Open source GPS-disciplined frequency generator.

[rcls]

Hi,

I made an open hardware & open source GPS disciplined frequency generator with output frequencies from 1Hz to 1.6GHz.  https://github.com/rcls/gpsfreak

The device outputs four separately programmable frequencies (from 2 PLLs) onto SMA connectors, with GPS used as the source for a reference time base.

The clock generator is a TI LMK05318b bulk-acoustic wave oscillator.  This device is pretty flexible and has very low phase noise, in the same ball-park as a decent crystal oscillator (I don't have equipment capable of measuring it).

Short term frequency accuracy is determined by GPS and should be under 1ppb.  It can be built with either a U-Blox MAX-M10S (low cost) or MAX-F10S (dual
frequency) GPS module.

There is firmware, plus python scripts for controlling the device from a PC. (Configurations can be saved to flash for disconnected operation).

Manufacturing with JLCPCB and hand-soldering select components (connectors and the more obscure passives), you can build two units at approx US $50 each.

Cheers,
Ralph.

[DimitriP]

I like a grand entrance!
Welcome to the forum

[5U4GB]

Nice, and nice detailed writeup!  Have you thought about submitting it to Silicon Chip as a project?

[rcls]

Thanks for the suggestion, I'm looking into it.

[5U4GB]

Just realised my note was a bit ambiguous, I meant it'd quality as a full-blown article, not just their reader-contributed circuit ideas at the back of the issue.

[SCSKITS]

Would be nice to have a PDF in the GITHUB of the schematic and maybe the board layout.

[rcls]

Hi,

I have added PDFs to the git repo:
https://github.com/rcls/gpsfreak/blob/main/gpsfreak-sch.pdf
https://github.com/rcls/gpsfreak/blob/main/gpsfreak-pcb.pdf

Cheers,
Ralph.

[dobsonr741]

Nice. Datasheet of LMK05318b say the max output frequency is 800MHz. How did you squeeze 2x of that? How’s the spectrum look like?

[Cyclotron]

This looks very interesting. I'll add it to my project list. Have you posted it to the time-nuts mailing list?
http://leapsecond.com/time-nuts.htm

[rcls]

The TI docs are vague and contradictory on the max frequency. The web page sez 800Mhz. The data sheet sez 1250MHz is the max output frequency, but Footnote 5 on the datasheet "electrical characteristics" says:

An output frequency over fOUT max spec is possible, but output swing can be less than VOD min specification.

On "what does the spectrum look like", it looks good on my oscilloscope but the reality is I don't have equipment capable of characterising it. Although I can confirm that if you set the output frequency to 1575.42MHz and attach an appropriate piece of wire to an output then nearby GPS units notice the interference