[sdiy] CA3046 Heater / Substrate Question

[Nils Pipenbrinck]

On 05/06/2016 10:44 AM, Roman Sowa wrote:
> You say, Q3 base is at -1.8, but with no heating current (at start) Q3
> emmiter is at about -0.7, determined by 2 small resistors and BE
> junction of Q2. So the substrate is higher than Q3 base and all
> transistors are connected now, no substrate isolation.
> Loose the LED and it should work, but substrate will be at very low
> negative voltage now, I'm not sure if it will not degrade expo pair
> performance.

Hi Roman,

thanks, that makes sense. I'll have thoroughly think through it, but 
it's clear that I messed up the substrate somehow. I just don't know 
where exactly the s. hits the f.

> BTW, "better" is an enemy of "good". What is wrong with just 1
> transistor heater?

There is nothing wrong with it, but it's not ideal because the 1 
transistor heater can only dissipate 300mW. The cooling of the package 
takes away a big chunk of the energy we put into heating. So reaching 
tuning stability is slower than it could be.


I did some measurements yesterday, measuring the Vbe drop directly while 
heating with two different target temperatures. I faked a bit, left the 
substrate transistor connected as usual and used one of the matched pair 
transistors as my second heater:

    http://torus.untergrund.net/synth/heater_65deg.png

Here I aim at 65°C. You'll see that the two transistor circuit reaches 
stability after 4 seconds while the one transistor circuit needs about 
100 seconds:

   http://torus.untergrund.net/synth/heater_90deg.png

Here I aim at about 90°C. The two trannie method reaches stability after 
half a minute while the single transistor never reaches it's target 
temperature and asymptotically drifts to somewhere around 70°


I think this is a great improvement over the one transistor circuit, and 
the substrate transistor is unused, so why not use it as a second heater?


Best,
   Nils