPS reception and you want your standard to maintain service through the outages.  Other than that, you can build an OCXO-equipped GPSDO that provides short-term performance that's limited only by the OCXO itself and long-term performance that's limited by GPS itself.  Both are better than the Rb's performance.

Ed
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Economical option for precision frequency reference?
« Reply #344 on: January 04, 2016, 10:23:27 pm »
I did the quick tuning using my frequency counter and then for fine tuning I compared the GPSDO output to the Rb output on the scope using infinite persistence. Once they are very close it will take a few minutes to see it start drifting either direction. Obviously you can't do this really fast or jitter in the GPSDO itself would mess you up. No matter how much adjustment you do it will always eventually drift with respect to the GPSDO. Ideally I'd like it so that a full cycle of drift took an hour +. Not sure it would last day to day though, in my Temex Rb's you can adjust frequency electronically but also via an analog voltage set by a pot - which will drift some on its own.
VE7FM
 

Online Vgkid

  • Super Contributor
  • ***
  • Posts: 2710
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #345 on: January 04, 2016, 10:38:22 pm »
I bought a gpsdo in addition to the symmetricom. From what i can see: Furuno receiver + 8663-xs receiver. Hopefully it artives intact. Update in a few months.
If you own any North Hills Electronics gear, message me. L&N Fan
 

Offline Bryan

  • Frequent Contributor
  • **
  • Posts: 618
  • Country: ca
Re: Economical option for precision frequency reference?
« Reply #346 on: January 05, 2016, 09:47:41 am »
I did the quick tuning using my frequency counter and then for fine tuning I compared the GPSDO output to the Rb output on the scope using infinite persistence. Once they are very close it will take a few minutes to see it start drifting either direction. Obviously you can't do this really fast or jitter in the GPSDO itself would mess you up. No matter how much adjustment you do it will always eventually drift with respect to the GPSDO. Ideally I'd like it so that a full cycle of drift took an hour +. Not sure it would last day to day though, in my Temex Rb's you can adjust frequency electronically but also via an analog voltage set by a pot - which will drift some on its own.

That is pretty much the procedure that I have done, but with some OCXO's. For the home lab I don't think one can do much better, nor need to. What I do is watch the output of the GPSDO on Lady heather and when I see it at the lowest oscillator fluctuations is when I will do an adjustment. We are talking >+/- 50PPT fluctuations in Lady Heather so think that is about as good as I will get it. If I had a better view of the sky I might be able to get better ADEV readings.



« Last Edit: January 05, 2016, 09:54:11 am by Bryan »
-=Bryan=-
 

Offline FlyingHacker

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: us
  • You're Doing it Wrong
Re: Economical option for precision frequency reference?
« Reply #347 on: March 01, 2016, 07:08:31 pm »
How are you using these GPSDOs? Do you leave them up all the time as your lab reference?

Could you simply turn it on, let it warm up and lock, and then calibrate a regular OCXO to it? Then use that regular OCXO as your lab reference frequency?

How accurate a 10MHz do you get with one of these if you only turn it on for a few hours?

My issue is that an OCXO is plenty accurate for a few months, but I need some way to calibrate it periodically. I doubt I can get a GPS signal on a regular basis in my basement hobby lab.

Thoughts?
--73
 

Offline Macbeth

  • Super Contributor
  • ***
  • Posts: 2571
  • Country: gb
Re: Economical option for precision frequency reference?
« Reply #348 on: March 01, 2016, 09:19:16 pm »
How are you using these GPSDOs? Do you leave them up all the time as your lab reference?

Could you simply turn it on, let it warm up and lock, and then calibrate a regular OCXO to it? Then use that regular OCXO as your lab reference frequency?

How accurate a 10MHz do you get with one of these if you only turn it on for a few hours?

My issue is that an OCXO is plenty accurate for a few months, but I need some way to calibrate it periodically. I doubt I can get a GPS signal on a regular basis in my basement hobby lab.

Thoughts?
The GPSDO has an OCXO onboard (that it disciplines, hence "Lady Heather" the dominatrix  :-DD ). I don't have any other reference because I am not a time-nut (yet).

Yes, if you turn it off then it will need to re-establish itself, but the 10MHz OCXO will still be available with a good degree of confidence for quite a while during GPS blackouts. That Figure of Merit is available via the serial terminal.

Also, your basement lab shouldn't be a problem as an external active "GPS mouse" is all you need - I have mine attached to a drain pipe outside my window. If you need more than a couple of meters of cable then there are trim factors you can specify to the GPSDO to let it correct for that.
 

Offline FlyingHacker

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: us
  • You're Doing it Wrong
Re: Economical option for precision frequency reference?
« Reply #349 on: March 01, 2016, 10:19:27 pm »
Thanks for the reply.

So whenever you turn off the GPSDO do you have to re-establish a GPS lock for it to be accurate again, or is the control voltage stored in some non-volatile medium, allowing it to boot up to the same accuracy as when it shut down?

I guess I am trying to figure out if you need a GPS signal every time you power cycle the unit.

Is this comparable:
http://www.ebay.com/itm/121530825744?_trksid=p2060353.m1438.l2649&ssPageName=STRK%3AMEBIDX%3AIT

I mainly want something to calibrate to.

Thanks.
--73
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf