#progrock GPS threshold adjustment #progrock #qlg1


Victor 4X6GP
 

I'm trying to adjust the GPS threshold on my Progrock. It is programmed to produce 24 MHz which serves as an external clock for an SDRplay RSPdx SDR. The Progrock is connected to a QLG1.

With the default threshold of 5 Hz, the error when receiving WWV or RWM on 10 MHz is around 2-3 Hz. I thought I could improve this by setting the threshold to 1, which should adjust the 27 MHz clock in the Progrock to within 1 Hz. I did this by entering 031F with the Progrock's pushbutton.

But the result was that the error seems to be worse, around 3.5 Hz. Did I do something wrong? Should I have entered 0301F to get 1 Hz?
Victor


Mike Berg
 

I think your syntax is right.   I use the serial comms to tweak mine and just left the GPS param at 5.

Reading the docs says a setting of 0 (continuous correcting) may not work.

I'd try different numbers from 0-5 and see what works.
73
Mike N0QBH


Victor 4X6GP
 

Thanks, Mike. I'll do some more experiments today.

Victor

On Tue, Jan 5, 2021, 02:45 Mike Berg <mikeberg@...> wrote:

I think your syntax is right.   I use the serial comms to tweak mine and just left the GPS param at 5.

Reading the docs says a setting of 0 (continuous correcting) may not work.

I'd try different numbers from 0-5 and see what works.
73
Mike N0QBH


Alan G4ZFQ
 

the error when receiving WWV or RWM on 10 MHz is around 2-3 Hz. I thought I could improve this by setting the threshold to 1,
Victor,

I do not normally use a GPS with my Progrock.
But when I tried it I never saw it lock exactly, it was always about 2Hz off frequency.
There was a discussion about this then, maybe a few years ago. I do not remember a conclusion.

73 Alan G4ZFQ


Victor 4X6GP
 

This is interesting! My progrock plus qlg1 have been running 24/7 for a few days, and the error appears to be about 2 Hz, or slightly less.

I will try searching the forum.

Victor
On 05/01/2021 9:55, Alan G4ZFQ wrote:

the error when receiving WWV or RWM on 10 MHz is around 2-3 Hz. I thought I could improve this by setting the threshold to 1,

Victor,

I do not normally use a GPS with my Progrock.
But when I tried it I never saw it lock exactly, it was always about 2Hz off frequency.
There was a discussion about this then, maybe a few years ago. I do not remember a conclusion.

73 Alan G4ZFQ






Giuseppe Marullo[iw2jww]
 

Yes,

I have the same problem (around 2Hz less with thresold of 0), but I only have another GPSDO so cannot tell for sure.

Does anybody have more than 1 (better more than 2) Thunderbolt/GPSDO  or similar to test?

TIA


Giuseppe Marullo

IW2JWW - JN45RQ


On 1/5/2021 9:24 AM, Victor 4X6GP wrote:

This is interesting! My progrock plus qlg1 have been running 24/7 for a few days, and the error appears to be about 2 Hz, or slightly less.

I will try searching the forum.

Victor
On 05/01/2021 9:55, Alan G4ZFQ wrote:
the error when receiving WWV or RWM on 10 MHz is around 2-3 Hz. I thought I could improve this by setting the threshold to 1,

Victor,

I do not normally use a GPS with my Progrock.
But when I tried it I never saw it lock exactly, it was always about 2Hz off frequency.
There was a discussion about this then, maybe a few years ago. I do not remember a conclusion.

73 Alan G4ZFQ






Victor 4X6GP
 

To continue this thread, I have tried different values in register 3 for the past few days. The most recent experiment was setting the GPS threshold to 0. Several hours later the error was somewhat reduced, to between 1 and 2 Hz at 10 MHz, so I thought it was perhaps going to improve with time. But this morning, I am seeing between 2 and 3.5 Hz. I suspect the ambient temperature has something to do with it.


I found a thread from 2019, in which Hans wrote, "A GPS disciplined ProgRock should be accurate well within 0.05ppm (note that this parts-per-million specification means that the absolute precision in Hz, scales with frequency)." (https://groups.io/g/QRPLabs/message/36182).


This means that the 24 MHz signal should be "well within" 1.2 Hz of the nominal value. I am assuming that the SDR multiplies and divides this value to provide its local oscillator signal, so the local oscillator frequency should not be less accurate than this.


But the final determination of frequency of a received signal depends on the measurement of an audio frequency in the computer sound card, does it not? So could this not be the source of the remaining error?


Other methods that have been discussed here, like setting the progrock output to 1 kHz different from the standard frequency broadcast and then measuring the audio frequency of the beat with Spectran (etc.) would also depend on the sound card.


