[sdiy] hot stuff (oven)

Czech Martin Martin.Czech at Micronas.com
Thu Apr 19 10:37:37 CEST 2001


Ahh, maybe this "subject" line will leed to many people deleting
it (I get lots of this spam in my hotmail account).
The last days I thought about temperature.
Heated circuits are not dead. Precision quartz osc. are all
in a controlled oven. Look @ www.elv.de for their heated 
quartz. This is no stupid option for analog precision work.

Now, how many power will that take?
How long will it take to heat? 
How bad will the temperature distribution be?

So I went and bought a little plastic enclosure (3x4x5 cm^3).
And put my favourite medium power tranny in, with heat
spreader (BD139). Emitter circuit, large base resistor
to prevent  runaway. Uce=30V, stable. Ic max 200mA.
That's 6W. Ok, this could go as far as 70C (Ta=18C).
But: the plastic material came out as big mistake:
very high temperature gradients inside. Very very hot
tranny...

So I changed to alu case of about that size, tranny screwed
to it inside. Now I needed some thermal insulation, two
old socks did it (clean socks, no poisonous steams or
gases evaporating ;->) Yep, again 70C after 20 minutes.
That's much too much. I guess that 3W will be ok for 
30 degree inside/outside difference. And: very low thermal
gradients inside the alu enclosure. Also the tranny keeps
relatively cool.

Next step was to scetch up the power/temp difference
relation. Perfectly linear, I measured at 50, 100, 150 and 200
mA, after very long settling time. I've forgotten the data,
have it at home. Paris-Roubaix was on Euro-Sport in 
the meantime...

Things to improove:

-reduce air volume as much as possible
-replace socks with black conductive foam (please, no styro foam!!!)
-build stable PI-regulator loop
-look for DTBS behavoiur (Device To Be Stabilized)

I think that 3W seems to be reasonable (hey, how much
power is needed for a tube filament?) , and this could be way lower
if better insulation was applied. The thermal gradients inside are
lower then 1 degree. Due to low base current/ large base resistor
no extra anti run away circuit is needed, I guess one op amp
and the BD139 will do it (servo loop).

This temperature compensation business got boring.
So I had to come up with something new.

Ahh, what kind of device I might want to heat in that little oven?
Easter eggs?

m.c.






More information about the Synth-diy mailing list