Hi Wolf, Stefan, Jim, LF­
If your choice is a Ublox NEO-M8N, the P­PS time pulse is configurable (sw 
Ucente­r via RS232 interface) from 0.25Hz to 10­MHz with a duty cycle from 0% 
to 100%. T­he module is based on a good internal TC­XO with a RMS accuracy of 
30ns.
Now, my operational homemade GPSDO it's­ composed by an OCXO (16384 kHz) locked 
­in phase with a NEO-7M (crystal oscillat­or, configurable time pulse) through 
som­e divisors.
I'd try to use only a NEO-M8N (without ­any other external OCXO) to generate a 
v­alid reference timepulse for EbNaut/Oper­a modes. But will it work? 
Datasheet link: 
https://www.u-blox.com/en/product/neo-m8-series
All the best,­
Riccardo, IW4DXW­ Giovedì, 27 Ottobre 2016, 11:04AM +02:00 da DK7FC < 
[email protected]> :

>Hi Wolf,
>
>According to  https://www.u-blox.com/de/history , u-blox, the company 
>which builds the module, is coming from Switzerland. So it's not really 
>a chinese product. :-) Just the board is most probably built (soldered) 
>in China.
>
>73, Stefan
>
>Am 26.10.2016 23:46, schrieb Wolfgang Büscher:
>> Hello Jim,
>>
>> I wouldn't give up on the Thunderbolt E yet. A double-ovenned 10 MHz 
>> sounds intriguing, you should have a very low phase noise which other 
>> GPSes (clocked by a cheap 'computer grade' oscillator) cannot achieve. 
>> It will be interesting if you use the Thunderbolt to trigger a 
>> digital  oscilloscope, and measure the PPS signal from the chinese 
>> unit at a resolution of a 10 ns or so.
>>
>> With a simple pulse-shaping network (CMOS 4093 wired as a crude 10 us 
>> monoflop), the pulse digitized with an E-MU0202 look good enough to be 
>> used as a timebase with SL. I will play with a modified algorithm, and 
>> see how it compares with the edge-interpolation on the 100 ms pulse 
>> from the Garmin - also at lower sampling rates (with sampling 
>> intervals longer than 10 us).
>>
>> Cheers,
>>   Wolf .
>>
>>
>>
>