[time-nuts] Temperature controlled TCVCXO

[David]

Is the performance limited by other than temperature induced drift of
the TCVCXO like the properties of the AT cut itself?  It would seem
odd if holding the temperature of the TCVCXO to a very narrow band did
not significantly improve drift.

Before the advent of better than OP-07 type low drift operational
amplifiers and even for a time after they were introduced, you could
buy little temperature controlled ovens made to fit over TO-99
packages; they made a huge improvement in operational amplifier drift
performance even though the little ovens were not really that good.
Even with OP-07s, we were grading the amplifiers and *then* using the
little ovens; chopper amplifiers were too noisy.

On Fri, 13 May 2016 07:37:14 -0400, you wrote:

>Hi
>
>Yes, it can be done. The performance will be poor, but it will work. 
>
>Bob
>
>> On May 13, 2016, at 4:15 AM, David <davidwhess at gmail.com> wrote:
>> 
>> Has anybody tried using a thermoelectric cooler to maintain a TCVCXO
>> at a constant temperature in lieu of using an OCXO?
>> 
>> I was thinking about this after reading some application notes in
>> connection with constant temperature control using TECs with bipolar
>> drive.  The power required would be pretty low so I think a linear
>> control circuit could be used avoiding the noise problems associated
>> with a switching controller.  Temperature sensor placement would be a
>> challenge.
>> 
>> I suppose as part of an automated calibration, the temperature could
>> be adjusted over a few degrees from ambient to find any flat spots in
>> the oscillator's ppm/C curve.
>> 
>> One place I have run across this application of TECs is with laser
>> diodes where coherence length depends on very tight temperature
>> control.  Typical laser diodes without temperature control have a
>> coherence length of millimeters making them useless in more demanding
>> applications like interferometry.