Tube screamer, well, screaming.

Started by Kesh, November 09, 2012, 08:26:19 AM

Previous topic - Next topic

Kesh

I have a tube screamer I've made, un-boxed, on my bench that oscillates when no input is connected. All the pots alter the frequency of the oscillation.

I'd like it not to, even though it won't when I box it because of the input jack depowering the board.

If you omit the fet switching it is pretty much RG's schematic, which I hope he won't mind me posting here. Only difference is closest E12 series for the E24s, poly films for NP electrolytics, the .22 electrolytic and the tantalum. Buffers are 2n5089s. Output follows the 808 values.



And I also have a big M resistor to ground at input. I'm thinking I should lower this? Any other ideas?

Am curious to know why this happens too, electronically.

R O Tiree

I read a thread on here a couple of months ago about a high-gain circuit squealing because the Input/Output traces ran side by side for a little way, less than 0.1" apart.  I guess it would apply to wires as well?  Worth a little thought?  Without seeing your layout, it's hard to know.
...you fritter and waste the hours in an off-hand way...

Kesh

Stray capacitance coupling.

It's on vero. Everything runs side by side  :icon_mrgreen:

But I'll look into it.

Now i realise why the non-inverting input is never next to the output in op-amps.

R.G.

Yep. Sometimes all the odd little things add up the wrong way. Wire routing, power supply impedance, wire impedance, stray capacitances, all that stuff.

For some things to mess with, you might stick a 0.1uF ceramic from the V+ to V- pin right at the opamps. Sometimes they just sing if everything isn't right, especially TL072s.
R.G.

In response to the questions in the forum - PCB Layout for Musical Effects is available from The Book Patch. Search "PCB Layout" and it ought to appear.

Kesh

I think I've found the vero traces. I'm going to cut some unnecessarily long traces that go nowhere and would give positive feedback if they were coupling, and see if that helps. Lowering the input shunt resistor to 1M helped, as did a 33pF input to ground. Now it only squeals at max gain with the clipping diodes out. (I'm building one where the diode clipping can be eased off with a series resistance, so this matters.)

Op amp is currently a CA3260, but the thing squeals with 4558s too.

Maybe it's because half of it's on breadboard, half on vero.

Thanks guys.