I think I will need to program the Progrock to 10 or 9.996 MHz, attenuate the output to match the received signal, and try to count the beats by ear!


Does this make sense or am I missing something?


Victor


On 04/01/2021 10:03, Victor Rosenthal 4X6GP wrote:

I'm trying to adjust the GPS threshold on my Progrock. It is programmed to produce 24 MHz which serves as an external clock for an SDRplay RSPdx SDR. The Progrock is connected to a QLG1.

With the default threshold of 5 Hz, the error when receiving WWV or RWM on 10 MHz is around 2-3 Hz. I thought I could improve this by setting the threshold to 1, which should adjust the 27 MHz clock in the Progrock to within 1 Hz. I did this by entering 031F with the Progrock's pushbutton.

But the result was that the error seems to be worse, around 3.5 Hz. Did I do something wrong? Should I have entered 0301F to get 1 Hz?
Victor


Alan G4ZFQ
 

I think I will need to program the Progrock to 10 or 9.996 MHz, attenuate the output to match the received signal, and try to count the beats by ear!
Victor,

Yes, that is one method I use.
I often use a GPS source https://sites.google.com/site/g4zfqradio/temp/u-blox_neo-6-7
And, not so much now a 10MHz GPS reference. They all agree within my comparison methods, better than 0.1Hz.

73 Alan G4ZFQ


geoff M0ORE
 

Can I suggest the following technique which does not involve the use of a computer soundcard.

Connect the reference signal to channel 1 of a dual beam scope and trigger to that channel. Connect the signal to be checked to channel 2. Adjust the Y gains so that the two traces are of nearly equal amplitude but this is not critical.

The frequency difference between the two signals will cause the trace on channel 2  to drift either to the left or right depending on whether the unknown signal is higher or lower in frequency.  You can count the number of cycles that the trace drifts past a fixed reference point such as the zero crossing point if a sine wave or the leading edge if a square wave.

This method does not require any measurement, just a direct comparison.


On 06/01/2021 07:05, Victor 4X6GP wrote:

To continue this thread, I have tried different values in register 3 for the past few days. The most recent experiment was setting the GPS threshold to 0. Several hours later the error was somewhat reduced, to between 1 and 2 Hz at 10 MHz, so I thought it was perhaps going to improve with time. But this morning, I am seeing between 2 and 3.5 Hz. I suspect the ambient temperature has something to do with it.


I found a thread from 2019, in which Hans wrote, "A GPS disciplined ProgRock should be accurate well within 0.05ppm (note that this parts-per-million specification means that the absolute precision in Hz, scales with frequency)." (https://groups.io/g/QRPLabs/message/36182).


This means that the 24 MHz signal should be "well within" 1.2 Hz of the nominal value. I am assuming that the SDR multiplies and divides this value to provide its local oscillator signal, so the local oscillator frequency should not be less accurate than this.


But the final determination of frequency of a received signal depends on the measurement of an audio frequency in the computer sound card, does it not? So could this not be the source of the remaining error?


Other methods that have been discussed here, like setting the progrock output to 1 kHz different from the standard frequency broadcast and then measuring the audio frequency of the beat with Spectran (etc.) would also depend on the sound card.


I think I will need to program the Progrock to 10 or 9.996 MHz, attenuate the output to match the received signal, and try to count the beats by ear!


Does this make sense or am I missing something?


Victor


On 04/01/2021 10:03, Victor Rosenthal 4X6GP wrote:
I'm trying to adjust the GPS threshold on my Progrock. It is programmed to produce 24 MHz which serves as an external clock for an SDRplay RSPdx SDR. The Progrock is connected to a QLG1.

With the default threshold of 5 Hz, the error when receiving WWV or RWM on 10 MHz is around 2-3 Hz. I thought I could improve this by setting the threshold to 1, which should adjust the 27 MHz clock in the Progrock to within 1 Hz. I did this by entering 031F with the Progrock's pushbutton.

But the result was that the error seems to be worse, around 3.5 Hz. Did I do something wrong? Should I have entered 0301F to get 1 Hz?
Victor


Alan G4ZFQ
 

Geoff

This method does not require any measurement, just a direct comparison.
Nor does a soundcard. It is just a comparison. As Victor says, count the beats by ear if it is close.
If not too close how do you estimate an error of several Hz using a scope?

73 Alan G4ZFQ


Victor 4X6GP
 

But what I'm trying to compare is the frequency of my progrock with the signal from a standard frequency station. How do I get that weak RF signal into my scope?

