Author Topic: Economical option for precision frequency reference?  (Read 117853 times)

0 Members and 1 Guest are viewing this topic.

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #175 on: August 03, 2015, 04:48:54 am »
You know, I just don't believe these TI numbers. Perhaps I don't understand what they represent. I hooked up the Lucent GPSDO 10MHz output to my 53131A, which has its own OCXO that has been calibrated to the Lucent.

Now if the Lucent was really drifting by 10 - 50ns per 1 second interval, unless the timer OCXO was drifting in sync (impossible), I would expect to see variation in the 0.1Hz digit. Instead, it is rock solid all the way out to the 0.0001 digit.

I guess maybe I need to sign up for the time-nuts mailing list and ask someone to explain it (or maybe Ed can explain here).

Believe your counter.  The GPSDO is reporting the TI between its PPS and the calculated PPS based on the received signals.  Problem is, the received signal gets distorted by the ionosphere and maybe by multipath.  This creates jitter in the calculated result.  Over a long period, this jitter averages out to zero.  The OCXO's function is to act as a flywheel to filter out the short-term jitter.  The  PPS is then generated from the OCXO.  The 'gotcha' here is that in the RFTG, the averaging appears to degrade the OCXO's performance - or doesn't completely filter out the short-term jitter.

There are a few things that can be checked to ensure that the jitter is minimized.  Like they say, "Location, location, location".  If the GPSDO doesn't have its location figured correctly, this can be a major cause of jitter.  I don't remember if the RFTG has this feature, but some GPSDOs give you the option to set the length of time for the location survey.  Longer is better!  The survey is just as vulnerable to jitter as the PPS.  A longer survey helps average out the jitter.  Give it a day or two - compare the before and after results. 

Note that measuring the output with a frequency counter isn't a very precise measurement because the counter's gate time results in an averaging function that hides things.  It's much more effective to measure the time interval between the GPSDO and another local reference.  Of course, you have to have another reference!  You also need to be able to capture the readings and log them to a PC.

Ed

Thanks, that is a very helpful explanation.

One interesting thing just happened. I was sitting here watching one of robrenz' videos, and all of a sudden the HUP value reset to its starting point (4.32E-4), and the EFC started going nuts. Checking the logs, it looks like it rebooted. After the reboot, it started another survey. I've turned off the survey on boot feature for now. I'm a bit worried as to why it rebooted. It's plugged into a UPS, and I wasn't moving it or doing anything that should have disturbed the (albeit iffy) power connection.

 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #176 on: August 03, 2015, 06:10:23 am »
I discovered why it rebooted. The UPS is faulty. It's a small Eaton that I bought used, so no surprises there.
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #177 on: August 04, 2015, 04:23:22 am »
So I think that UPS had been acting up for some time, perhaps very brief cutouts that were enough to disrupt the power supply to the GPSDO. Since removing that from the circuit, the EFC corrections have really settled down, and HUD was as low as 1.1us today  - currently at 1.7us.  The EFC statistics in Z38XX say the EFC slope is 5.025E+01 per day resembling a relative OCXO drift of 5.024E-11 (Hz) per day.
 

Offline TSL

  • Regular Contributor
  • *
  • Posts: 240
  • Country: au
Re: Economical option for precision frequency reference?
« Reply #178 on: August 04, 2015, 04:49:58 am »

What is actually on the bg7tbl adapter board - just connectivity to power and the SMA connectors, or is there actually some logic there (i.e. a microcontroller). What I am mainly interested in is if these units are subject to the frequency problem (not exactly 10MHz) that the other BG7TBL design is.

I did a tear down and test of the BG7TBL unit here..

https://drive.google.com/a/skybase.net/file/d/0B9Oysj7clpT7YTBWSjlfNTI1aTQ/view

While I don't have the resolution of KE5FX's test rig ( he has a Cesium standard!) my Thunderbolt based test setup did not show that error.

regards

Tim
VK2XAX :: QF56if :: BMARC :: WIA :: AMSATVK
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #179 on: August 04, 2015, 05:21:15 am »

What is actually on the bg7tbl adapter board - just connectivity to power and the SMA connectors, or is there actually some logic there (i.e. a microcontroller). What I am mainly interested in is if these units are subject to the frequency problem (not exactly 10MHz) that the other BG7TBL design is.

I did a tear down and test of the BG7TBL unit here..

https://drive.google.com/a/skybase.net/file/d/0B9Oysj7clpT7YTBWSjlfNTI1aTQ/view

While I don't have the resolution of KE5FX's test rig ( he has a Cesium standard!) my Thunderbolt based test setup did not show that error.

regards

Tim

Very nice. From which eBay seller did you purchase this unit?
 

Offline TSL

  • Regular Contributor
  • *
  • Posts: 240
  • Country: au
Re: Economical option for precision frequency reference?
« Reply #180 on: August 04, 2015, 05:38:18 am »

Very nice. From which eBay seller did you purchase this unit?

