GPS/Galileo (or both) for GPSDO use?

[Solder_Junkie]

I am building a GPSDO based on the Arduino pulse counting controller idea by Lars Walenius and further refined by Murray Greenman. There is a lengthy discussion on the projects part of this forum. Unfortunately Lars and Murray are no longer around to ask.

I have already found, as have others, that a proper “timing” GPS module (Ublox Neo M8T) gives noticeably better results than a fake Ublox Neo 6. In a simple GPSDO the overall frequency stability (jitter) is around half the level when using the M8T in fixed position timing mode.

I have been using both GPS and Galileo satellites together. For best frequency accuracy from a GPS disciplined oscillator, using long term counting/correction techniques, is there any advantage in sticking with GPS, Galileo, or using both? In theory there can the a few nS time difference between the constellations and if the receiver swaps between them it may introduce some jitter.

The data sheet for the M8T says to not use SBAS position corrections but suggests contacting their regional office over the choice of satellite networks.

The only GPS based time server manuals I can find seem to only refer to “GPS” and cable length between the antenna and receiver.

Thanks for any advice.

SJ

[thinkfat]

I have found that using GPS and Galileo together helps with SV coverage when you don't have access to a roof with full sky view, while at the same time not degrading the timing stability. But also not improving it.

I also found that enabling GLONASS has a major negative impact on timing stability. Weird phase jumps, overall less stable. Might be a firmware issue with the M8T, but that is what I have.

SBAS - same. However, SBAS is quite helpful during the survey-in, because it will allow the true position to converge faster. Therefore, in my own GPSDO software, I will switch on SBAS when starting a survey-in and also increase the measurement frequency to 4Hz, which is something the M8T can easily handle. Once the survey-in is complete, I switch off SBAS again and also go back to 1Hz measurement frequency.

Hope this helps.

[Solder_Junkie]

Thanks for the reply. I know the position of the antenna to within 2m or so. Running fixed mode with GPS and Galileo at the moment. I have an outdoor antenna with a pretty good sky view.

The Lars/Murray Greenman GPSDO is locked and working, I just need to fine tune the settings.

Regards
SJ

[Solder_Junkie]

In answer to my own question, just as thinkfat found, with a good sky view there is no advantage using GPS+Galileo over GPS or Galileo alone. I haven't bothered with the Chinese or Russian SVs. I am using a Ublox Neo-M8T. I have set the M8T for fixed mode and I am not using SBAS.

SJ

[olepr01@gmail.com]

Without knowing how you have come to the conclusion that GAL adds nothing, from a purely theoretical point of view it *should* - for two reasons: 1) GAL has more stable clocks on board the SV's, but perhaps more importantly 2) it broadcasts ephemerides more often - meaning the receiver does not need to extrapolate the satellite clock models so far into the future. YMMV, I guess.

[Solder_Junkie]

Without knowing how you have come to the conclusion that GAL adds nothing

I am using a GPSDO that averages out the 1 second GPS pulses, and oscillator frequency, over long timescales. Perhaps if it were a time reference, as opposed to a frequency reference, there may be some advantage in using both sets of satellites.

As far as I can determine higher accuracy units than my home made one, have in the past only used GPS.

My home made Murray Greenman (ZL1BPU) GPSDO produces results in the region of no more than +/- 2 parts in 10^11 frequency difference, measured against a rubidium oscillator over any time span.

Details of the GPSDO are on this web site, and were discussed at great length on the "Projects, Designs, and Technical Stuff" part of this forum. https://www.qsl.net/zl1bpu/PROJ/NGPSDO/New%20GPSDO.htm

SJ

[MIS42N]

Do a survey-in for several hours for each of the three options you give (GPS, Galileo, both) and use the one with the best error figure. If you've got a good sky view, it will likely give a figure 2m or better for one of GPS or Galileo. With a poor sky view it can be more than 2m and sometimes that gives advantage to choosing both.

[kj7e]

I use both, one GPSDO on GNSS (Galileo) and another GPSDO on GPS and compare between them.  If they both agree, then I know everything is right in the world.

The BG7TBL unit is one I modified and installed a genuine u-blox M8T and a OSCILLOQUARTZ 8663 OCXO.  Its setup for GNSS Galileo.  The ZYT is a Trimble Thunderbolt on GPS.

Both are run though a GPS signal splitter (DC pass on the Trimble side) and fed from an outdoor PCTEL GNSS Timing antenna;
https://www.pctel.com/antenna-product/global-gnss-timing-reference-antenna/

[radar_macgyver]

I'm using a Jackson Labs Micro-JLT GPSDO with an M8T receiver, their docs recommend not enabling SBAS. From their user guide:

As of November 2019, JLT received multiple reports on SBAS satellites causing timing issue on the GPSDO units around the world. Until further notice, JLT recommends disabling the SBAS in GPS:SYST:SEL command to avoid potential timing issue or monitoring the timing stability of Micro-JLT GNSS™ unit if SBAS is desired.

Not sure if it's a Micro-JLT firmware thing or related to the M8T itself.

GLONASS uses different frequencies, so you should verify that your antenna phase center doesn't move too much with frequency, and also that it's specified to receive it. As @thinkfat reported, enabling it causes the estimated frequency error to increase (tested with a PCTEL compact antenna, and a Trimble GA810)

[thinkfat]

It is related to the M8T. Ublox even mentions it in their manual for the receiver.

[Solder_Junkie]

SBAS is not needed for fixed use as you can "survey in" your fixed location initially and then turn SBAS off.

There have been studies into the timing accuracy of all 4 GNSS systems and both GPS and Galileo offer better timing than BDS or Glonass. The link below is typical:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9002469/table/sensors-22-02486-t003/?report=objectonly

My own measurements were against my own rubidium standard over typically 30 minute periods of plotting the frequency "jitter" of the GPSDO, which as stated earlier showed no difference between GPS + Galileo, or just one of the systems. My GPS module is a genuine M8T with SBAS off.

SJ