[time-nuts] Adapting my GPSDO to the FE-5680A

[Nick Sayer]

> On Apr 22, 2016, at 1:58 AM, GandalfG8--- via time-nuts <time-nuts at febo.com> wrote:
> 
> Thanks. I’ve taken your suggestion for the sine-to-square converter.
> I  believe there are two separate commands for tuning the 5680 - one is “
> temporary”  and one writes through to the EEPROM. I’ll be using the latter, 
> of course.
> 
> ---------------
> I'm surprised nobody else seems to have  commented on this but I'm  pretty 
> sure that using the write to EEPROM option on an automatically  and 
> regularly repeated basis could be somewhat akin to applying the  kiss of death.

I mis-typed. I meant to say that I’d be using the *former* - the temporary command. No writes to EEPROM. I do have code in all of my current GPSDO code that skips the DAC/serial write if the value isn’t changing. I did that for the DAC because there’s a change glitch that’s good to avoid if it’s not absolutely necessary.

Anyway, I’ve ordered the first set of boards. In addition to Atilla’s suggestion for the oscillator output conditioning (the DC-block + self-biased inverter), I’ve also broken out the oscillator’s PPS pin to a connector, in case folks might want to experimentally measure it against the GPS PPS or whatever other use one might desire. The physical form factor is designed to be the same width as the FE-5680A, with the connector spaced so that it looks like an extension of it. If you put, oh, 1/3” feet on the four corners of the board, it ought to sit flat adjacent to the oscillator and, well, I’m hoping it looks rather handsome. :) The connectors are the same as the crystal based GPSDOs - a mini-DIN-4 diagnostic connector with the same pinout, an SMA GPS antenna connector, 2.1mm power jack - but this one takes 16-24 VDC @ 25-30W (easily obtainable from a surplus laptop power supply), and two .1” JST output jacks that get separately buffered CMOS square wave outputs (NB3N551 + 33Ω series resistor). The boards should get back early May and I’ll post reports to the Hackaday.io project. My results may not be very useful, as I don’t have a significantly better reference for comparison. My control results just sort of look like I’m really measuring the Thunderbolt against the 5680A rather than vice-versa. I suppose I could attempt to buy a second 5680A and square the two off against each other, but buying 5680As is fraught with danger, as there are significant feature and pinout differences that are not always properly called out in eBay auction descriptions.