Help with oscilation running Thunderchief at 18 volts

[Atodovax]

Hi everyone!
Does anyone know why this circuit will start oscilating badly when powered with 18 volts? It works great with 9 volts but using 18v and adjusting all the fets to + - 1/2 VCC, would introduces some severe oscilation...

[antonis]

I presume you didn't alter Sources resistors values..

You do realize that the double the PS voltage the double the current through Drain-Source channel, hence the double the V_GS , don't you..??

[Atodovax]

Upss  .
I knew i had to do it! Would it be as simple as doubling each source resistor value and then adjusting the drain back?

[antonis]

Double current and double resistance result into quadruple voltage drop..
(halving them perhaps..??)

[Atodovax]

Double current and double resistance result into quadruple voltage drop..
(halving them perhaps..??)

Let me try that and come back. I've seen other circuits like the Supreaux used by JHS at double the voltages, using lamost the same Source resistors, so i though i would be ok to leave them as they are

[Atodovax]

Double current and double resistance result into quadruple voltage drop..
(halving them perhaps..??)

Would you also half the 4k7 and 470r at the source of Q3 and the junture from the NFB patch?

[antonis]

I might confused you by "halving" those resistors because the current drawn by the FETs is unlikely to change linearly with the voltage increase.

You have to estimate/find the optimum V_GS for the new current and set Sources resistors values according to Ohm's Law..

P.S.
Or just leave as it is, place Miller caps on all JFETs and cross your fingers..

[Atodovax]

I might confused you by "halving" those resistors because the current drawn by the FETs is unlikely to change linearly with the voltage increase.

You have to estimate/find the optimum V_GS for the new current and set Sources resistors values according to Ohm's Law..

Can i just measure the Vgs off using 18volt instead of 9volts and use the Fetzer calculator?

[Atodovax]

I might confused you by "halving" those resistors because the current drawn by the FETs is unlikely to change linearly with the voltage increase.

You have to estimate/find the optimum V_GS for the new current and set Sources resistors values according to Ohm's Law..

P.S.
Or just leave as it is, place Miller caps on all JFETs and cross your fingers..

I think the problem might be coming from the input being extremely sensitive? if i turn the knob on my guitar down, there is some craclking and the oscilation stops. The input is super sensitive when i put my hand close to it

[antonis]

Try a 10nF input cap and a 220pF cap across 1M resistor..

[Atodovax]

Try a 10nF (or bigger) input cap and a 220pF cap across 1M resistor..

Let me check that!

[Atodovax]

Well.. That didnt do it . If i put a lower gain i Q1, i can get rid of the oscilation but the distortion is much less... Any idea how to easely calculate the source resistors? I have tried two different PCB layouts and with 9 volts and the problem only seems to be when powering it with 18v. I also tried with batteries. Same problem.. I tried with miller caps but had to go too big in values and that resulted in too many high frecuencies lost.. This is driving me crazy

[PRR]

It is a "distortion pedal", isn't it?

Think. More distortion comes from LESS supply voltage. Why is everybody jacking their supplies today?

[Atodovax]

It is a "distortion pedal", isn't it?

Think. More distortion comes from LESS supply voltage. Why is everybody jacking their supplies today?

Thanks PPR! Yes, im not looking for more distortion than at 9volts. Im just saying that in order to use it with no oscilation, i have to set it qith to little distortion (using a low gain Q1).
The main goal is to use this as a Preamp for a Germanium power amp section (all powered together by 18volts).
Now im wondering why it oscilates so badly at 18 volts.. The input is so so sensitive. Also, if i leave the input with no load, it squeal even at almost minimum. I had this same problem with a BSIAB II before. The only eay to kill it is to put a very large cap across thh input. Much more than the capacitance of a bad cable or even a guitar. So, the resulta are quite obvious, too much treble roll of .

[Atodovax]

Try a 10nF input cap and a 220pF cap across 1M resistor..

Well.. It turned out to be the input pad very close to the 470k resistor at Q2 gate. That and a filter capacitor that was right over one of the signal paths that was causing to much inductance and crossover over the input and it was selfoscilating. Thanks Antonis for the help!