I could of course get an audio beat note. But how do I get a standard audio frequency to compare it with?

Victor
On 06/01/2021 12:13, geoff M0ORE via groups.io wrote:

Can I suggest the following technique which does not involve the use of a computer soundcard.

Connect the reference signal to channel 1 of a dual beam scope and trigger to that channel. Connect the signal to be checked to channel 2. Adjust the Y gains so that the two traces are of nearly equal amplitude but this is not critical.

The frequency difference between the two signals will cause the trace on channel 2  to drift either to the left or right depending on whether the unknown signal is higher or lower in frequency.  You can count the number of cycles that the trace drifts past a fixed reference point such as the zero crossing point if a sine wave or the leading edge if a square wave.

This method does not require any measurement, just a direct comparison.


On 06/01/2021 07:05, Victor 4X6GP wrote:

To continue this thread, I have tried different values in register 3 for the past few days. The most recent experiment was setting the GPS threshold to 0. Several hours later the error was somewhat reduced, to between 1 and 2 Hz at 10 MHz, so I thought it was perhaps going to improve with time. But this morning, I am seeing between 2 and 3.5 Hz. I suspect the ambient temperature has something to do with it.


I found a thread from 2019, in which Hans wrote, "A GPS disciplined ProgRock should be accurate well within 0.05ppm (note that this parts-per-million specification means that the absolute precision in Hz, scales with frequency)." (https://groups.io/g/QRPLabs/message/36182).


This means that the 24 MHz signal should be "well within" 1.2 Hz of the nominal value. I am assuming that the SDR multiplies and divides this value to provide its local oscillator signal, so the local oscillator frequency should not be less accurate than this.


But the final determination of frequency of a received signal depends on the measurement of an audio frequency in the computer sound card, does it not? So could this not be the source of the remaining error?


Other methods that have been discussed here, like setting the progrock output to 1 kHz different from the standard frequency broadcast and then measuring the audio frequency of the beat with Spectran (etc.) would also depend on the sound card.


I think I will need to program the Progrock to 10 or 9.996 MHz, attenuate the output to match the received signal, and try to count the beats by ear!


Does this make sense or am I missing something?


Victor


On 04/01/2021 10:03, Victor Rosenthal 4X6GP wrote:
I'm trying to adjust the GPS threshold on my Progrock. It is programmed to produce 24 MHz which serves as an external clock for an SDRplay RSPdx SDR. The Progrock is connected to a QLG1.

With the default threshold of 5 Hz, the error when receiving WWV or RWM on 10 MHz is around 2-3 Hz. I thought I could improve this by setting the threshold to 1, which should adjust the 27 MHz clock in the Progrock to within 1 Hz. I did this by entering 031F with the Progrock's pushbutton.

But the result was that the error seems to be worse, around 3.5 Hz. Did I do something wrong? Should I have entered 0301F to get 1 Hz?
Victor


Giuseppe Marullo[iw2jww]
 

Hi Victor,

But what I'm trying to compare is the frequency of my progrock with
the signal from a standard frequency station. How do I get that weak RF signal into my scope?
first of all, How do you know your radio is perfectly tuned to properly receive such signal?

I have a Trimble with Timing antenna(yes, you do need a specific antenna for timing gps) and I do use Lady Heater to check it, usually under 100-300ppt (parts per trillion), depending on how stable the room temperature is, survey done and usually at least a week operating time.

I use a Racal Dana 1998 FC (10/11 digits) with compensated oven, calibrated using the Trimble (and can use the Trimble as timebase).

Now, with this setup I should be fairly confident that the ProgRock (that is usually set at 10/125Mhz output) is 2Hz below 10000000). My equipment is old and such, but seems working.

Despite LH giving assurance that the Trimble precision is way better than needed, I could not be sure which one is to "blame".

I asked to the Timenuts ml gurus, and their answers was like I need at least three GPSDO to be sure (they are on another level of course).

The other question is for Hans:

- Which is the equipment used to verify 0.05ppm performance? I would expect at least 2-3 reliable sources else you could not be in any better position to "certifiy" than us.

BTW, I ordered a new firmware chip, just to sort out the idea that the calibration data could be wearing the eeprom due to keeping the ProgRock several months powered on, but even if set to 0 Hans told that is not writing data too often. Didn't have time to return on the subject but I am vey interested to have ProgRock and Trimble on the same page, that should be 0.05pps.

We really need at least one OM with at least two GPSDO to check this. Else I would have to send a ProgRock to some folks on the Timenut ml.

Giuseppe Marullo
IW2JWW - JN45RQ


geoff M0ORE
 

Alan, up to about 10Hz deviation should be easily seen on scope display any more than that and it would be difficult to count the audio beats.

