Author Topic: Economical option for precision frequency reference?  (Read 117849 times)

0 Members and 1 Guest are viewing this topic.

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Economical option for precision frequency reference?
« on: July 11, 2015, 06:18:22 pm »
I picked  up an HP 53131A a while back. After fixing a few small issues, the thing seems to be working great. However, this one came with the low-end oscillator. I've ordered an OCXO for it (off eBay), and when that comes in, I'd like to calibrate it.

I've read through a number of threads on this topic, and I see people have Rubidium frequency references, and GPS-disciplined references. On the GPS option, I see there are some surplus units that used to be inexpensive, but now the price is creeping up. Also, I am not sure what exactly is needed there.
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.

Thanks!
 

Offline mazurov

  • Frequent Contributor
  • **
  • Posts: 448
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #1 on: July 11, 2015, 06:54:33 pm »
If you need one-time setting of coarse adjustment of the unit you can cobble up a manual gpsdo from a GPS module with 1pps output like the one Adafruit carries plus PIC32 micro clocked from the unit being calibrated. You can PLL input clock of the micro to 50 MHz and then count between GPS 1PPS pulses. Using this technique I was able to get a HP 10544A within less tan 10 Hz on couple occasions.
« Last Edit: July 11, 2015, 06:57:45 pm by mazurov »
 

Offline Lightages

  • Supporter
  • ****
  • Posts: 4302
  • Country: ca
  • Canadian po
Re: Economical option for precision frequency reference?
« Reply #2 on: July 11, 2015, 07:02:50 pm »
Economical is a very relative term. Something like this is ready to go:
http://www.ebay.com/itm/281701995263
I have no experience with this unit nor the seller but it looks good.

Another option is this:
http://www.ebay.com/itm/FEI-fe-5680b-rubidium-oscillator-With-1pps-20mhz-output-ONLY-10mhz-NEED-to-MOD-/291419889143?pt=LH_DefaultDomain_0&hash=item43d9fa9df7
I purchased one from this seller and have modified it to output 10MHz. It was only a handful of cheap parts to do so. The information is mentioned in his auction. If you need cheaper than $65, then the options get rather few.
 

Online edpalmer42

  • Super Contributor
  • ***
  • Posts: 1820
  • Country: ca
Re: Economical option for precision frequency reference?
« Reply #3 on: July 11, 2015, 07:24:53 pm »
The answer really depends on the precision you want.  The following are in increasing order of cost / complexity / performance.

1.  Zero-beat the OCXO against WWV at 10 MHz.  Requires a 10 MHz shortwave receiver.  Maybe 1 Hz accuracy.

2.  Measure the carrier frequency of WWVB at 60 KHz.  Requires a tuned LC circuit / antenna and amplifier for 60 KHz.  Maybe 0.1 Hz.

Now you jump to accuracy levels in the 1e-9 to 1e-12 level.

3.  Rb standard.  Assumes that the unit is working and has been calibrated against a known standard.  Drift is in the range of 5e-11 per month.  Jitter isn't an issue.

4.  Bare GPS receiver.  Measure the output frequency.  Usually 1 PPS although some units have an auxiliary output with a programmable frequency.  Requires a decent antenna location.  No drift.  Jitter will be in the range of 5 - 100 ns rms depending on which unit you get.  Make sure you get a timing GPS rather than a navigation GPS.  Navigation GPS units can have 1 PPS jitter in the microsecond range.

5.  GPSDO.  Requires a decent antenna location.  No drift.  Jitter is < 1 ns rms.
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #4 on: July 11, 2015, 08:52:52 pm »
Economical is a very relative term. Something like this is ready to go:
http://www.ebay.com/itm/281701995263
I have no experience with this unit nor the seller but it looks good.

Another option is this:
http://www.ebay.com/itm/FEI-fe-5680b-rubidium-oscillator-With-1pps-20mhz-output-ONLY-10mhz-NEED-to-MOD-/291419889143?pt=LH_DefaultDomain_0&hash=item43d9fa9df7
I purchased one from this seller and have modified it to output 10MHz. It was only a handful of cheap parts to do so. The information is mentioned in his auction. If you need cheaper than $65, then the options get rather few.

