[time-nuts] FE5680A Calibration

[Bob Camp]

Hi

Consider that the noise on the 10 MHz output of the FE5680 is pretty bad.
You may / may not want to propagate it all over your shop. At the very
least, a simple band pass filter is needed. A better solution would be to
lock a crystal oscillator up to the FE and use the output of the oscillator.


It's a reasonable bet that you unit is within +/- 1x10^-10 of the right
frequency. If you are lucky, it's > 10X better than that. To observe it's
frequency adequately to set it well, you will need to average it's output
for at least a couple hundred seconds. You don't want to make it worse
trying to set it closer....

If you are comparing to a normal GPS receiver, they have noise on their PPS
output as well. To get to adequate stability / resolution you will need to
average them for a while. Just how long depends on which one you have and a
few other things. 

Best approach / lazy approach / provides adequate time for a beer: Look at
the edge of the FE's pps relative to the pps out of the GPS on a scope. Note
the offset on a piece of paper along with the time. Come back in an hour and
repeat the process. Get a couple of points this way. The drift (if any) will
show up as a linear trend. The noise will likely be in the 20 to 200 ns. At
this point all you really are looking for is a noise estimate.

Say the noise is 200 ns. If you do observations at a 200 second spacing you
would get 1x10^-9. To get to 1x10^-12 you need to observe for about 1000X
longer. 200,000 seconds is a couple of days. Yes indeed there are fancy math
things try to speed things up. They don't fit under the "lazy approach"
(there also are a few other reasons to take your time). 

Based on the noise estimate you get, and the calculations above, come up
with a schedule. No need to get obsessive about it. All you are looking for
is: read every day / every three days / once a week sort of spacing. Start
logging your phase offset at what ever spacing makes sense. Take at least
ten readings. 

Based on your readings, take a stab at correcting the FE. Adjust it and then
go back to taking readings. Eventually it will be "close enough".

So much fun...

Bob

-----Original Message-----
From: time-nuts-bounces at febo.com [mailto:time-nuts-bounces at febo.com] On
Behalf Of Chris Stake
Sent: Wednesday, March 14, 2012 11:17 AM
To: 'Discussion of precise time and frequency measurement'
Subject: [time-nuts] FE5680A Calibration

Hello all,

I purchased a FE5680A from a Chinese  Ebay vendor. I have connected it to a
16.5V laptop supply, added 7805-based 5V rail and a PMOS Fet switch to drive
a LED for the "locked" signal. I can communicate with it using the excellent
Fe5680Calibrator software. As received the frequency offset is set to Zero
and the unit seems to work well.

I plan to put it into service as a workshop frequency reference and so would
like to set it close to 10.0Mhz. I have various oscillator modules, signal
generators, SDR receivers, GPS receivers, a scanner, some uncalibrated
test-gear. I live in a river valley in North Devon UK so radio reception is
a bit restricted but I receive digital television via a masthead amplifier
and long downlead.

Could someone please suggest a way of going about this?

 

Regards

Chris Stake

 

_______________________________________________
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.