WOW Signal Fuzz. Ideas anyone?

[digi2t]

Hi all. Here's a little brain teaser for the more talented here. Here are the gutshots of the Spaceman WOW Signal Fuzz;











I finally got around to tracing the PCB, but they went the extra mile to grind the IC's, transistors, and Vactrol. There's one cap that I'm not sure of the value, but that's a blip on the radar.

I have some voltages, except for the tranny on the Vactrol, which bounces up and down all the time. Here's the PCB trace;



Any guesses on what the IC's and trannies (other than the Russian PNP) might be?

[Quackzed]

tl072 ic's ,npn 2n5088 trannies?
i dont THINK theres a 555 or 386 hiding under that tar, so... op-amps!
the bottom right tranny is being used as a diode, so it dont mattah so much theah...
btw i'm not talented, so my 2c is probably worth just that. if not less.   

[digi2t]

Yeah, so far from what I've found on the net, I'm betting on opamps too. The transistor on the Neg leg of the vactrol is reminding me of a current control maybe?

The transistor arrangement on the input is throwing me right now. I haven't found anything similar yet.

[dorrisant]

Oh yeah! Here we go with another one!

Tony

[Ronan]

This is a possible schem of half of the circuit, have not done the LFO/tremolo side yet. Could be wrong, at best its a guess. There seems to be a track or tracks or something missing around IC1 (the one at the top and to the right in your drawing Dino). I would guess that IC1B does the tremolo bit, but it isn't connected to the rest of the circuit. Also the 2 trannies on top of each other could be either way (pnp on top or npn on top), I don't know enough about electronics to work out what's most likely from the given voltages.

If nothing else, its a start, and maybe someone more knowledgeable is saved from some legwork and can help out.

[slacker]

If the first two trannies were flipped round they look like a Lockhart frequency tripler as used in Tim E's Triple Fuzz and the Zvex Machine.
Simulator says that would work and that the way it's drawn now basically does nothing and just passes signal.

[Ronan]

Thank you Ian, that's probably cracked the hardest bit. The rest of it doesn't look too difficult, at this stage.

[digi2t]

God I love you guys. Going to sleep less stoopid again. Nonselective frequency tripler, first time I`ve heard of that. Worst part is that I looked at Tim E`s Triple Fuzz, but it didn`t register with my brain. Going to breadboard it, along with the pnp germ, and see what it does. Just to be clear, the two transistors on the board are not the same. One has the offset legs, while the other has straight ones, so for sure two different trannies.

Would this principle work with a jfet / bjt combo? Just throwing that out there.

Correction to the layout I posted above; The 4.7K resistor, situated between the IC and the Vactrol, in series with the 100K and 1K resistors DOES NOT go to ground. It was a mis-draw on my part, but looking at it, I thought that it was weird having one side of the LDR going to ground. I`ll fix it tonight.

[slacker]

Just noticed your schemo has voltages, they pretty much match up with what the sim says the voltages would be for the frequency tripler arrangement simming it with 2N3906 and 2N3904.

Should be easy enough to tell if that's what it is if you build it, it's a pretty unique sound.

[digi2t]

Just noticed your schemo has voltages, they pretty much match up with what the sim says the voltages would be for the frequency tripler arrangement simming it with 2N3906 and 2N3904.

Should be easy enough to tell if that's what it is if you build it, it's a pretty unique sound.

Thanks Ian. For those about to fuzz, we salute you! 

Going to clear the Harmonic Percolator off the breadboard, and do some tests with this puppy.

[Ronan]

If you haven't already breadboarded it Dino, here's a rough drawing without the LFO. IC pinout given is for dual opamp TL072 etc.

[digi2t]

Thanks Ian. I've decided to vero and build this, but with two mods. First, a second footswitch to kick the LFO on and off, and the LFO doubler will be an external toggle.

I really like the fuzz, it's unique. It would be nice to have it on it's own, without the LFO. The original has the LFO going all the time, and all one can do is turn the rate and depth to minimum. I believe a footswitch would be better. I'll probably just tap the 9v+ feeding the LFO IC and Vactrol, a la Uglyface.

Done deal!

[digi2t]

Well I putz'd around with it in TinyCAD, and this is what I've come up with;



I apologize for not drawing proper op amp symbols. I use these drawings more for assembly purposes, when I'm proof tracing my veros.