The GPS option there looks good. I wonder if anyone here has any experience with that unit, though. I'm a bit leary about buying something from China with no confirmation that it's not junk.
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #5 on: July 11, 2015, 08:56:09 pm »
The answer really depends on the precision you want.  The following are in increasing order of cost / complexity / performance.

1.  Zero-beat the OCXO against WWV at 10 MHz.  Requires a 10 MHz shortwave receiver.  Maybe 1 Hz accuracy.

2.  Measure the carrier frequency of WWVB at 60 KHz.  Requires a tuned LC circuit / antenna and amplifier for 60 KHz.  Maybe 0.1 Hz.

Now you jump to accuracy levels in the 1e-9 to 1e-12 level.

3.  Rb standard.  Assumes that the unit is working and has been calibrated against a known standard.  Drift is in the range of 5e-11 per month.  Jitter isn't an issue.

4.  Bare GPS receiver.  Measure the output frequency.  Usually 1 PPS although some units have an auxiliary output with a programmable frequency.  Requires a decent antenna location.  No drift.  Jitter will be in the range of 5 - 100 ns rms depending on which unit you get.  Make sure you get a timing GPS rather than a navigation GPS.  Navigation GPS units can have 1 PPS jitter in the microsecond range.

5.  GPSDO.  Requires a decent antenna location.  No drift.  Jitter is < 1 ns rms.

This is a great list. I really want something that I don't have to periodically recalibrate. There's a window right where I would want to place a GPSDO, so I don't think antenna location will be an issue. So given that, a GPSDO seems like the way to go.

Any suggestions on options for that? What about the unit that Lightages posted a link to in the thread above? From the pictures in the listing, it looks like it is using a GPS chip intended for navigation (uBlox NEO-6M). Data sheet on that is here:

https://www.u-blox.com/images/downloads/Product_Docs/NEO-6_DataSheet_(GPS.G6-HW-09005).pdf
« Last Edit: July 11, 2015, 09:01:30 pm by motocoder »
 

Online blueskull

  • Supporter
  • ****
  • Posts: 14199
  • Country: cn
  • Power Electronics Guy
Re: Economical option for precision frequency reference?
« Reply #6 on: July 11, 2015, 09:28:51 pm »
I have an old X72 Rb clock, which is rated for 20 years operation, and it has been ran for about 8 years. I would say it will run happily for another at least 5 years. Currently it locks in 40 seconds.

Also, I have 2 new old stock stratum 3 20Mhz OCXOs from CTS.

If you are interested in them, please PM me a price you want to pay for one or all of them.

I am currently in Raleigh, NC, and I ship to any lower 48 states.
 

Online edpalmer42

  • Super Contributor
  • ***
  • Posts: 1820
  • Country: ca
Re: Economical option for precision frequency reference?
« Reply #7 on: July 11, 2015, 10:21:37 pm »
This is a great list. I really want something that I don't have to periodically recalibrate. There's a window right where I would want to place a GPSDO, so I don't think antenna location will be an issue. So given that, a GPSDO seems like the way to go.

Any suggestions on options for that? What about the unit that Lightages posted a link to in the thread above? From the pictures in the listing, it looks like it is using a GPS chip intended for navigation (uBlox NEO-6M). Data sheet on that is here:

https://www.u-blox.com/images/downloads/Product_Docs/NEO-6_DataSheet_(GPS.G6-HW-09005).pdf

Remember that recalibration intervals are related to required accuracy.  If you have a calibrated Rb standard that drifts at 5e-11 per month, and your accuracy needs are only (!) 1 ppb, you only have to recalibrate every 20 months.  Maybe not a problem.

The BG7TBL GPSDO has been evaluated here:  http://www.ke5fx.com/gpscomp.htm .  It puzzles me a little.  Based on the graphs, it isn't performing as well as the other GPSDOs that I'm familiar with.  Using the NEO-6M would have an effect, but that's not enough to explain the performance.