IoW covid doesn't look good, please take care. I heard today (Wednesday ) that the island was "invaded" over the Christmas / New Year period by second homers from mainland.

On 06/01/2021 16:08, Alan G4ZFQ wrote:

Geoff

This method does not require any measurement, just a direct comparison.

Nor does a soundcard. It is just a comparison. As Victor says, count the beats by ear if it is close.
If not too close how do you estimate an error of several Hz using a scope?

73 Alan G4ZFQ







Mont Pierce KM6WT
 

On Tue, Jan 5, 2021 at 11:05 PM, Victor 4X6GP wrote:
To continue this thread, I have tried different values in register 3 for the past few days. The most recent experiment was setting the GPS threshold to 0. Several hours later the error was somewhat reduced, to between 1 and 2 Hz at 10 MHz, so I thought it was perhaps going to improve with time. But this morning, I am seeing between 2 and 3.5 Hz. I suspect the ambient temperature has something to do with it.
I have a few ProgRocks I've been playing with GPS calibrations.  The frequency varies quite a bit, probably due to room temp, and I have it sitting out in the open.

I bought some FOX TCXOs, at DigiKey:   https://www.digikey.com/en/products/detail/fox-electronics/FT5HNBPK27.0-T1/1024780

Worked really well.  Frequency calibration was just a few hertz.  And very stable.  Only needed to use the GPS calibration once after installing the TCXO.

BTW, I used the 27MHz TCXO, and not the 25MHz that Hans puts on the QCXTCXO boards.  Mainly just to avoid the mismatch when doing a Factory Reset...  Either Frequency TCXO should do.


73,
km6wt


Alan G4ZFQ
 

I bought some FOX TCXOs, at DigiKey: https://www.digikey.com/en/products/detail/fox-electronics/FT5HNBPK27.0-T1/1024780 >
Are these the ones with digital compensation?
There was a discussion some time ago about these, instead of a slow analogue variation they jump, often 1Hz steps. https://sites.google.com/site/g4zfqradio/incircuitprogramatmel168#h.pbu1b1vv7jm6

This can be a problem with narrow bandwidth modes like WSPR on higher frequencies.

73 Alan G4ZFQ


Alan G4ZFQ
 

Alan, up to about 10Hz deviation should be easily seen on scope display
Geoff,

OK, it's some time since I used that method. I find a SDR or Speclab display easier and gives sufficient accuracy.

I heard today (Wednesday ) that the island was "invaded" over the Christmas / New Year
That's what we expected. Could be why we are now on even terms with the rest. Apart from the fact there is only one hospital that does not involve ferry trips.

73 Alan G4ZFQ


Victor 4X6GP
 

I am continuing my experiments to try to get my Progrock plus QLG1 to provide the accuracy they are  supposed to.
In my latest test, I programmed it to 10 MHz and compared it directly to WWV. I can clearly hear the beats, and the result agrees with my SDR: it is off between 1 and 3 Hz. That is 0.3 ppm and I think it should be better.
Victor
On 06-Jan-2021 09:05, Victor 4X6GP via groups.io wrote:

To continue this thread, I have tried different values in register 3 for the past few days. The most recent experiment was setting the GPS threshold to 0. Several hours later the error was somewhat reduced, to between 1 and 2 Hz at 10 MHz, so I thought it was perhaps going to improve with time. But this morning, I am seeing between 2 and 3.5 Hz. I suspect the ambient temperature has something to do with it.


I found a thread from 2019, in which Hans wrote, "A GPS disciplined ProgRock should be accurate well within 0.05ppm (note that this parts-per-million specification means that the absolute precision in Hz, scales with frequency)." (https://groups.io/g/QRPLabs/message/36182).


This means that the 24 MHz signal should be "well within" 1.2 Hz of the nominal value. I am assuming that the SDR multiplies and divides this value to provide its local oscillator signal, so the local oscillator frequency should not be less accurate than this.


But the final determination of frequency of a received signal depends on the measurement of an audio frequency in the computer sound card, does it not? So could this not be the source of the remaining error?


Other methods that have been discussed here, like setting the progrock output to 1 kHz different from the standard frequency broadcast and then measuring the audio frequency of the beat with Spectran (etc.) would also depend on the sound card.


I think I will need to program the Progrock to 10 or 9.996 MHz, attenuate the output to match the received signal, and try to count the beats by ear!


Does this make sense or am I missing something?


