[time-nuts] Ebay FE-5680A Rb: Are they that good?
smither at c-c-i.com
Sun Dec 11 11:31:47 EST 2011
Bob Camp wrote:
> If you dig into the circuit, the resolution of the DDS is indeed 7x10^-13 per LSB.
Thanks for the confirmation.
> On Dec 9, 2011, at 9:51 PM, Bob Smither wrote:
>> Bob Smither wrote:
>>> My documentation states that the full range of adjustment is:
>>> 7f ff ff ff => +383 Hz
>>> 80 00 00 00 => -383 Hz
>> This scale factor is clearly wrong - using it and applying an appropriate correction changes the frequency way too much.
>> (thanks Jose!) a better estimate is 140000 / Hz.
>> Using data after 20 hours:
>> the estimated frequency offset is 1.36E-3 Hz. Using the above scale factor gives a correction of -190 counts. Here is the command I used to (temporarily) program the new offset:
>> 2e 09 00 27 ff ff ff 42 bd
>> (the bd is the exclusive-or checksum of the offset)
>> Now I wait ...
After 13 hours with the above correction the offset appeared to be ~ +3E-4 Hz or
3E-11. This is close to the expected aging of the unit (2E-11 / day).
A plot showing the reduced drift from the offset correction of -190 is here:
Adding a correction for the +3E-4 Hz to the previous correction gives a new
offset of -229 counts. Shortly after I got this set up, a power line glitch
reset things so I took the opportunity to adjust the time interval to ~ 100
usec. I then applied the -229 counts as:
2e 09 00 27 ff ff ff 1b e4
(the e4 is the checksum) and starting measuring again.
It looks like this is chasing the aging of the unit. More later.
Bob Smither Smither at c-c-i.com
A well regulated Militia, being necessary to the security of a free State,
the right of the people to keep and bear Arms, shall not be infringed.
-- Second amendment to the Constitution of the united States
Circuit Concepts, Inc. 281-331-2744
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 140 bytes
Desc: not available
More information about the time-nuts