However, due to a happy coincidence, there is another option.  I suggest you look at this one:  http://www.ebay.com/itm/221777430088 .  Since it's on sale, the price is very close to the BG7TBL unit.  With names like Lucent and Symmetricom behind it, you know it's going to be good!  There was a long discussion on this unit on the Time-Nuts mailing list some months ago.  Lots of reverse-engineering and info.  The mailing list archive is here:  https://www.febo.com/pipermail/time-nuts/ .

« Last Edit: July 11, 2015, 10:24:10 pm by edpalmer42 »
 

Offline Electro Fan

  • Super Contributor
  • ***
  • Posts: 2542
Re: Economical option for precision frequency reference?
« Reply #8 on: July 11, 2015, 10:36:40 pm »
Economical is a very relative term. Something like this is ready to go:
http://www.ebay.com/itm/281701995263
I have no experience with this unit nor the seller but it looks good.

Another option is this:
http://www.ebay.com/itm/FEI-fe-5680b-rubidium-oscillator-With-1pps-20mhz-output-ONLY-10mhz-NEED-to-MOD-/291419889143?pt=LH_DefaultDomain_0&hash=item43d9fa9df7
I purchased one from this seller and have modified it to output 10MHz. It was only a handful of cheap parts to do so. The information is mentioned in his auction. If you need cheaper than $65, then the options get rather few.

The GPS option there looks good. I wonder if anyone here has any experience with that unit, though. I'm a bit leary about buying something from China with no confirmation that it's not junk.


I think member TSL has experience either with the GPSDO unit or a similar unit - I think he did some extensive testing.  Maybe he will surface and see this - he is very experienced with precise clocking.
 

Offline G0HZU

  • Super Contributor
  • ***
  • Posts: 2576
  • Country: gb
Re: Economical option for precision frequency reference?
« Reply #9 on: July 11, 2015, 11:48:06 pm »
Quote
2.  Measure the carrier frequency of WWVB at 60 KHz.  Requires a tuned LC circuit / antenna and amplifier for 60 KHz.  Maybe 0.1 Hz.

That's the best all round option IMO in terms of cost, speed and reliability. I tried testing a similar cheap and cheerful circuit for 198kHz (BBC Droitwich, UK) some time ago and got very good results with just a simple antenna, amplifier and a hard limiter chip. All of the above can be built very quickly from surplus parts for next to zero cost. The performance wasn't as good as my old homebrew offair standard but it was good enough :)

When used with a reciprocal counter it should completely outclass the GPSDO and Rb options in terms of cost and speed/convenience to calibrate as you should be able to get a good result in less than 1 minute from cold. Obviously, the offair std isn't as accurate as a well settled GPSDO but I can't think of a valid reason why I'd need the extra accuracy because nothing I have here in terms of test gear or things I've designed can maintain better accuracy over a few months than the simple offair standard can deliver when averaged over a few minutes.

Note:

I've got a (Jackson Labs?) GPSDO here inside some comms equipment and apart from powering up the GPSDO module once to see how painfully slow it is to lock up and how much of a PITA it is to deploy the antenna outside I haven't used it since. I think I've had it here several months now.

By comparison, a simple LF offair standard can lock and be useful in about 1 minute from cold and it runs on flea power and can be placed anywhere in the house :)





« Last Edit: July 11, 2015, 11:50:38 pm by G0HZU »
 

Offline Electro Fan

  • Super Contributor
  • ***
  • Posts: 2542
Re: Economical option for precision frequency reference?
« Reply #10 on: July 12, 2015, 12:00:23 am »
 a quick Google search found these:

http://www.leapsecond.com/pages/sony-wwvb/

http://www.joejaworski.com/wwvb/

https://www.sparkfun.com/products/retired/10060

- maybe someone can post some better info/instructions for building something
 

Online edpalmer42

  • Super Contributor
  • ***
  • Posts: 1820
  • Country: ca
Re: Economical option for precision frequency reference?
« Reply #11 on: July 12, 2015, 12:11:18 am »
If you're going to use WWVB (or similar low frequency broadcast) you don't really want one of the 'atomic' clocks.  They decode the signal and give you time of day.  But typically, the carrier frequency is locked to an atomic clock, while the modulation may not be.  As a result, any signal derived from the clock information may not be as stable as the carrier itself.

