vco7c thermal results

[tomg]

----------
From: jorgen.bergfors at idg.se
To: vco at mindspring.com; synth-diy at mailhost.bpa.nl
Subject: Re: vco7c thermal results
Date: Tuesday, January 26, 1999 2:18 AM

>I did a few experiments for the fun of it. I set the vco at 440hz then
>sprayed freeze on the 3046 for 1 sec. freq went up to 582hz . It took it
>about 60secs to recover to 440hz then after a little minor drifting
>stabilizing at 440hz in about 3mins. Then I took my 45w iron and placed it
>in the middle of the 3046 for 30secs. The freq went down to 382hz again it
>took it about 60secs to recover to 440hz stabilizing in about 3mins. So I
>got out the hair dryer full bore and max heat about 6" from the board for
>30secs freq went down to 332hz and took about 5mins to fully recover.
 >-tg

This might indicate that the temperature compensation does not work, but on
the other hand these tests are rather extreme. You should perform some
tests that more resemble normal use. Around ten degrees change inside the
synth is probably realistic when you do a live gig. This corresponds to a
larger change in free air, as the synth case isolates  more than you would
expect. For studio use the temperature variation ought to  be less.
Below is an example how you can do a real world temperature test:

I put the VCO board (size 50 x 75 mm, 2 x 3") in a shoe box and put on the
lid. 20 cm (8") from the box I placed a 300 watt infrared heater. After 15
minutes the temperature inside the box had rised from 24 to 34 degrees C.
This was less than expected, as the box was just bearable to touch on the
outside (the side facing the heater).
With the tempco the frequency of the VCO increased from 518 to 523 Hz. This
corresponds to 17 cents or 1/6 of a semitone. With the tempco replaced by a
normal 1% metal film resistor, the frequency increased from 424 to 470 Hz.
That's 190 cents or almost two semitones.

/Jorgen

I think it's ok. I just push things too far, I wanted to put it in the oven
but the wife said no! It was more like NO!  Ambient temp changes of +10-20o
with a hair drier and cooling had no real effect on the pitch, about like
your tempco results. I didn't do or post the results of non-heater test
because I thought the heater test were so good. I have been forced to
change my opinion about heaters, I guess it is worth the supply drain for
the stability. Gonna have to put a bigger heat sink on those regulators
though. Anyway I pulled the 470 ohm resistor to shut the heater off and let
it run for 5mins. I set the freq for 440hz. I sprayed freeze on the 3046
for 1sec. The freq went up to 528hz and after about 15mins I gave up and
used my finger to heat it up enough to get it stable. It never made it back
to 440hz at any point on it's own, I had to re-adjust it. I let it run for
another 10mins before I did the iron test. After 30secs the freq went down
to 425hz and again I gave up after about 15mins. It never made it back to
440hz on it's own.
These are not the kinds of tests you do no a non-tempcomped
circuit....obtusely. I wonder how your tempco would react to such extremes.
Anybody want to hurt your vco (compensation of your choice) a little for a
comparison? BTW when I said "then after a little minor drifting" maybe
+-5hz .It was useable at the initial frequency within 60secs of being
thermally slammed. I though that was pretty good. 

-tg