[time-nuts] GPSDO - was - oscillator choice question
lists at rtty.us
Mon May 17 00:40:09 UTC 2010
The gotcha is that the Z38xx and TBolt do a pretty good job.
To do as well:
1) Build a counter with a < 1 ns resolution to compare the pps ticks. You want the resolution to be below the floor of a good timing receiver.
2) Set up a roughly 20 bit D/A to drive the EFC. More bits are always better. You can get away with 16 bits and tweaking fairly often. The objective is a LSB in the parts in 10^-13.
3) Put a good reference on the D/A. Voltage tuning on the EFC makes this tough to get around.
4) Set up a PIC or micro of your choice (Coldfire works quite well). You can get a *lot* of processor for $5 these days.
5) Probably lay out a pcb to put everything on.
Hook that all up to a good modern high sensitivity 12 channel timing grade GPS and start tuning your loop. You will need a good local reference to really get the tuning right.
This is in addition to the usual divide to 1 pps on the 10811, buffering of outputs, power supplies and regulation, and packaging.
Software wise, most people seem to wind up with some sort of software PID, possibly with coefficients that change as the unit stabilizes (Z38xx family) or not (TBolt ?).
Lots of work to get to "as good as". Of course you will be as good as the combined performance of the two, which is indeed better than either one by it's self.
On May 16, 2010, at 5:30 PM, chris at yipyap.com wrote:
> Thanks for all of your kind encouragement.
> I gave up on using the NC100's oscillator for my GPSDO home-lab freq standard. I was able to find a 10811 oscillator on ebay by
> purchasing a non-working 5328A counter with the correct option.
> It was very inexpensive and the 10811 appears to be alive.
> Now I am reading Time-nuts history and I see that
> there is no consensus on which of the popular/published GPSDO
> configurations is the best return on effort and expense.
> That is rather disappointing.
> Any comments on that before I just flip a coin?
> Bob Camp wrote:
>> Microwave radio suggests low spurs and often low phase noise. That's on top of the stability requirements.
>> In terms of stability a good Z3801 will run rings around a garden variety Thunderbolt. Both are significantly better on phase noise than a "garden variety" OCXO. You will need a *very* good OCXO to get to the performance level of either one. The 10811 is probably the easiest thing to both identify and find. You see them selling anywhere from $100 to $200 depending on just how patient you are.
>> On May 2, 2010, at 4:13 PM, chris at yipyap.com wrote:
>>> Bob asks a reasonable question:
>>>> What kind of accuracy are you trying to obtain?
>>> I would like to end up with something that is usable
>>> on a home electronics workbench. Something like a
>>> Z3801. Something I can use if I ever try to do some
>>> goofing around with microwave radio operation.
>>> I'm not serious enough to want to spend hundreds
>>> on a working unit. On the other hand, lashing up
>>> something that works would count toward my
>>> hobby/education so then it is ok to spend a little
>>> bit of money.
>>> 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.
>> 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.
>> No virus found in this incoming message.
>> Checked by AVG - www.avg.com Version: 9.0.814 / Virus Database: 271.1.1/2849 - Release Date: 05/02/10 01:27:00
> 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.
More information about the time-nuts