What you need is the RF carrier frequency (i.e. 60 KHz).  You need to ignore or filter out whatever modulation is present.  So you use a 60 KHz tuned RF circuit (maybe a ferrite rod), coil, and tuning capacitor plus an amplifier / limiter to boost the signal and block out the modulation.  I've seen various web sites that talk about this, but can't come up with any links right now.

« Last Edit: July 12, 2015, 12:19:57 am by edpalmer42 »
 

Online PaulAm

  • Frequent Contributor
  • **
  • Posts: 722
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #12 on: July 12, 2015, 12:49:20 am »
The GPSDO that Edpalmer42 mentioned is down to $100 which is $25 less than when I bought one.

That's got to be the best deal around at the moment.  It's NOS still in the original packing containers.  Add a power supply, an antenna and you're good to go.
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #13 on: July 12, 2015, 04:14:42 am »
The GPSDO that Edpalmer42 mentioned is down to $100 which is $25 less than when I bought one.

That's got to be the best deal around at the moment.  It's NOS still in the original packing containers.  Add a power supply, an antenna and you're good to go.

For $100, I think I'll give it a go. It seems like it will be a fun project to get that up and working. Any suggestions on PS and Antenna?

Edpalmer42 - thanks for the recommendations. I may also do some experiments with WWV, but I'd like to get the GPSDO working first.
 

Offline Wuerstchenhund

  • Super Contributor
  • ***
  • Posts: 3055
  • Country: gb
Re: Economical option for precision frequency reference?
« Reply #14 on: July 12, 2015, 11:12:10 am »
Economical is a very relative term. Something like this is ready to go:
http://www.ebay.com/itm/281701995263
I have no experience with this unit nor the seller but it looks good.

I have this one, and while I haven't cross-checked it with another source I'm sure it does what it says on the tin, although as described at http://www.ke5fx.com/gpscomp.htm it doesn't seem to run at exactly 10MHz but 9.9999999998MHz instead.

The main reason I went for this instead of an used HP/Agilent or whatever GPS disciplined OCXO were price, the absence of need for modification before use, and that it's rather compact and not in some odd slot-in module format with backplane connector. I did replace the GPS antenna that came with the unit with a better one which gives me better receiption, though.
« Last Edit: July 12, 2015, 11:19:40 am by Wuerstchenhund »
 

Offline jpb

  • Super Contributor
  • ***
  • Posts: 1732
  • Country: gb
Re: Economical option for precision frequency reference?
« Reply #15 on: July 12, 2015, 12:23:38 pm »
If you can set a long gate time on the counter and it measures low frequencies (1Hz) then you can check a 10MHz OCXO by using it as a reference and measuring the 1pps from a cheap (£30) GPS module.

I did this with my TF930 counter. Of course you need to do it over a long time as there is quite a lot of jitter  from pulse-to-pulse.

The module I use is this one from Adafruit as it has an external antenna and is quite an accurate one ( 10nsecs according to the spec sheet - others are up to 50 nsecs or more):

http://proto-pic.co.uk/ultimate-gps-breakout-66-channel-w-10-hz-updates-mtk3339-chipset/

The other method is to use a scope, use the 1pps to trigger the scope and observe the phase shift on the 10MHz (assuming the phase shift is less than the 100 nsec period which it should be).
 

Online edpalmer42

  • Super Contributor
  • ***
  • Posts: 1820
  • Country: ca
Re: Economical option for precision frequency reference?
« Reply #16 on: July 12, 2015, 04:35:19 pm »
For $100, I think I'll give it a go. It seems like it will be a fun project to get that up and working. Any suggestions on PS and Antenna?

Any good quality power supply should be fine.  People tend to prefer linear supplies for frequency references, but since most of the commercial units have internal switching supplies I don't think it makes much difference.  It looks like each of the two units needs 24V at 2 amps.  You don't have to run both units - remember that this was for Lucent who always specifies redundant systems - but it's probably best to start with both to make sure everything is working properly.

Any 5 volt GPS antenna will work.  Some of the newer antennas are 3.3 volt only so look carefully.  However, you'll get the best performance with a timing-grade antenna.  They typically have narrower filtering than navigation antennas. Just search for <gps antenna timing> and start browsing.

