Help with Tonebender MKI Oscillation/Howling

[middleburyad]

Hey everyone! First post, hope it goes alright! I've been building properly for about three years and have been getting more and more into doing the "mojo" builds; they're just a ton of fun and pleasantly ingredient dependent.

I'm working on a Tonebender MKI for myself and ran into a funny problem when I get it boxed up. On the breadboard, it sounds ace! I got some transistors from Small Bear that really kick ass, and proved reliable when breadboarded. However, when I put my vero build into an enclosure with a 9V DC jack (I didn't leave space for a battery, since I personally typically just use an isolated power supply), I got this howling/oscillation not dissimilar to the howl of some silicon Fuzz Faces, that encroaches on the end of notes on some attack settings and remains persistent on others.  I've been trying different things to filter this out without luck, whether it's decoupling with a 100uf cap between negative 9V and positive ground, or using a limiting resistor in tandem with this.

Is this just an issue of sloppy vero work/criss-crossed wires in a small enclosure? (truthfully, I built up the board over some sleepy nights, so it wasn't an A+ job underneath, but I always go through and make sure to clean up my joints). Is it an issue of using a 9V DC jack instead of a battery? I only wonder about this being the issue because I've built Buzzarounds, Rangemasters, etc. running off of an adapter with no issues, even without decoupling, honestly.

In case of common questions:
Q1 is a CV5712 with an hfe of ~50, unfortunately unsure on leakage, though I assume fairly leaky cause it lets a generous amount through (compared to the low leakage OC45 I have laying around which prematurely gates the whole thing)
Q2 is some GE device with hfe of 114 and leakage 75
Q3 is a OC75 with hfe of 84 and leakage 338

-9V and positive GND are appropriately flipped on the DC jack.
Pedal is appropriately grounded

TLDR: The pedal sounds awesome on breadboard being powered by a 1Spot but gets cranky when boxed up on vero.

[antonis]

First of all, Welcome.. 

Try a 47-100pF cap between each transistor Collector-Base and/or a low value resistor between each individual stage supply..

[middleburyad]

Thanks, I'll have to try that! I suppose I should mention that I used metal film on the breadboard and carbon composition on the vero, but would that really make **that** much of a difference? Seems too excessive to just be an issue of carbon comp, especially since I've used them before without any noticeable change (I'm just using them for the fun of it, really)

[middleburyad]

Thanks, I'll have to try that! I suppose I should mention that I used metal film on the breadboard and carbon composition on the vero, but would that really make **that** much of a difference? Seems too excessive to just be an issue of carbon comp, especially since I've used them before without any noticeable change (I'm just using them for the fun of it, really)

More for archival purposes than anything else; just checked this, carbon comp makes no difference on the breadboard. I would've thought as much but it seemed worth the detour. Feel like I'm hunting gremlins a little bit, I guess the layout just wasn't happy?

[11-90-an]

Welcome...

Carbon comps are noisier..(at least that's what i heard)

Pictures of your vero? Back and Front....

[willienillie]

-9V and positive GND are appropriately flipped on the DC jack.

Please explain.

And carbon comps may add a little hiss in the background, but not howling.

[middleburyad]

Will have a photo of the vero up in a sec! Ended up rebuilding it; I tried MacGyvering solutions too many times and became annoyed with the war-torn appearance haha.

As far as the DC jack flip, what I mean is that since the transistors are germanium, (+) and (-) are appropriately "flipped" on the DC jack so that the pedal can be properly powered. Basically, positive still goes to the positive tab, and negative still goes to the negative tab, but I've taken care that negative is regarded as "hot" and positive is used as ground.

[willienillie]

If you're using a plastic-body DC jack, one that doesn't short to the enclosure, then there is no relationship between the DC jack wiring and whether the effect is positive ground or negative ground.  Most of us use the (near) industry standard of center pin negative on our DC jacks, so almost any common pedal power supply will work.  If you were using a metal-body DC jack, that would still work out for a positive ground circuit.

But yeah, pictures would help.

[middleburyad]

Apologies for late reply! I ended up building a new board and salvaging the new transistors because all my futzing with it just got it to a point of being an ugly nightmarish hack jab. I tried to get this board nice and clean but the howling oscillation persists. I just don't understand why it wouldn't be a problem on breadboard but would be boxed up? I've read some mojo-y wisdom that it's an issue with wall power, but the breadboard didn't yield the same kind of oscillation.

[moterrac]

Looking at your photos it looks like you used the layout from tagboardeffects. I could never get this layout working no matter how much I tried, I would get the circuit working great on the breadboard and then it would be a terrible screeching mess. I ended up using the layout on dirtboxlayouts instead and it worked first time with no oscillations or howling, even with adjusting the layout for transistor sockets. I could never work out where the problem was with the layout I was using previously, even with shielded wire I had no luck with it.

[willienillie]

I don't see any obvious errors on the tagboardeffects layout.  It seems to match the D.A.M. schematic I have.  On your build I see you have a 180K in place of a 470K (for biasing I assume) and a 1M in place of a 2M2 (for more volume I assume).

Something else here is how sharply you've bent the legs on the axial caps to fit them into the layout.  I know from experience that the smaller Sprague Atom electrolytics don't tolerate that well on the + lead, and I've even had it cause squealing in a Fender amp reverb circuit once.  I've also ruined Mallory 150s doing that too, but more of a no-connection situation.  Not sure if that relates to what you're experiencing, but for a test maybe try a radial electro in place of that Sprague Atom.

[anotherjim]

You can get howling due to positive feedback between input and output. Physical separation is a cure. There is a lot of amplification, so just a little output picked up at the input sets off the howling. That's why it's called positive feedback, since it reinforces itself, not because its a good thing -  unless you want an oscillator.

Just the type of stompswitch wiring can do it. Where you have a 3PDT switch, move the wires over so the middle row of contacts aren't used and connect the unused middle pin to audio ground. Keep the input and output jack wires well apart. I always use a screened cable for the input jack wiring. The screen only needs connecting to the input jack ground at one end.

[middleburyad]

So I tried replacing both the Sprague Atom, and physically distancing the leads on the footswitch, and neither yielded any better results. It's annoying cause the fuzz itself sounds awesome, but that hiss is too much to even get excited about on an "unruly/unstable badass fuzz" basis, haha.

The two thoughts I have left are either that the pedal does need a battery instead of wall wart, as according to some stuff I've read about other Tonebender and Analogman pedals doing this? (In which case, I have 9V barrel/battery adapters on the way), or perhaps as moteracc said, it's just an issue with the layout? Which would be a humongous pain, but not outside the realm of possibility? My buddy built me a TB MKI with the same layout and it works great, but it's battery only and on wide-pitch stripboard, so those are other factors to consider.

Thanks for weighing in everyone! It seems like a common enough problem with Tonebenders (moreso the MKII, evidently, but regardless) that hopefully this'll be useful for someone else down the road.

[anotherjim]

If its ok on battery, you may need a bypass/smoothing capacitor on the 9v supply when used with a PSU. Most early fx only had a battery so didn't include a supply capacitor.
Like here...

[middleburyad]

Yeah I tried a 100uf cap between positive and negative but that didn't help either. It's a total shot in the dark honestly, I don't have the background to really understand power filtering (though this has had me doing some more reading).

[willienillie]

So is it a hiss or an oscillation?

[jimitrader]

I had that issue with a creepy fingers fuzz bud...super simple fuzz ..breadboard was great... in the box...massive squeal super loud insane sound?  so went back through it and it was a bad solder joint. that was it...peace

[middleburyad]

So is it a hiss or an oscillation?

Apologies if my terminology isn't clear, I wasn't sure if there's a specific name for it. It's that high-pitched sort of "wizzing" that on some attack settings comes in in the decay, and in others is just present. It's similar to the "silicon whine" that some Fuzz Faces get.

[willienillie]

SMPS wall warts can have a high-frequency noise component, which may not filter out well with a big electrolytic cap.  It should be fairly easy to temporarily hook a battery up to your build while it is still in the box.  If the noise goes away, then you know it's coming from the power supply.

[brazdj]

Hello,

Just some observations (and in no way am I an expert):
1. It looks like you almost have a solder bridge between the second and third row, on the right side where the solder looks real shiny. Surely you have done so, but if not make sure to check the rows for (dis)continuity.
2. It is hard to see, but it looks like the top of your board opposite of the two structural posts is touching the enclosure (at the radial electro cap). A potential short?
3. I know it is not always necessary, but why not try soldering positive ground to your output jack instead of leaving it floating? In theory should not matter since the input jack is positive ground...