bg7tbl gpsdo master reference

[Ohm_My]

PS, what's a typical timeframe for the Self Survey to complete upon first startup?

The site survey should complete in about 24 hours. No provisions to change the interval.

[Fennec]

Fennec, just to confirm, you found the M8T to be both a direct replacement and better performance than the M7?

Yes, simple because my M7N was a fake modul and the M8T can handle rtk raw data to play around with SAPOS.
For a simple upgrade the M8N is good enough. It's way more stable than the M7N.
My next steps are the ublox ZED-F9P and change the cheap internal  ublox M8T TCXO to an OCXO.
https://www.u-blox.com/en/product/zed-f9p-module
Just to play arround with.... I do not really need that accuracy, but's interesting what you can reach with those "cheap" hardware.

[bugi]

... change the cheap internal  ublox M8T TCXO to an OCXO.

Hmm.. been a few weeks since I read about those... Seems those modules have an option for external clock input so one doesn't have to replace that internal clock, just input the OCXO into it (divided down enough). Then again, I'm not sure, the replacement might give better results than just the external clock input.
(See M8T datasheet 1.15.2 "Aiding" with EXTINT.)

[Electro Fan]

PS, what's a typical timeframe for the Self Survey to complete upon first startup?

The site survey should complete in about 24 hours. No provisions to change the interval.

On my Star 4+ it seems to complete the site survey in less than 24 hours - maybe closer to 6-12 hours?  I haven't kept very good track of the time but it's been started and then restarted two more times in the first day and it's completed the survey twice, and is now on it's third startup survey.   I realize it could be somewhat warm from prior starts but even the very first survey completed in well less than a day.  Not sure why.

Lots of other questions developing.....
(I've read and searched the LH manual but so far haven't found the answers.)

1. Is there a way to display the Star 4+ 10 MHz signal in LH?

2.  Doppler looks ver (completely) flat – is this normal?  (Kinda expected to see some variation over time.)

3. What does holdover perf POOR mean?  (What causes POOR? How can it be improved?)

4. Dly is -77ns, that’s a default set in the unit by the seller for the supplied antenna cable length, or something else?

5. Discipline mode is Normal - is that good or is there something better than Normal?

6. When the survey completes the plots for lat, long, and altitude in the ADEV graph disappear; the values remain displayed above the graph area but the plots disappear; GV toggles the values on and off above the graph area but not the graph plots.  Kind of strange.

7. When the survey completes DAMP shows 0.000, GAIN shows -3.500 Hz/V, INIT shows 0.000 V; they are all grayed out.... good/not so good?

8. When the survey completes TC says 200.0 sec - is that good/bad/something that needs adjustment?

Thanks for any info/tips.

[kj7e]

It seems the Timing specific module may perform better for use as a GPSDO rather than precision positioning module.  The phase comparison for the Disciplined Oscillator comes from the 1pps timing pulse.   From what I have been reading, the Timing specific module employ extra features to minimize time pulse jitter.

I'm still poring over the M8T's datasheet and reference information and noted some settings I want to play with when I get home later.

[jpb]

3. What does holdover perf POOR mean?  (What causes POOR? How can it be improved?)

I think that this is just a reflection of the amount of data it has received. Assuming you have a good signal the holdover performance should move onto being "GOOD" at least that is what mine show.

It is just a measure of how good it thinks its learned (Kalman Filter?) parameters are.

Personally I'm not that worried about holdover performance as I've never known the GPS system stop transmitting and if there were a temporary blip I'd just wait for good signal again.

[Ohm_My]

4. Dly is -77ns, that’s a default set in the unit by the seller for the supplied antenna cable length, or something else?

For precision timing you would want to compute the antenna cable delay for your particular setup based on the length of coax between your  GPS antenna and the receiver.

There are several online calculators available to assist you in doing this for both RF delay and signal loss.

[Electro Fan]

4. Dly is -77ns, that’s a default set in the unit by the seller for the supplied antenna cable length, or something else?

