Have GPSDO - Now What?

[GigaJoe]

hm , same things i made, with my own code , and simplified

direct count up to 400 sec (32 bit)
and ocxo control over 12bit dac ,  vref - lt1021, and sum. ampl that   summing from multiturn resistor and dac.
using multiturn -do adjust coarse, when dac set in the middle;  then dac\code do fine adjustment.

no learning , just count difference and proportional steps in adjustment,  as dac change fraction of output, it about 0.5mV step. seems sufficient.

stunning prototype
napkins left - ocsx and voltage thingi
foil right - ublox clone
3 colour led -  switching in blue in a few hours, means calculated and expected counts for 400 sec  timeframe match , seems sufficient for me.

[MIS42N]

Sounds nifty!

If you think of it, please drop a link to your Github repo.

Thanks,

Kirk, NT0Z

Not fleshed out. I hope people will read and say 'what about ....' and I'll add more.
https://github.com/ajcashin/budget-gpsdo

[tkamiya]

If your GPSDO is in 3D locked status, it's accuracy and precision are order of magnitude better than one can compare against WWV and WWVB over an airwave.  A benefit of GPSDO is its self-calibrating nature.

[MIS42N]

If your GPSDO is in 3D locked status, it's accuracy and precision are order of magnitude better than one can compare against WWV and WWVB over an airwave.  A benefit of GPSDO is its self-calibrating nature.

I think this is an assumption that people make, and in most cases it's true. But I think the OP's point was how do you know. If you have marginal GPS signal, the GPS module usually gets a lock but the 1 pulse per second that most GPSDOs rely on can wander ±150ns causing the instantaneous frequency of a PLL oscillator to wander. Which is OK if you are making period measurements as the long term frequency averages out to be what you want. But if you are using the GPSDO to provide a reference to a transverter for a 100GHz receiver (and I believe there are people who do this) then those second by second deviations are unacceptable.

What happens if some failure causes your lock light to come on (a software bug?) even though there isn't a lock? Does the GPS examine the NMEA messages of the GPS module to see if it is still receiving valid GPS signals or is the module "guessing" when to output a 1PPS (some do). Just because the oscillator is locked to the 1PPS doesn't mean the 1PPS is valid.

About the only way to check most GPSDOs is to compare with something else. Which sort of defeats the purpose, because if you have a 'something else' you'd probably use it.

[tkamiya]

You are talking about quite an edge case.

I've compared mine to Cesium (luckily, I have one in my lab) and except for minor jitters, I really didn't find anything unusual.  I believe as long as you'd stick with proven commercial products, you'd be fine.  (Thunderbolt and such)