[time-nuts] WTB: GPSDO

Chris Albertson albertson.chris at gmail.com
Wed Mar 15 00:25:19 EDT 2017


Yes, that is a problem.  How to test a 10 MHz oscillator of any kind
(GPSDO or not) if you only have one oscillator.

I call it the "bootstrap problem"  and I think it's worth some
discussion.  The easy way is to buy a few known good clocks before you
set out to make a clock but that is the "easy way".   that

It is not hard to test its long term stability.  Simply count how may
cycles your new device makes in (say) 10 seconds.   Now all you need
is a way to measure 10 seconds.    GPS receivers are good at this.
For short term stability you have yo have some faith.  Buy a $25
oversized crystal oscillator that was made by the name brand with a
decent reputation and it will likely be reasonable

All the GPSDO does is adjust the electronic frequency control of that
$25 eBay oscillator until there are EXACTLY 10,000,000 counts between
every 1 Hz pulse.   The GPSDO is self testing.  Every second it
decides if that $25 crystal is running to fast or to slow and then
does something.   One of the things it should do it tell its owner if
the crystal is reusing much to fast or slow or if it is just right.
You will know the GPSDO has decent long term stability because the
green :just right" LED stays on all the time.

You can also verify the GPSDO is working correctly by looking at the
EFC voltage.  That is the control signal the controller sends to that
$25 eBay crystal.  You'd expect that the ESFC volts follows some kind
of theory and tracks the environment and you'd not expect a random

In short you can verify the correct operation of a home build GPSDO
because you can see inside of it.  When you see the inside parts are
working as expected you can have high confidence it is all working
well.     But for proof you need at least to more known-good
oscillators and you can measure beat frequencies between them.



On Tue, Mar 14, 2017 at 3:33 PM, Tim Lister <listertim at gmail.com> wrote:
> On Fri, Mar 10, 2017 at 2:35 PM, Chris Albertson
> <albertson.chris at gmail.com> wrote:
>> A GPSDO is not hard to make.  All you need is some way to compare the
>> phase of two signals, an XOR gate can do that.  Then a small $2
>> process moves the control voltage on the crystal.    I tried one to
>> build the simplest GPSDO that could still work.   Got the parts count
>> down to about four or five and the cost well under $10 plus the OXO
>> which was about $20.  The simplest dumb one I could make keeps about
>> e-10.  Not great but enough for many uses.   I compared to my
>> Thunderbolt and I could see the phase advance and retreat.  Just a
>> little most sophistication and I likely could do much better but my
>> goal was to prove to myself that a GPSDO could be build VERY simply
>> with cheap parts
>>
>
> Hi Chris, that's good news that a GPSDO is that easy to make (at least
> a basic one) as that is exactly my medium term plan ! The issue of
> course is having something to test the newly built GPSDO against... I
> got one of the rehoused Trimble UCCM-based GPSDOs off ebay a while ago
> but haven't been super happy with it. It's quite a bit less sensitive
> than more modern GPS receivers and it often struggles to get even 1
> satellite with the indoor patch antenna. At one point both red alarm
> LEDs came on and stayed on despite power cycles - I eventually fixed
> that by taking it apart and finding and hitting a reset button on the
> board. Currently although I can talk to the unit over serial and it
> seems to respond, Lady Heather is not seeing any output from it.
>
> Combined these things don't give me a great deal of confidence that
> this unit will act as a stable master reference. I was wondering if a
> second GPSDO like Russ linked to would work better (I have a ublox
> LEA-6T GPS already which I plan to use as the basis of the homebuilt
> GPSDO and it consistently sees many more satellites than the UCCM
> with a similar indoor antenna)  or put the money to getting an outdoor
> antenna mounted (don't feel happy drilling holes in the house myself)
> by someone. Do 2 GPSDOs tell you much more or just that each is
> different and you need a third to adjudicate ? (I can see a slippery
> slope looming from here...)
>
> Cheers,
> Tim
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.



-- 

Chris Albertson
Redondo Beach, California


More information about the time-nuts mailing list