[time-nuts] SRS SR620 External Source Issue -- Help Request -- A thought
Dr Bruce Griffiths
bruce.griffiths at xtra.co.nz
Wed Feb 21 20:25:33 EST 2007
TheInfamousFlavio at hotmail.com wrote:
> Here's a thought. Perhaps there is a problem with the rear reference input;
> A and B channels are fine since they read the same error. The question is
> "Is there a way of adjusting the Reference Input so that it correctly read
> the 10Mhz input?"
> I tried the same setup with my other counter and it read 10,000,000.0000 or
> thereabouts correctly. So rule out the GPSDO and the cables.
> ----- Original Message -----
> From: "Dr Bruce Griffiths" <bruce.griffiths at xtra.co.nz>
> To: "Discussion of precise time and frequency measurement"
> <time-nuts at febo.com>
> Sent: Wednesday, February 21, 2007 19:47
> Subject: Re: [time-nuts] SRS SR620 External Source Issue -- Help Request
>> Hej Magnus
>> The Frequency measurement logic (before the interpolators) appears to
>> use a pair of 2 bit synchonisers, one for the rising edge and one for
>> the falling edge of the internally generated frequency gate. The
>> synchronisers synchronise the respective transitions of the frequency
>> gate signal to the external frequency being measured.
>> One synchroniser uses U508A + U509A the other uses U508B + U509B, so
>> that any mismatch in propagation delays between U509A And U509B will
>> result in a small error in the time interval. Even for ECL flipflops a
>> propagation delay difference between U50(A and U509B of 150ps wouldn't
>> be unusual. As far as I can tell the calibration procedures do not
>> address this issue. The uncalibrated differential delay through the
>> interpolation circuit source selection multiplexer will also contribute
>> an offset.
>> Perhaps a better assignment of propagation delay differences may be
>> U509A and U509B 50ps, multiplexer and wiring delay mismatches 100ps.
>> time-nuts mailing list
>> time-nuts at febo.com
> time-nuts mailing list
> time-nuts at febo.com
However if you run a test with say a 0.1 sec gate time and average
several readings, then if my model is correct the apparent offset in the
mean will be increased by a factor of 10 over that for a a 1 second gate
More information about the time-nuts