Thanks - I bought from fly-xy, he has a bunch of interesting RF stuff. URL for the GPSDO I bought here...

http://www.ebay.com/itm/10MHZ-OUTPUT-SINE-WAVE-GPS-DISCiPLINED-CLOCK-GPSDO-GPS-Antenna-Power-supply/121530825744


And here is a newer model with Huawei internals ...

http://www.ebay.com/itm/HW-HUAWEI-GPS-DISCIPLINED-CLOCK-GPSDO-OSCILLATOR-Antenna-power-supply-/111716558224

and yet another model with Trimble internals...

http://www.ebay.com/itm/Trimble-GPS-Receiver-GPSDO-10MHz-1PPS-GPS-Disciplined-Clock-Antenna-power/131564496688

regards

Tim
VK2XAX :: QF56if :: BMARC :: WIA :: AMSATVK
 


Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #182 on: August 04, 2015, 06:07:48 am »
« Last Edit: August 04, 2015, 06:09:24 am by motocoder »
 

Offline TSL

  • Regular Contributor
  • *
  • Posts: 240
  • Country: au
VK2XAX :: QF56if :: BMARC :: WIA :: AMSATVK
 

Offline usagi

  • Frequent Contributor
  • **
  • Posts: 366
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #184 on: August 04, 2015, 06:30:51 am »
i have a bg7tbl huawei (oscilloquartz star 4), and i'm waiting for delivery of a bg7tbl trimble.

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #185 on: August 04, 2015, 06:23:47 pm »
i have a bg7tbl huawei (oscilloquartz star 4), and i'm waiting for delivery of a bg7tbl trimble.

Excellent. Looking forward to some comparisons.
 

Offline EV

  • Frequent Contributor
  • **
  • Posts: 517
  • Country: fi
Re: Economical option for precision frequency reference?
« Reply #186 on: August 05, 2015, 11:28:08 am »
I'm assuming most of these posts are referring to the

LUCENT/SYMMETRICOM/HP Z3810AS, KS24361 L101 REF0 / L102 REF1, GPSDO TIMING SYSTEM

This REF0 unit has no GPS antenna connector. However it has "NO GPS" led. Can it be used as GPS reference separately without REF1 unit? Are there any instructions or manuals for these?
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #187 on: August 05, 2015, 03:00:49 pm »
I'm assuming most of these posts are referring to the

LUCENT/SYMMETRICOM/HP Z3810AS, KS24361 L101 REF0 / L102 REF1, GPSDO TIMING SYSTEM

This REF0 unit has no GPS antenna connector. However it has "NO GPS" led. Can it be used as GPS reference separately without REF1 unit? Are there any instructions or manuals for these?

There was some discussion about this on the time-nuts mailing list. There was speculation that an external GPS could be hooked up and providing only the 1PPS signal to REF0, to allow a 21st century GPS chip to work with it and avoid the crappy insensitive chip that's in there now. However, I read every single message on time-nuts about this unit, and no one ever reported back as having even tried this.

AFAIK, there is no manual publicly available for these units. They were created as a drop-in replacement for an earlier unit, the KS-24019, and it is similar to a later unit, the Z3805A, for which I believe there is some documentation.
 

Offline usagi

  • Frequent Contributor
  • **
  • Posts: 366
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #188 on: August 05, 2015, 10:10:27 pm »
There was some discussion about this on the time-nuts mailing list. There was speculation that an external GPS could be hooked up and providing only the 1PPS signal to REF0, to allow a 21st century GPS chip to work with it and avoid the crappy insensitive chip that's in there now. However, I read every single message on time-nuts about this unit, and no one ever reported back as having even tried this.

why bother, just get a bg7tbl oscilloquartz or bg7tbl trimble and be done with it.

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #189 on: August 05, 2015, 11:35:14 pm »
There was some discussion about this on the time-nuts mailing list. There was speculation that an external GPS could be hooked up and providing only the 1PPS signal to REF0, to allow a 21st century GPS chip to work with it and avoid the crappy insensitive chip that's in there now. However, I read every single message on time-nuts about this unit, and no one ever reported back as having even tried this.

why bother, just get a bg7tbl oscilloquartz or bg7tbl trimble and be done with it.

Yes, I agree. Just answering his question :)
 

Offline dadler

  • Supporter
  • ****
  • Posts: 848
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #190 on: August 06, 2015, 12:51:14 am »
So, I have some information about the RFTG units.

I've shared my personal directory here:

https://www.dropbox.com/sh/osgb29wvgnxag4v/AABzYRcHALPExnaa6hU7MLeua?dl=0

There is some software and an RFTG user manual from 1996. You guys might already have this stuff, but just thought I'd share just in case.
 

Offline Vgkid

  • Super Contributor
  • ***
  • Posts: 2600
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #191 on: August 06, 2015, 03:22:38 am »
I have a rftgm-II at home if anyone wants a teardown.
If you own any North Hills Electronics gear, message me. L&N Fan
 

Offline EV

  • Frequent Contributor
  • **
  • Posts: 517
  • Country: fi
Re: Economical option for precision frequency reference?
« Reply #192 on: August 06, 2015, 05:39:33 am »
Thanks for the info to motocoder and dadler!
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #193 on: August 06, 2015, 05:49:27 am »
So, I have some information about the RFTG units.

