[sdiy] Tempco vs. "Substrate Oven" in VCO and VCF

[Rick Jansen]

Well, talking about a blast from the past :-)

It's hardly my circuit though, it's from the AN-299 application note.

Heating up the 3046 does not cost much power, it's mass is just a few grams, and the chip 
is not hot to the touch.

This is what I wrote in my documentation:

1. Heater adjustment

First make sure the heater is switched off by shorting D2. This way the base of the heater 
is grounded and Q5 switched off. Now measure the voltage present at the +-input of A1. 
This must be done with a digital voltmeter (DVM) or another voltmeter with high input 
impedance. A regular multimeter will not do. The voltage measured must be about 650 to 
700mV, being the Ube of the sensor Q3.

Do not apply power unless the heater is adjusted or switched off!

With every degree temperature rise this be drops 1.9mV. In operation A1 will make Q5 heat 
up the 3046 array until the voltage at the plus-input is the same as the (set) voltage at 
the minus input of A1. So, you now need to adjust P1 so that the minus-input of A1 is fed 
with the voltage that goes with the elevated temperature. should with the elevated 
temperature. To compute this set voltage subtract the temperature rise (in Celcius) times 
1.9mV from the measured voltage.
Example: you want a 30 degrees temperature rise and the measured Ube is 695mV. The desired 
drop in Ube therefore is 30x1.9=57mV. The minus-input of A1 must be set (with P1) to 
695-57=638mV.

rick

On 29/02/2016 01:33, sleepy_dog at gmx.de wrote:
> Hey there,
>
> first a question about thermistors (this is what "tempco" are, correct?)
> In circuits for voltage control of a VCO or VCF which use a tempco, how
> much could one deviate with the coefficient, say for instance in Thomay
> Henry's newest VCO based on the TLC555 timer? I think it was spec'd at
> 3500 - what if I could get one with 3470 or 3600 more easily?
>
> Also, I find Rick Jansen's Moog VCF schematic interesting, which uses 3
> remaining transistors in a 3046 transistor array as a controlled "oven"
> for the NPN matched pair, to get rid of temperature sensitivity.
> How come I don't see this kind of solution more often? Is the tempco
> solution "better"?
> How much current will the "oven" draw, and does this reduce the lifetime
> of the IC?
>
>
> Regards,
> Steve
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20160229/a002e0fc/attachment.htm>