[time-nuts] OCXO Retrace

Charles Steinmetz csteinmetz at yandex.com
Thu Feb 11 11:11:59 EST 2016


Bob wrote:

>The Trimble TBolts were made over a long period of time. If you 
>compare "early" (90's) units to the "late" (~2005) units there are 
>differences in performance.

If I'm not mistaken, that is largely because they used several 
different (part number) OCXOs.  I'm not sure it's so much a 
progression over time as it is evidence of "special" models built to 
user specifications.  The ones I've seen with any specific OCXO seem 
to perform similarly to others with that same OCXO. [1]

[1]  Except that, as I've noted before, there seem to be two distinct 
groups of the P/N 37265 OCXOs, one group with tempco about 100x 
better than the other, larger group.  Other than the tempcos, these 
two groups seem to perform similarly.  My working hypothesis is that 
all 37265s are supposed to perform like the low-tempco units, but 
that the ovens in the high-tempco units are not adjusted correctly to 
the turnover temperature of the crystals.  I did hack into one of the 
high-tempco units and, after some experimentation, was able to get 
its tempco down into the ballpark of the low-tempco units by 
adjusting the oven temperature (but an anecdotal result for one unit 
doesn't prove anything).  I have not been able to discern any pattern 
WRT date codes or serial numbers between these two groups, but my 
sample size is probably too small to reveal one even if it were there.

Best regards,

Charles




More information about the time-nuts mailing list