[time-nuts] LT1763

John Green wpxs472 at gmail.com
Mon Oct 5 21:37:23 UTC 2009


Poul Henning wrote:
You should pick your poison carefully here:

Low noise regulators are usually not very stable, typically they
are quite sensitive to temperature.

Precision regulators are stable, but tend to be noisy.

If your OCXO is voltage sensitive, make sure to connect the sense
resistor to the terminal, right at the OCXO VCC terminal.

If the heater has a separate VCC terminal on the OCXO, give it its
own supply.

Another trick: if you don't plan to run your OCXO in high temperatures
is to insulate it thermally.

That will reduce the current draw from the heaters, making regulation
easier, and if done right, will also low-pass-filter the temperature
excursions.

But don't overdo it, it has to be able to get rid of excess heat
and you may shorten the electronics life if it gets too hot.



The LT3080 seems to be the best of both worlds, at least according to the
spec sheet. Low noise and excellent regulation, temperature stability. This
particular OCXO doesn't have a separate heater connection. It also doesn't
have EFC. Just a coarse and fine adjustment. I plan on building a little
thermal isolation housing around it. It will be in a non cooled and heated
environment. The voltage regulator will also be inside the housing. I will
have to leave access to the adjustments so I can't get it completely
thermally self contained. Since it will be at a location with limited
access, I hope it will be stable enough not to require adjustment but
perhaps every 2 years or so. For this reason, I am also considering using a
Rubidium standard instead. I don't have enough history with Rubidium
standards to yet feel comfortable putting one where I won't be able to get
at it very often.  There are other parts of the project that have to be
completed so I plan on building the reference oscillator first and letting
it just sit and run for a couple of months to see how it works.


More information about the time-nuts mailing list