For precision timing you would want to compute the antenna cable delay for your particular setup based on the length of coax between your  GPS antenna and the receiver.

There are several online calculators available to assist you in doing this for both RF delay and signal loss.

Thanks.  That's kind of what I was thinking but I was wondering what would have set the current setting to -77 ns.  Maybe it is a default that comes up when you set LH for /rx4 (Oscilloquartz Star-4).  In any event, I followed your advice and found a Times Microwave calculator and entered RG174, ~5 meters (~16.4'), and I tried both the L1 and L2 frequencies (not sure what was intended there so I tried both as shown in the attachments below).  The L1 freq calc gave a warning msg saying the freq exceeds the max tested frequency; both calcs showed the same result:  ~25 ns, so I've adjusted that in Lady Heather (using the minus sign as per the LH documentation for Oscilloquartz Star).  Of course it's pretty difficult to tell if the new delay value is having any impact on the 10 MHz reference signal, or anything else.  But it's fun to try


- I guess that makes sense.  I think light travels about 30 cm or about 1 foot in a ns, so 25 nanoseconds would be 25' but with a .66 Velocity Factor we'd be back to ~16.4 feet.

[kj7e]

The delay calculations are only needed for absolute phase alignment and will have no bearing on the 10MHz accuracy or stability.

[Ohm_My]

The warning about the max tested frequency is due to the fact that RG174 coax is rated for frequencies up to 1,000 MHz. Beyond that the signal attenuation is greatly amplified as the frequency increases.

RG174 or equivalent is offered up due to the cheap price per foot. You might would want to consider upgrading your coax as this will greatly benefit SNR an d C/N0 signal levels at the L1 portion of the spectrum for satellite acquisition / tracking.

[Electro Fan]

The delay calculations are only needed for absolute phase alignment and will have no bearing on the 10MHz accuracy or stability.

Yep, it would seem that phase is the next mountain after frequency.  It would be cool to be able to watch the phase shift as the values get entered into LH but I'm guessing there is some system level propagation delay, and I don't know that any of my gear would reflect the phase shift accurately - but it's fun to contemplate.

[Electro Fan]

The warning about the max tested frequency is due to the fact that RG174 coax is rated for frequencies up to 1,000 MHz. Beyond that the signal attenuation is greatly amplified as the frequency increases.

RG174 or equivalent is offered up due to the cheap price per foot. You might would want to consider upgrading your coax as this will greatly benefit SNR an d C/N0 signal levels at the L1 portion of the spectrum for satellite acquisition / tracking.

Roger that.  The next step might be to upgrade the antenna with a better antenna - and an antenna with detachable coax, so the coax could be easily changed out.  There's a lot of performance improvement chasing to be done with this amplitude, frequency, and phase nut stuff

[kj7e]

My HP 53132A will calculate phase angle to 0.1deg between inputs A and B.  But phase alignment is only needed for absolute timing or phase sync of the reference signal.  Fun to play with but phase is only important if your have something to compare it against.  The 10MHz output will be 10Mhz at all 360deg of phase angle with no loss in accuracy.

[Electro Fan]

The delay calculations are only needed for absolute phase alignment and will have no bearing on the 10MHz accuracy or stability.

Just to be sure I understand, at some point, if the frequency is very close but the phase is off by more than 360 degrees the frequency would have to be (at least) one unit off, right? 

So for example, at 10,000,000.1 Hz if the phase is off 180 degrees, the frequency might be 10,000,000.05 Hz or 10,000,000.15 Hz?

[kj7e]

No, at 360 or 180 deg of phase shift, between lets say a frequency counters reference input and the measured signal, the count will still be perfect as the number of waves are still the same.  With older counters that count the waves and use the ref as the gate, phase shifts other than 360 or 180deg could cause an error of one count.  Modern reciprocal counters like the 53132a, the phase of the test signal has no bearing on the measurement as they don't count the number of waves, but measures the period then inverts to display the frequency.  So phase has no effect of the period of the measured signal compared to the reference.

