[time-nuts] GPS lock of the FE5680. Current experiment and question
lists at rtty.us
Fri Feb 10 23:13:31 UTC 2012
Without seeing the schematic it's a little hard to be sure this is all correct...
If it is all based on 10 MHz, then yes you are doing it all modulo 100 ns. To get 1 part in 1x10^13 you would need 100,000 seconds. If that drives the LSB of a 16 bit counter for the DAC you would take a very long time to get to mid scale on the DAC. Even if you drive at the 100 second point, you would take about 3,200,000 seconds to get to mid scale.
On Feb 10, 2012, at 5:53 PM, paul swed <paulswedb at gmail.com> wrote:
> Just for fun have built up a simple multi-chip project to try my hand at
> locking the FE5680 using the external efc connection mod. mentioned in
> time-nuts. Typical counter micro rs 232 port stuff.
> Essentially the system is 2 X 24 bit counters that toggle back and fourth
> every interval selectable from 1 to 65000 seconds accumulating counts. I am
> measuring the 10 Mhz of the RB and using a novatel superstar II GPS rcvr. I
> do have a better timing rcvr but this was handy and at this point most
> likely good enough.
> My question is this. It seems that the system must use pretty large
> accumulation times like 1000 or 10000 seconds and higher to see the drift
> in 1 sample.
> I think this actually all makes sense. Since the frequency accumulating is
> 1 x 10^ 7 and the gate so far is 1 X 10^3 or the total resolution is 1 X 10
> 10th. Essentially accumulating 3 or so counts and averaging most likely
> only gets the lock down to 1 X10^10 or a bit more perhaps. Using a
> accumulate of 10,000 would be the 11th.
> Is my logic on target here.
> I only see a few counts one way or the other when I purposely increase or
> decrease the RB offset. But it does indeed go in the correct directions. I
> can adjust the dac by the drift and time method to get down in the -12th
> region. So essentially the eyeball method is pretty good.
> If this is all correct then for a given accumulated time there is an exact
> 24 bit number to always drive the RB to. I have an excel that figures out
> that number for any accumulation time is.
> By the way took Berts suggestion some weeks back of using the LTC1655 16
> bit Vdac. Talk about easy to work with and in a 8 pin DIP.
> Appreciate any guidance. Thanks.
> 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