On 18/07/11 06:40, Chuck Harris wrote:
I understand that, but you didn't answer my question:
Has anyone seen any failures that are attributable to heat?
50 year theoretical life vs 75 theoretical year life probably
isn't going to be too significant. 1 year life vs 10 year would
be.
These GPSDO's are made to run with their ovens powered, and hot.
Surely the manufacturer took a little time to select parts that
would survive the intended operating environment.
Cooling down an oven is not a good thing, as it simply ups the
power the oven control circuitry dumps into the oven to keep it
hot.
Actually it is two-folded...
You do need a certain amount of cooling such that the oven remain in a
linear control-state. If you have too little cooling, the oven will turn
off completely for periods and then click in and heat up and then turn
off again. I've seen what this does to the frequency and it is not nice.
You do not want too much cooling, or you will draw a lot of current in
the heating regulator and this will shorten the life of that
transistor(s) significantly as it is both hot and operating with high
current, a double-bad situation.
So, instead of just slabbing a big passive radiator there or a strong
cooling fan, what I was discussing and what Charles and Bert has been
implementing is a temperature controlled cooling.
I agree that the temperature needs to be higher than ambient, but with
an active cooling you can have higher cooling dynamics than otherwise be
premisable, so the baseplate temperature may be higher than for a
passive radiator.
If you want to use the iron stabilisation proposed by Poul-Henning,
which can be an additional trick to use, mounting pre-heating resistors
on the iron body to pre-heat it up to operating range may be
recommended, since the heaters of the rubidium oven isn't that powerful
and the heat capacity of a big iron lump can be quite significant.
Essentially, it behaves like an oven, just that the main controlling
mechanism lies in cooling rather than heating. Pre-heating of iron blob
just fits the picture.
Cheers,
Magnus
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
