Help with Tonebender MKI Oscillation/Howling

[middleburyad]

Hey all,

Sorry for late reply! Tropical Storm Isaias knocked the power out at my house for a couple days; crazy times! So I tried a couple more things, just to chase every possibility.

- I reflowed the entire board
- I grounded the output jack as well, just in case
- I powered the pedal exclusively off battery,

And the issue remains regardless. Given that I've chased basically every possible problem I'd be aware of or that this thread brought attention to, the only conclusion I can come to is that the tagboardeffects layout is just too close and allows for the wizzing. The issue seems to be ameliorated with wide pitch stripboard or with a different layout, according to others. I'm gonna futz about a bit longer, see if I can use Miller caps with any luck, etc., but it seems like the issue may be inherent to the layout, at least on standard pitch stripboard with the leaky transistors the MKI likes.

Thanks for the help everyone!

[Smeddling]

I too have had a good Mk1 build on breadboard only to have it all go to rats**t when put onto vero.
I'm using Russian MP42B, with tweaks to the Biasing resistors as the transistors are low gain (45-55) and low leakage (50-60uA)
The point is I have plenty of voltage gain and it worked well on breadboard. I even took it to band practice in a small pizza box!
I didn't use any published layout, and I know it's not the enclosure wiring as I've measured it remotely (disconnected from the wiring and just powered up, putting in varying resistance to ground to simulate attack pot) on a scope.
The frequency of the oscillation varies with the attack pot setting (2-3kHz at full, up to 13kHzish at the bottom).
I've isolated all long strips where not needed with a reamer as I thought it was due to coupling strip to parallel strip. No dice.
Tried decoupling power even though I'm running on battery (even a brand new one). No dice.
I've inspected the board very closely and do not detect any shorts or dry joints.
I'm going to get the trannies back on breadboard to check again whilst probing around for any stray resistances on the vero (whilst transistors are out)
The only major differences I've made to the circuit are:
Q2C load is 5k6, 22uF coupling Q2 to Q3, 2M2 Q3 base bias to give it a trickle of bias current To help it out due to low leakage, 47k in Q3C to get the voltage close to -8V.
Will also try the 47pF C-B on Q1-3 when I re-build it though after weeks of tinkering I ain't feeling too confident 😌
Any further thoughts would be most welcome kind folk 😄
Had the transistors back out and retested. All good.

[bmsiddall]

Late to the party but this might be useful anyway.

I've built about 5 mki TBs, mostly the same tagboard layout (apart from one silicon version). One particular build for a customer oscillated from the get-go when soldered up.  Tried swapping transistors (had these on flying leads initially), soldering transistors in place to get rid of flying leads, power supply filtering, various c-b capacitors and moving other wire leads - no luck.  On a whim i stole the mkii input cap to ground idea (10nf) - success!  No oscillation and negligible loss of high end.