Victor


On 04/01/2021 10:03, Victor Rosenthal 4X6GP wrote:
I'm trying to adjust the GPS threshold on my Progrock. It is programmed to produce 24 MHz which serves as an external clock for an SDRplay RSPdx SDR. The Progrock is connected to a QLG1.

With the default threshold of 5 Hz, the error when receiving WWV or RWM on 10 MHz is around 2-3 Hz. I thought I could improve this by setting the threshold to 1, which should adjust the 27 MHz clock in the Progrock to within 1 Hz. I did this by entering 031F with the Progrock's pushbutton.

But the result was that the error seems to be worse, around 3.5 Hz. Did I do something wrong? Should I have entered 0301F to get 1 Hz?
Victor


Victor 4X6GP
 

Here is a picture of the 1 pps pulse from my QLG1, connected to the GPS input of the Progrock:



Note the "grass" growing after the pulse. I believe that this is crosstalk from the GPS data that is also carried in the cable from the QLG1.


I wonder if this could be the cause of the poor accuracy of the Progrock's output frequency?


I would disconnect it at the GPS, but I need it for my U3S, which I also use the QLG1 with. I need to think about this.

Victor
On 08/01/2021 6:35, Victor 4X6GP via groups.io wrote:

I am continuing my experiments to try to get my Progrock plus QLG1 to provide the accuracy they are  supposed to.
In my latest test, I programmed it to 10 MHz and compared it directly to WWV. I can clearly hear the beats, and the result agrees with my SDR: it is off between 1 and 3 Hz. That is 0.3 ppm and I think it should be better.
Victor
On 06-Jan-2021 09:05, Victor 4X6GP via groups.io wrote:

To continue this thread, I have tried different values in register 3 for the past few days. The most recent experiment was setting the GPS threshold to 0. Several hours later the error was somewhat reduced, to between 1 and 2 Hz at 10 MHz, so I thought it was perhaps going to improve with time. But this morning, I am seeing between 2 and 3.5 Hz. I suspect the ambient temperature has something to do with it.


I found a thread from 2019, in which Hans wrote, "A GPS disciplined ProgRock should be accurate well within 0.05ppm (note that this parts-per-million specification means that the absolute precision in Hz, scales with frequency)." (https://groups.io/g/QRPLabs/message/36182).


This means that the 24 MHz signal should be "well within" 1.2 Hz of the nominal value. I am assuming that the SDR multiplies and divides this value to provide its local oscillator signal, so the local oscillator frequency should not be less accurate than this.


But the final determination of frequency of a received signal depends on the measurement of an audio frequency in the computer sound card, does it not? So could this not be the source of the remaining error?


Other methods that have been discussed here, like setting the progrock output to 1 kHz different from the standard frequency broadcast and then measuring the audio frequency of the beat with Spectran (etc.) would also depend on the sound card.


I think I will need to program the Progrock to 10 or 9.996 MHz, attenuate the output to match the received signal, and try to count the beats by ear!


Does this make sense or am I missing something?


Victor


On 04/01/2021 10:03, Victor Rosenthal 4X6GP wrote:
I'm trying to adjust the GPS threshold on my Progrock. It is programmed to produce 24 MHz which serves as an external clock for an SDRplay RSPdx SDR. The Progrock is connected to a QLG1.

With the default threshold of 5 Hz, the error when receiving WWV or RWM on 10 MHz is around 2-3 Hz. I thought I could improve this by setting the threshold to 1, which should adjust the 27 MHz clock in the Progrock to within 1 Hz. I did this by entering 031F with the Progrock's pushbutton.

But the result was that the error seems to be worse, around 3.5 Hz. Did I do something wrong? Should I have entered 0301F to get 1 Hz?
Victor


Alan G4ZFQ
 

I would disconnect it at the GPS, but I need it for my U3S, which I also use the QLG1 with. I need to think about this.
Victor,

Have you checked the accuracy of the U3S?
I think WSPR TX is still 2Hz out but I think that is for a different reason.
I guess other modes on the U3S should be accurate.

73 Alan G4ZFQ


Victor 4X6GP
 

I have used the U3S on WSPR with 200 mW and got good reports worldwide on 30, 40 and 15m. By adjusting the parking frequency I have got consistent drift reports of zero and 1. So it works well.

Does the precise output frequency matter if it is within the band?

Victor
On 09/01/2021 11:37, Alan G4ZFQ wrote:

I would disconnect it at the GPS, but I need it for my U3S, which I also use the QLG1 with. I need to think about this.

Victor,

Have you checked the accuracy of the U3S?
I think WSPR TX is still 2Hz out but I think that is for a different reason.
I guess other modes on the U3S should be accurate.

73 Alan G4ZFQ