The antenna should be mounted outdoors with a clear view of the sky towards the equator (since that's the direction where you'll see more satellites).  You may be able to get away with an indoor antenna, but it will definitely reduce the signal levels and the number of satellites tracked.  It may cause the unit to lose lock and go into holdover.  More gain is better.  These older GPSDOs tend to have poorer sensitivity than current ones.  This could become an issue if you have a long cable run from the antenna to the receiver.  In the manual for the Trimble Thunderbolt they recommend RG-59 or RG-6 cable between the antenna and the receiver.  They basically say "Yes, we know it's 75 ohms and the equipment is 50 ohms.  It doesn't matter.  The cable is good, cheap, and widely available.  Shut up and use it!"  ... or words to that effect.  :)

 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #17 on: July 12, 2015, 11:06:35 pm »
For $100, I think I'll give it a go. It seems like it will be a fun project to get that up and working. Any suggestions on PS and Antenna?

Any good quality power supply should be fine.  People tend to prefer linear supplies for frequency references, but since most of the commercial units have internal switching supplies I don't think it makes much difference.  It looks like each of the two units needs 24V at 2 amps.  You don't have to run both units - remember that this was for Lucent who always specifies redundant systems - but it's probably best to start with both to make sure everything is working properly.

Any 5 volt GPS antenna will work.  Some of the newer antennas are 3.3 volt only so look carefully.  However, you'll get the best performance with a timing-grade antenna.  They typically have narrower filtering than navigation antennas. Just search for <gps antenna timing> and start browsing.

The antenna should be mounted outdoors with a clear view of the sky towards the equator (since that's the direction where you'll see more satellites).  You may be able to get away with an indoor antenna, but it will definitely reduce the signal levels and the number of satellites tracked.  It may cause the unit to lose lock and go into holdover.  More gain is better.  These older GPSDOs tend to have poorer sensitivity than current ones.  This could become an issue if you have a long cable run from the antenna to the receiver.  In the manual for the Trimble Thunderbolt they recommend RG-59 or RG-6 cable between the antenna and the receiver.  They basically say "Yes, we know it's 75 ohms and the equipment is 50 ohms.  It doesn't matter.  The cable is good, cheap, and widely available.  Shut up and use it!"  ... or words to that effect.  :)

I'm not worried about the cable. I have my ham radio AE license, and used to tinker around with radios, so I should have some good 50 ohm coax around.

Thanks for the tips on the PS and antenna. I can probably find a 24V dongle on eBay, and will look for an antenna with the criteria you suggest.
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #18 on: July 12, 2015, 11:07:27 pm »
If you can set a long gate time on the counter and it measures low frequencies (1Hz) then you can check a 10MHz OCXO by using it as a reference and measuring the 1pps from a cheap (£30) GPS module.

I did this with my TF930 counter. Of course you need to do it over a long time as there is quite a lot of jitter  from pulse-to-pulse.

The module I use is this one from Adafruit as it has an external antenna and is quite an accurate one ( 10nsecs according to the spec sheet - others are up to 50 nsecs or more):

http://proto-pic.co.uk/ultimate-gps-breakout-66-channel-w-10-hz-updates-mtk3339-chipset/

The other method is to use a scope, use the 1pps to trigger the scope and observe the phase shift on the 10MHz (assuming the phase shift is less than the 100 nsec period which it should be).

That sounds like a good way to double-check the reference as well. Thanks
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #19 on: July 13, 2015, 11:01:24 pm »
For $100, I think I'll give it a go. It seems like it will be a fun project to get that up and working. Any suggestions on PS and Antenna?

Any good quality power supply should be fine.  People tend to prefer linear supplies for frequency references, but since most of the commercial units have internal switching supplies I don't think it makes much difference.  It looks like each of the two units needs 24V at 2 amps.  You don't have to run both units - remember that this was for Lucent who always specifies redundant systems - but it's probably best to start with both to make sure everything is working properly.