I've shared my personal directory here:

https://www.dropbox.com/sh/osgb29wvgnxag4v/AABzYRcHALPExnaa6hU7MLeua?dl=0

There is some software and an RFTG user manual from 1996. You guys might already have this stuff, but just thought I'd share just in case.

So this is the manual for the older unit that the KS-24361 is patterned on, right?

For anyone thinking of downloading - the software does not run on recent versions of Windows.
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #194 on: August 06, 2015, 06:05:13 am »
So I posted in the other thread, but forgot to post here that I obtained one of the BG7TBL 2015-07-17 "Trimble" units. This is basically a board NOT designed by BG7TBL, grafted onto a board of his design that provides connectivity to front-panel LEDs, DB9 RS-232 connector, power and RF connectors. My unit is all sealed up in an aluminum box, although I am a bit concerned about it overheating and may pull it out (the box doesn't feel hot at all, so maybe OK). It came with a puck antenna and power supply. Notably the antenna is labelled 3.3V, but the "documentation" that came with this unit clearly says it uses 5V antennas.

If the antenna is placed outside (my "lab" is apparently a GPS dead zone - only the most sensitive of GPS chips will lock up indoors there), it locks up within just a few minutes. The box puts out a very nice and clean sine wave. I don't have the ability to measure phase noise on it, so can't comment about how clean it is. If I hook it up to my scope, and trigger on the signal from the Lucent, I can see a very consistent and repeatable drift in phase between the two references of about 27ns. I need a third unit of some type to see what is responsible for this phase drift.

I will have to fashion an adapter cable to hook up RS-232 and see what can be done with that. The software that came with the unit will not run on my machine; it just errors out looking for some ancient VB6 OCX component (MSCOMM32.OCX).

Pros
  • Reasonably priced
  • Nice 10MHz sine wave output
  • Quick GPS lock-up

Cons
  • No usable software
  • Effectively no documentation
 

Offline dadler

  • Supporter
  • ****
  • Posts: 848
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #195 on: August 06, 2015, 06:10:30 am »
So, I have some information about the RFTG units.

I've shared my personal directory here:

https://www.dropbox.com/sh/osgb29wvgnxag4v/AABzYRcHALPExnaa6hU7MLeua?dl=0

There is some software and an RFTG user manual from 1996. You guys might already have this stuff, but just thought I'd share just in case.

So this is the manual for the older unit that the KS-24361 is patterned on, right?

For anyone thinking of downloading - the software does not run on recent versions of Windows.

Yeah it's an old manual for an ancestor revision, but I was told that most of the communication protocol stuff and other configuration info should be relevant to the newer boxes.

And yeah, I never got the software to work myself but I didn't try that hard. The guy I got the Lucent box from gave it to me--and said he didn't know if it was compatible with the particular Lucent RFTG box or not. I have some more info in paper form, let me "scan" it and upload the PDF.
 

Offline dadler

  • Supporter
  • ****
  • Posts: 848
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #196 on: August 06, 2015, 06:43:27 am »
So I never really looked at the paperwork before, I guess it's not that useful but I scanned it anyways:

https://www.dropbox.com/s/n4nt3yhdt5mc10r/Lucent_RFTGm_II_XO_Notes.pdf?dl=0

Keep in mind that the unit I have is the RFTGm-II-XO, which is different than those with the open enclosures. It has a similar front plate, but the entire enclosure is a big, sealed heat sink. However I believe these are all similar architecture devices.



 

Offline EV

  • Frequent Contributor
  • **
  • Posts: 517
  • Country: fi
Re: Economical option for precision frequency reference?
« Reply #197 on: August 06, 2015, 07:07:07 am »
Cons
  • No usable software
  • ...

Have you tried tboltmon.exe? You can get it here under references:

http://www.prc68.com/I/ThunderBolt.shtml#Tboltmon
 

Offline TSL

  • Regular Contributor
  • *
  • Posts: 240
  • Country: au
Re: Economical option for precision frequency reference?
« Reply #198 on: August 06, 2015, 07:46:27 am »

Have you tried tboltmon.exe? You can get it here under references:

http://www.prc68.com/I/ThunderBolt.shtml#Tboltmon

That's actually a good idea. Trimble have their own language for their GPSDO's called TSIP. It's the only language Thunderbolts talk. Other Trimble's can be made to talk NMEA but probably only after you've spoken to it in TSIP!

regards

Tim
VK2XAX :: QF56if :: BMARC :: WIA :: AMSATVK
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #199 on: August 06, 2015, 08:47:49 pm »
dadler - thanks for scanning and uploading the docs on the RTFG.

EV and TSL - I will give tboltmon a try. I need to fashion an adapter cable before I can connect it up. Fortunately, the pinout is printed right on the front panel. I have to look up the pinout of the standard PC serial port - it may just need a gender changer (aka "null modem").

(Edit) Looks like a standard USB to serial adapter will work. I'll have to order one from Amazon, though, as all the ones here are USB to Serial/Null-Modem (i.e. a female DB9 connector).
« Last Edit: August 06, 2015, 09:07:52 pm by motocoder »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf