GPSDO Tutorial?

[Segrest]

Hello Scott,

Thank you for the correction. I should know better than post when I can't get to sleep by now...

I will try the WWV verification this evening.

Your guidance on the distribution amplifier is appreciated. I will put it on my toy list.

Bob
KO2F

[WA0VY]

Well Scott, you ALMOST have me convinced that I can't live without one of these things. I had no idea there was such a variance in these devices and so may different things you can do with them. So, that commentary was very interesting.

I currently manually calibrate and my correction factor is 0.99999962. Auto worked one time, early on, and never worked again. I have no idea why. Working back and forth on 10 mHz WWV from LSB to USB and DSB, I can detect a difference between 0.99999958 and 0.99999966, a tiny WOWWOW at the edges. Once calibrated, the radio is very stable and does not require readjustment. I am wondering how close I get doing a manual calibration. If anyone can figure it, I'd love to see the calculation.

[w-u-2-o]

Brent,

You don't really need one. Few people do. It probably does make things a lot easier for folks who are transverting, or EME'ing, or, you know, the sort of things where starting 1Hz off is already a problem.

If you look at the phase display when manually calibrating against WWV in DSB mode, you will see the phase point spinning around in a circle. The rate at which it goes around is the amount you are off frequency by. Most people can easily get this to be less than 1Hz manually.

73,

Scott

[WA0VY]

Scott, I assume you mean Phase 2? Wow, I was very close. At 0.99999960, it barely moves counterclockwise. At 0.99999959, it reverses and is noticeably faster. So, my ear did a decent enough job.

[WA0VY]

Ok, missed the part about needing to be in DSB. However, I got the exact same result using phase and DSB as I did using USB and Phase 2 - just that the display looked somewhat different. Very nice little trick.

[Buck]

Hi Bob,
As for distribution to multiple radios, since the 10MHz inputs on the Apache Labs radios are high impedance, you can generally get away with using T connectors. However, to do it in a more proper fashion a distribution amplifier is the best way. There is a matching BG7TBL designed distribution amplifier that can be found on eBay.

73,

Scott

I put an ANZAC T1000 coupler between the GPSDO and the feed to the 200D. It worked fine, showing zero correction on WWV. These T-couplers cost $20-30 on ebay.

Buck ko0y

[Jay_N2MGA]

the DSB/phase display works very well!

jay
n2mga

[w-u-2-o]

Buck has already reported that he is within 3 milliHertz--3 thousands of one Hertz. I think he's all set!

[kd8tut]

Side note for people playing in the time domain:

I've jury rigged an Android tablet to move it's NMEA data via the local network to provide a local clock for an NTP daemon running on a PC. it provides a suitable backup clock for when the internet is unavailable and provides a few ms time offset from NIST. Because of the latencies involved (tablet---> wireless---> wired LAN---> PC) it's not good as a stratum 0 clock. Though when forced to be a stratum 3 clock source in NTPD it can act as an "automagical" failover for folks running their own local stratum 2 server.

This arrangement runs + or - 15ms from the NIST standard- well within the needs of a ham. This is still an active project though, it is possible with NMEA filtering, fudge time1/time2, and PPS implementation that it could provide stratum 1 accuracy. Still working with it.

Secondarily I've jury rigged a Windows 10 tablet with the Broadcom GNSS 4752 Geolocation Sensor to do the same thing as above- and it's a waste of time. The jitter and offset are terrible when converting data from the Windows location API to NMEA.

kd8tut@arrl.net

[Tony EI7BMB]

I sere LeoBodnar now has a cheaper single output unit at STG 99 http://www.leobodnar.com/shop/index.php ... cts_id=301

[WA2DVU]

Don't get upset if your GPS and WWV don't match. WWV at 10MHZ can be off at your QTH 1HZ or more due to our friend doppler. I am one of the "FMT Nuts" group and when measuring fx to a few Millihertz, a dart board is your friend! 73, Bill, WA2DVU

[Tony EI7BMB]

Thanks Bill, good to know

[w-u-2-o]

If I'm within 1Hz without having to lift a finger, and I am , then I'm happy.

Plus I get to bitch about how badly everyone else is off frequency

[WA0VY]

I received my BG7TBL unit today with a Bliley unit and am delighted to report a near perfect, effortless calibration. WWV at 10 mHz shows a completely centered meatball on Phase 2 in the panadapter with no correction dialed in. I could never accomplish that with a manual calibration. I have three devices capable of taking the 10 mHz reference signal. So, the first question is whether a distribution unit is needed or whether I can get a decent signal distributed with a splitter. Info in the forum suggests that a distribution unit is the way to go but that's another $100 and I see on eBay that there is a long lag time in delivery on those units.

Second question is re: the 1PPS out to the computer for NTP timing. I have read Scott's references in the link. At first glance, this does not look like an easy setup. I Googled the issue to see if there was anything out there that was more straightforward and mostly saw nonsense in blogs or the same setup info which Scott has provided. I will give it a try and see if I can get it working, but if anyone has done this and has a shortcut procedure, I'd love to hear about it.

Thanks in advance,

[w-u-2-o]

Hi Brent,

You could probably get away with a hybrid, 3dB, two-way power divider, but for more than a two-way split you really need a distribution amp.

As for rolling your own NTP server using the 1PPS output, there are no shortcuts. I don't think it's all that hard. However, that said, it is kind of a labor of love. If you've got the interweb, Dimension 4 is more than adequate, and, if you are a purist, Meinberg NTP is definitely even more adequate. I'm a Meinberg kind of guy

73!

Scott

[WA0VY]

Ok Scott, I will give a try this weekend. Having reread your posts on this point, I wonder what you mean by "[t]he 1PPS needs to be properly buffered ...." Are you saying this requires something on the serial line connection other than the cable? Or are you referring to something internal to the GPSDO unit? If the latter, then can I assume the Chinese BG7TBL unit is good to go?

BTW, before I installed the GPSDO unit, I was only able to get my radio to auto calibrate one time during which it showed a correction. Thereafter, the correction never moved from 1.00000000 (not sure I have the right number of "0's" here but you get the point) and was noticeably off. I always had to do a manual calibration to WWV, usually getting a slight and always repeatable downward offset. I measured the actual difference of my 8000 at 1.00000000 from WWV and it was 4 Hz. That seemed like a lot and was most noticeable. I wonder why I could never again achieve an auto calibration.

[w-u-2-o]

Brent,

The 1PPS signal needs to be brought to RS-232 voltage levels if you are using the software that uses an RS-232 control pin as a 1PPS input.

On my BG7TBL unit it appears that this condition is already met. However, if someone were using a board with a 3 or 5V TTL 1PPS output it might not work without a proper TTL to RS-232 buffer.

73,

Scott

[Jay_N2MGA]

Hey Brent, if you're still looking to setup a distribution board for the GPSDO have a look at rolling ur own at the following site; http://www.medpants.com/distribution-amplifier

pretty straight forward...

jay
n2mga

[Jay_N2MGA]

While on the topic of external GPSDO devices, I have a question I'd like to throw out to anyone that has an answer regarding the 8000 and the 10Mhz external input. On the main board this input traces back to, there are two jumpers labeled J24 and J22, the schematic shows these as "EXT CLK GND" and "EXT CLK TERMINATION" respectively. From the factory, are these jumpers installed on the 8000? And what are their purpose regarding the use of an external reference?

Any ideas/information are greatly appreciated, thank you...

jay
n2mga

[WA0VY]

Thanks Jay. That's a nice solution. However, it looks like I only need two outputs, not three. So, I just ordered a t-connector for a few bucks. We'll see if works.

[w-u-2-o]

Thanks Jay. That's a nice solution. However, it looks like I only need two outputs, not three. So, I just ordered a t-connector for a few bucks. We'll see if works.

A BNC tee did not work for me to feed both the 8000 and 100D. Both radios had to be powered up for it to work. With one or the other radio powered off it loaded down the signal too much.

I've actually ordered a 3dB 0 deg. hybrid splitter from Mini Circuits and we'll see if it works. I got the hybrid instead of the resistive splitter because it has much better port to port isolation. And zero degree so that things remain in phase, of course!

[w-u-2-o]

While on the topic of external GPSDO devices, I have a question I'd like to throw out to anyone that has an answer regarding the 8000 and the 10Mhz external input. On the main board this input traces back to, there are two jumpers labeled J24 and J22, the schematic shows these as "EXT CLK GND" and "EXT CLK TERMINATION" respectively. From the factory, are these jumpers installed on the 8000? And what are their purpose regarding the use of an external reference?

I don't know if these are installed by default or not. I suspect that the termination jumper is not. At any rate, my BG7TBL reference works without doing anything to the radio.

The termination jumper allows you to choose between a high impedance input and a 50 ohm impedance input. This may or may not be important for proper output from your external reference. It can also be important for preventing reflections from the end of the line if you are daisy chaining your reference to multiple devices--terminate the last device in the chain.

The ground jumper changes the input from balanced to unbalanced.

73,

Scott

[WA0VY]

I've actually ordered a 3dB 0 deg. hybrid splitter from Mini Circuits and we'll see if it works.

Scott, let me know what model you ordered and how it works. Some of these are very expensive compared to a distribution amplifier, so I'm guessing I have not found what you ordered.

[w-u-2-o]

I ordered a ZFSC-2-4+

https://www.minicircuits.com/pdfs/ZFSC-2-4.pdf

I should be trying it over the weekend.

[WA0VY]

I ordered a ZFSC-2-4+

https://www.minicircuits.com/pdfs/ZFSC-2-4.pdf

I should be trying it over the weekend.

Excellent. I had a hard time navigating that site. I'll look forward to your report.

[Jay_N2MGA]

While on the topic of external GPSDO devices, I have a question I'd like to throw out to anyone that has an answer regarding the 8000 and the 10Mhz external input. On the main board this input traces back to, there are two jumpers labeled J24 and J22, the schematic shows these as "EXT CLK GND" and "EXT CLK TERMINATION" respectively. From the factory, are these jumpers installed on the 8000? And what are their purpose regarding the use of an external reference?

I don't know if these are installed by default or not. I suspect that the termination jumper is not. At any rate, my BG7TBL reference works without doing anything to the radio.

The termination jumper allows you to choose between a high impedance input and a 50 ohm impedance input. This may or may not be important for proper output from your external reference. It can also be important for preventing reflections from the end of the line if you are daisy chaining your reference to multiple devices--terminate the last device in the chain.

The ground jumper changes the input from balanced to unbalanced.

73,

Scott

Scott, thank you for confirming all that.

jay
n2mga

[w-u-2-o]

I ordered a ZFSC-2-4+

https://www.minicircuits.com/pdfs/ZFSC-2-4.pdf

I should be trying it over the weekend.

OK, I got this baby hooked up and it seems to work pretty good. I can power up both radios, or just one at a time, and the 10MHz external reference is working. The output of the BG7TBL I have here puts out +13dBm into a 50 Ohm load, so it's down to about +9 after the splitter, but that seems to be enough. If I had to feed more than two radios I'd go with a real distribution amplifier.

73,

Scott

[WA0VY]

OK, I got this baby hooked up and it seems to work pretty good.

Scott, good to hear. Not sure what I'm going to do here yet but I did note that the BG7TBL distribution amps can be way out on delivery times. I do have a t-connector so I will try that first, of course, but don't expect it will work any better for me than it did for you.

[w-u-2-o]

I had another idea that you might try, Brent, and I'm kind of sorry I didn't try it here before I purchased the power divider. And that idea was to use two T's and a 50 Ohm terminator. Put a T on each radio, run the 10MHz from the GPSDO to the first radio, then the second, and put a 50 Ohm terminator on the T at the end of the line.

[w-u-2-o]

There would appear to be a new, even less expensive, Chinese designed and produced GPSDO on the eBay marketplace now:



I have no technical details, and don't know anyone using one, yet, but I'm guessing it uses a steerable TCXO as opposed to a more expensive OCXO.