Any 5 volt GPS antenna will work.  Some of the newer antennas are 3.3 volt only so look carefully.  However, you'll get the best performance with a timing-grade antenna.  They typically have narrower filtering than navigation antennas. Just search for <gps antenna timing> and start browsing.

The antenna should be mounted outdoors with a clear view of the sky towards the equator (since that's the direction where you'll see more satellites).  You may be able to get away with an indoor antenna, but it will definitely reduce the signal levels and the number of satellites tracked.  It may cause the unit to lose lock and go into holdover.  More gain is better.  These older GPSDOs tend to have poorer sensitivity than current ones.  This could become an issue if you have a long cable run from the antenna to the receiver.  In the manual for the Trimble Thunderbolt they recommend RG-59 or RG-6 cable between the antenna and the receiver.  They basically say "Yes, we know it's 75 ohms and the equipment is 50 ohms.  It doesn't matter.  The cable is good, cheap, and widely available.  Shut up and use it!"  ... or words to that effect.  :)

Any thoughts on either of these antennas (i think they might be the same, barring the read color of the collar)?

http://www.ebay.com/itm/NIB-Lucent-PCTEL-GPS-TMG-HR-26NCM-26dB-Timing-Antenna-N-f-w-Collar-Mount-MaxRad-/301687250710?pt=LH_DefaultDomain_0&hash=item463df61f16
http://www.ebay.com/itm/Datum-2640NW-D-E-GPS-26dB-5V-Antenna-Telecom-Timing-N-Conn-PCTEL-Lucent-NEW-/361230143301?pt=LH_DefaultDomain_0&hash=item541afe8f45

I would also need to source a cable with the correct connectors. I have some cable, but not with these connectors on them. Looks like I would need a TNC male on one end and an N male on the other.

 

Offline Lightages

  • Supporter
  • ****
  • Posts: 4302
  • Country: ca
  • Canadian po
Re: Economical option for precision frequency reference?
« Reply #20 on: July 13, 2015, 11:08:39 pm »
I decided to get the unit I posted earlier from ebay. It should make a nice video when I get back to making videos.
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #21 on: July 13, 2015, 11:43:05 pm »
I decided to get the unit I posted earlier from ebay. It should make a nice video when I get back to making videos.

I am looking forward to seeing that.
 

Offline N8AUM

  • Regular Contributor
  • *
  • Posts: 131
  • Country: us
Re: Economical option for precision frequency reference?
« Reply #22 on: July 18, 2015, 09:27:27 am »
Heres my 2 cents worth.
After installing a High Stability Timebase Option (clone) for my 53131a I needed to calibrate it some how. I tried zero beating WWV's carrier but I quickly learned about the varying propagation delay caused by the variation of the ionosphere and receiving WWVH at the same time gave me too much error so my brother suggested using a GPS Disciplined Clock as a reverence source. The last time that I looked the cost was just way too much for me. Just recently I found the prices were getting reasonable so I purchased a Trimble Thunderbolt. After the 30 day warranty ran out it started to reboot on occasion so I found the BG7TBL GPSDO from China and decided what the hell, I will get one. I must admit, the build quality was pretty impressive and I liked the small size so I picked up 2 more as backups lol. With the el-cheapo supplied antenna placed on the north side of the house (north is bad) it tracks 12 birds most of the time. Just for giggles I stuck a resistor in the antenna connector and to my surprise it tracked 4 birds ! 

Screen shot of the BG7TBL:
 
« Last Edit: July 18, 2015, 09:54:41 am by N8AUM »
 

Online edpalmer42

  • Super Contributor
  • ***
  • Posts: 1820
  • Country: ca
Re: Economical option for precision frequency reference?
« Reply #23 on: July 18, 2015, 09:28:13 pm »
Here are the results of my tests on my Lucent RFTG-u GPSDO.

The first picture shows the phase measurements between the GPSDO and my Efratom FRT Rb standard.  For comparison, I've also shown the same measurement with my HP Z3801A GPSDO.  It's much better to measure phase - i.e. time interval - than frequency difference because frequency measurement implies averaging over a period of time or requires unobtainable accuracy while phase difference is a relatively simple measurement.  Also, many counters measure time interval much more accurately than frequency.  Odd, but true.  The graph shows that the RFTG-u is much busier than the Z3801A, but is otherwise working properly.  I don't know whether it will settle down over time or can be tweaked.  Most GPSDOs don't allow you to modify the parameters of the internal control loops.  The Trimble Thunderbolt is a notable exception.