[Electro Fan]

No, at 360 or 180 deg of phase shift, between lets say a frequency counters reference input and the measured signal, the count will still be perfect as the number of waves are still the same.  With older counters that count the waves and use the ref as the gate, phase shifts other than 360 or 180deg could cause an error of one count.  Modern counters like the 53132a, the phase of the test signal has no bearing on the measurement as they don't count the number of waves, but measures the period then inverts to display the frequency.  So phase has no effect of the period of the measured signal compared to the reference.

Roger
How about 361 degrees off?  Will that cause a modern counter to miss by 1 count (wave)?

[kj7e]

361 deg = +1 deg, only 360 deg in a circle.  It could, but + or - 1 count is always a given with any counter.

[Electro Fan]

Just seems like basically the counter is counting waves and then rounding to the nearest wave (not sure yet how it rounds up or down but I'll let that go).  But at some point an extra wave (more than 360 degrees past the previous wave) is going to have to be considered an extra count.  This might need a glass of wine or a beer.

[texaspyro]

The purpose of the cable delay parameter is to align the PPS output to UTC by compensating for the signal delay in the antenna cable between the antenna and the receiver input.   

Some receivers also have a way of compensating for the delay in the PPS output cable (usually the PO command in Heather).  Some even have ways of compensating for the delay in the receiver RF circuitry. 

Heather's cable delay command lets you specify the delay in nanoseconds (like -77)  or feet  (-50f) or meters (-15m).  You can also specify the velocity factor (like -50f 0.66v).

[Electro Fan]

Just bumping this back into the thread in case anyone has some of the answers - Thx

Lots of other questions developing.....
(I've read and searched the LH manual but so far haven't found the answers.)

1. Is there a way to display the Star 4+ 10 MHz signal in LH?

2.  Doppler looks ver (completely) flat – is this normal?  (Kinda expected to see some variation over time.)

3. What does holdover perf POOR mean?  (What causes POOR? How can it be improved?)

4. Dly is -77ns, that’s a default set in the unit by the seller for the supplied antenna cable length, or something else?

5. Discipline mode is Normal - is that good or is there something better than Normal?

6. When the survey completes the plots for lat, long, and altitude in the ADEV graph disappear; the values remain displayed above the graph area but the plots disappear; GV toggles the values on and off above the graph area but not the graph plots.  Kind of strange.

7. When the survey completes DAMP shows 0.000, GAIN shows -3.500 Hz/V, INIT shows 0.000 V; they are all grayed out.... good/not so good?

8. When the survey completes TC says 200.0 sec - is that good/bad/something that needs adjustment?

9. FW says 0149... any chance that's firmware?  If so, is that current?  What are the chances it can be updated from somewhere?

Thanks for any info/tips.

[N4KPI]

Got a beginner question. I want to get one of these and looking on EBay most of them show the 10 MHz output as for example “10M OUTPUT :SINE WAVE,1Vrms (10-15dBm),square wave,3.3/4.7Vpp”. This suggests it is selectable. Is that true and how do you set it, or is it a hardware option?

What model (date I guess) is best?

Thanks all

[usagi]

It's not settable. You have to buy the model with the fixed output you want.

[kj7e]

It's not settable. You have to buy the model with the fixed output you want.

The output of the CPLD is a square wave, the sine wave is via a low pass filter.  You could just jumper the filter for a square wave if you wanted.

[kj7e]

Just wanted to add, finally getting the ublox NEO-M8T configured, enabled Galileo and tracking 20+ sats now. After a few other configuration changes I can say it seems to lock down to 0.00ppb very quickly and is rock solid.  Still sorting through the 400 page book on this chip.

[Fennec]

Still sorting through the 400 page book on this chip.

lol, ya, me too :o)

....finally getting the ublox NEO-M8T configured, enabled Galileo and tracking 20+ sats now.

You just need to set the NMEA Protocoll / Version to 4.1, than you can use all 72 SATs and GPS, Gallileo and GLONASS together.

and check out this suite:
http://www.simonsdialogs.com/category/electronics/gpsdo-10-mhz/