[time-nuts] TECs in cooling below ambient

[Poul-Henning Kamp]

--------
Bob kb8tq writes:

> I think you will find that things work out a whole lot better if you target something
> just above room temp. If your room runs 22 +/- 3 C , a set point of 27C likely results
> in better operation than 17C. 

This decision should be based on expected abnormal situations.

If your ambient might reach 35-40°C when cooling fails, running lower than
ambient will cope, running higher than ambient will not.

If your ambient might drop to -20°C when the heating fails or the lid is
opened, then running higher than ambient is the robust thing to do.

And as others have said:  If you run lower than ambient, you have to plan
for condensation.

Bidirectional TEC setups need special attention: You have to take
into account the 3:1 efficiency difference between heating:cooling,
and only seldom, and then gently, switch direction, in order to
reduce thermal stress in the TEC element.

A sound bidirectional design use an inner oven to keep the payload
temperature constant, and use a more coarse outer TEC loop only
to keep the oven's operating conditions inside a narrow window.

That way the TEC can operate with a suitably slow time-constant.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.