The second picture shows the Allan Deviation results for both measurements.  If you're not familiar with Allan Deviation, it's the standard calculation used to evaluate the noise and stability of any kind of oscillator.  The math is beyond me, but I think I've got the operational basics figured out.

In these graphs, the X-axis value represents measurement duration or interval between measurements while the Y-axis represents stability or resolution.  For example, if you make a measurement that's one second long, your resolution can't be better than 3.6e-11.  If you extend that to 10 seconds with the Z3801A, you can resolve to about 4e-12.  The RFTG-u is noisier at 10 seconds, so your resolution will be 6e-12.  If you get into the underlying mathematics you'll find that my explanation isn't quite right, but it's good enough for now.

But wait, that's still not right!  Each graph is actually a composite measurement.  The measurement system, the device under test (the RFTG-u) and the reference (the FRT) all contribute to the graph.  At any point along the X-axis, the graph represents the combination of the three elements.  Typically, it represents the performance of the worst of the three at that point in time.

In my graphs, the straight-line segment on the left of the graph represents the limitations of my measurement system.  It's only when the graph deviates from a straight line that you see the performance of the oscillators inside the GPSDOs.  Hopefully, as time passes, the oscillator in the RFTG-u will settle down and the relatively flat section of the graph will drop lower, perhaps dropping as far as the Z3801A graph or beyond.

But just after the oscillators have started to show their worth, they run into the 'GPS line'.  That's my name for it.  It represents the approximate performance limit of the GPS system.  The oscillators are forced to turn and follow that line.  Every GPSDO will do the same thing.  Sometimes one side of the line, sometimes the other.  It will vary from unit to unit and maybe data run to data run.  Normally, the oscillator would continue toward the upper right of the graph as aging became significant so, in this case, the graph isn't really showing the worst of the three elements.

So where's the FRT in this explanation?  It's stability is good enough that it hasn't made an appearance yet.  It's running lower than the other elements in the test so it's not visible.  If I ran the test longer, both graphs would start to rise as the aging of the FRT became significant.  The upwards hook on the end of the black graph might be the FRT, but the end of these graphs flap around quite dramatically so you can't be sure.  This characteristic is inherent in the mathematics involved and doesn't represent a deficiency in the equipment, the measurement, or the calculations.

The software used to collect this data and generate the graphs is called Timelab, an excellent, freeware, open source program written and actively supported by John Miles and available from http://www.ke5fx.com/timelab/readme.htm .

Ed

« Last Edit: July 18, 2015, 09:30:15 pm by edpalmer42 »
 

Offline motocoder

  • Supporter
  • ****
  • Posts: 769
  • Country: us
  • Electrical Engineer
Re: Economical option for precision frequency reference?
« Reply #24 on: July 19, 2015, 12:01:21 am »
Heres my 2 cents worth.
After installing a High Stability Timebase Option (clone) for my 53131a I needed to calibrate it some how. I tried zero beating WWV's carrier but I quickly learned about the varying propagation delay caused by the variation of the ionosphere and receiving WWVH at the same time gave me too much error so my brother suggested using a GPS Disciplined Clock as a reverence source. The last time that I looked the cost was just way too much for me. Just recently I found the prices were getting reasonable so I purchased a Trimble Thunderbolt. After the 30 day warranty ran out it started to reboot on occasion so I found the BG7TBL GPSDO from China and decided what the hell, I will get one. I must admit, the build quality was pretty impressive and I liked the small size so I picked up 2 more as backups lol. With the el-cheapo supplied antenna placed on the north side of the house (north is bad) it tracks 12 birds most of the time. Just for giggles I stuck a resistor in the antenna connector and to my surprise it tracked 4 birds ! 

Screen shot of the BG7TBL:

I think it's a good option. I may end up with one of these at some point, but it will be fun to get the Lucent unit running.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf