Pepper Spray Diodes revisited

[das234]

I finished my pepper spray today and it sounds like a torn speaker cone in all three switch positions.  I followed the Albini specs and used the 1n695 for D1 and D2 and the specified transistors.  For D4 and D5, I plugged in some 1n914s I had. I planned to used OA126s but I don't have those yet so I socketed and used the 1n914s.  Like I said, the thing works but it sounds really ratty like a ripped speaker.  Funny thing is, I noticed I had a diode in backwards after my first test.  I switched it around and it still sounds the same. 

Any ideas on parts substitutions or why it doesn't sound like Jimi Photon's video??  I don't hear all the pretty harmonics percolating.

[midwayfair]

What are your voltages like?

Does the sound change when you change the gain control?

What's the gain and leakage on Q1?

Just to make sure ... it's negative ground despite the PNP germanium.

[Bret608]

Mine has Albini specs and the same diode combination as yours. The 1n695 position should have a tiny bit of germanium diode fizz, but otherwise should sound like a thick, dark distortion. The middle position should clean up quite a bit but be quite a bit louder. The 1n914 side should be like a really aggressive overdrive, and also louder and not as dark as the 1n695s.

I hope this helps, at least by giving you some kind of sonic reference point. Otherwise Jon's questions are the first place I would go as well, besides the usual checking part values and looking for bridges. I'm pretty sure Jimi Photon's version is the "stock" specs, but it shouldn't be too radically different.

[das234]

What are your voltages like?

Does the sound change when you change the gain control?

What's the gain and leakage on Q1?

Just to make sure ... it's negative ground despite the PNP germanium.

I wouldn't know what to do with voltages (or where to check them) so I haven't done that yet.

The sound seems to stay about the same, just the volume changes with either knob.  I'll give it another listen.

Q1 is a 2n404a that was labeled from Small Bear as 143, L=144.  I guess I can double check that it's installed correctly.

I didn't do anything different with the power supply so that should be ok.

[Bret608]

That 2n404 may be a bit higher gain than the circuit likes, but Jon would be able to comment on that better than me.

If you posted some pictures, we could help spot any potential problems such as a mis-oriented transistor. It's a pretty fuzzy circuit, but that doesn't sound quite right.

[das234]

This should show everything. The soldering isn't pretty but I believe I colored within the lines and reflowed joints that looked iffy.

[midwayfair]

I wouldn't know what to do with voltages (or where to check them) so I haven't done that yet.

Jacob gives a good explanation here:
http://www.madbeanpedals.com/forum/index.php?topic=902.0

Might want to consider moving this to tech help.

Your PNP is WAY too high gain for this circuit; I'm surprised Steve sent you one with gain that high for this circuit. It's possible he just doesn't realize what this circuit requires. Honestly I think 70 hfe is about tops for getting something decent, and I've had times when even that was too high once the temperature in the room went up a little. (This circuit is really an excellent place to use trash transistors that are too low gain for anything else.) You might try just flipping it around and using the leakage in place of the gain, or lowering the resistor from the collector to ground (socket it, then try something more like 22K or 10K).

However, I think something else might be wrong overall, because you should have been able to clean it up by rolling off the Harm, and too much gain in Q1 manifests as overcompression, oscillation, and other weirdness, not "torn speaker" sounds. Unfortunately it's too hard to tell -- I'm just guessing at this point -- without more information, so please check the tech help thread for how to do that.

EDIT: Now that I see the picture: The TO-92 package has the "bulge" backwards from the dot package -- that is, when you look at a dot package with the pins facing you, you see:
C      E
    B
But when you look at a PN3565 with the pins facing you and the bulge facing down, you see:

E B C

In short: Flip Q2.
And further: ALWAYS check the datasheet for the specific transistor you're using BEFORE soldering any transistor. Even if you've used that same part number before.

[Bret608]

I suspected a mis-oriented Q2. If this a PN3565 as Jon mentions, then by all means follow what he said; otherwise, let us know what you've got in there and we can help you figure out how to orient it.

Also, the 1n695 that appears in the upper left in your picture is facing the wrong way. They should both have the black stripe facing toward the interior of the board. I know it seems weird, but having those facing the same way with those series resistors is how the asymmetrical clipping works.

[das234]

I suspected a mis-oriented Q2. If this a PN3565 as Jon mentions, then by all means follow what he said; otherwise, let us know what you've got in there and we can help you figure out how to orient it.

Also, the 1n695 that appears in the upper left in your picture is facing the wrong way. They should both have the black stripe facing toward the interior of the board. I know it seems weird, but having those facing the same way with those series resistors is how the asymmetrical clipping works.

Ah yes.  I had that diode right the first time and flipped it because I thought they were supposed to be opposite like the other pair.

I will flip and resolder Q2 immediately and see what happens.  Thanks.

[das234]

Question... could I have wrecked Q2 by soldering it in backwards?  I'm not sure if that's possible, especially since Jon recommended flipping Q1.  I'm assuming they're safe either way.

Anyway, I flipped Q2 (and my reversed 1n695) but it didn't fix it.  It does clean up when I turn down the harm but I lose all my volume if I turn it more than a tiny bit.  I tried flipping Q1 also and it seemed to maybe make a slight improvement but it's still not what I'm expecting to hear.

I may take Jon's advice and move this debacle over to tech after I get some readings and such.

[midwayfair]

I lose all my volume if I turn it more than a tiny bit.

Define "all your volume." It's normal for the pedal to kind of shelve when the Harm is at max. A TON of fuzzes do this. You should have unity gain around 1:00 on the volume when the Harm is at 3:00 on the 1N695 setting. If you really are losing ALL of your volume, as in, the signal dies or gates, then there's something else wrong.

Also: Does flipping the diode switch do anything now?

[Bret608]

Actually, your volume loss sounds like what I was getting when I first finished mine and had the 2n3565 in backward--I had to have the harmonics at max to get much dirt, and the balance at max to get over unity. It would lose the sound completely if I took the balance or even the harmonics down to before about noon. Let me know if this is close to what's happening now.

Anyway, after you reported this, I googled the PN3565 datasheet and took a look at that with the schematic, etching layout, and your photo. I think you may have had it in correctly at first. Sorry I didn't check it this way sooner.

That leads me back to Jon's thoughts on the 2n404 being too high-gain for the circuit. I remember when I ordered mine from SB their site said they'd range from 50-ish all the way to 200 hfe, so Steve might not know this doesn't work well for the circuit. By any chance do you have another, lower-gain PNP germanium you could try if you're able to get your PN3565 flipped back the other way without any damage?

[midwayfair]

Socket the transistor.

And again, data sheets ...

[das234]

Bret, you're exactly right about what's happening now.  I checked with Balance at 1:00 and Harm at 3:00 and was no where even close to unity gain.  Way low.  And thanks for checking that data sheet for me. 
I will SOCKET and flip Q2 one more time. 
I have a GT309G with hfe of 79 that I could plug in for Q1 as well.  I didn't like it in my Fozzaround so it's just laying around.

Much appreciation for your help guys.

[midwayfair]

Sorry about that. I was looking at the etch mask side and comparing the pinout.