If I take a cue from the Tremulous Lune, I might try a TL072 for the audio side, and a 4558 for the LFO side.

I'm going to breadboard the fuzz side today, and see what sounds good. I have some P416A's, but I've also got a bunch of other Russian PNP germ's. You never know...

[digi2t]

Well, I dedicated my day to the fuzz section today. After lots of trial and error, a lot of voltage measuring, and raping my stash of transistors, here is the result;

The PNP/NPN transistors are REALLY finicky about who they're making whoopy with. I tried matched 5087/5088 sets, but I kept getting a crackly, crunchy, gatey decay. I wasn't impressed. I realize that in Tim E's diagram, the resistors are matched as well, but I didn't want to go that route, since I wanted to find how the original was making it's magic.

Finally, after lots of auditioning, Cupid's arrows ended up hitting the 2N5087 in the PNP spot, and a BC546B in the NPN spot. Measuring the gain, I saw that they were fairly close in gain. Sifting through them, I matched two of them at hFe 315. That was it... Nirvana! As a second choice for the NPN, the 2N4401 is good as well, though the decay is just (and I mean JUST!) not as sweet. I know that's what I'm using in the video, but while it was uploading, further testing sold me on the BC546B. Either way, you have to try and match them as close as possible. I tried different pairs, but as soon as I drift more than 20 points apart, the decay starts to suffer.

Now, with that said, my voltages don't quite line up with the original. This is what the original puts out;

PNP
E - 1.609v
B - 1.002v
C - 0.573v

NPN
E - 0.411v
B - 1.002v
C - 0.573v

PNP germanium
E - 5.87v
B - 5.25v
C - 128.5mV

My version is as follows;

PNP
E - 1.625v
B - 1.019v
C - 1.017v

NPN
E - 0.411v
B - 1.019v
C - 1.017v

PNP germanium
E - 5.73v
B - 5.25v
C - 129.7mV

I'm not quite sure why my collectors are pumping out twice the voltage than the original, but one thing for sure, the more I lower that collector voltage to the original spec (3904/3906, or 5087/5088 combo) the more gating I get. Which is weird, since the original really isn't that gatey at all, quite smooth actually. Anyway, I'm a true believer in letting my ears guide me in the end, and if double voltage is what it takes, then so be it! 

 

[Ronan]

Sounds really good Dino, another good job! Are you going to do the LFO as well? If not, a simple buffer after the Ge PNP and before the tone control would be worth trying.

[digi2t]

Sounds really good Dino, another good job! Are you going to do the LFO as well? If not, a simple buffer after the Ge PNP and before the tone control would be worth trying.

Oh, I'm going all the way on this one, plus a couple of mods. I'll be using a second footswitch for the LFO, like the Uglyface, and a toggle on the outside for the LFO frequency divider.

Then again, this might be cool in a wah shell as well, with the treadle controlling either the depth, or speed. Just a thought though, I'm going for a conventional box.

I checked the trimmer, to see what resistance I had for the germ tranny, and it shows exactly 27K. Not too far from the 29.4K that's in the original. Like I said in the video, you could use a trimmer to pin point the right voltages, and then swap in a resistor. Or, leave the trimmer, and tune it to your ear's delight.

[PRR]

> not quite sure why my collectors are pumping out twice the voltage

That joint-collector voltage is utterly *undefined*.

The lower emitter will be near 0.4V, the upper emitter 1.2V higher or 1.6V; this is forced by 9V supply, 0.6V+0.6V base-emitter junctions, and 22K and 1.2K resistors.

At dead-silent idle: The joint-collector point could be _any_ voltage between 0.4V and 1.6V. The DC impedance here is super-high, DC voltage gain is super-high, and nothing to enforce any specific voltage.

I don't think that voltage means much. If diddling it (how??) is changing the sound, you are probably changing something else.

With significant input signal: the audio bangs between 0.4V and 1.6V. For fairly symmetrical signals, the DC voltage observed should settle toward the center, about 1V.

[digi2t]

> not quite sure why my collectors are pumping out twice the voltage

That joint-collector voltage is utterly *undefined*.

The lower emitter will be near 0.4V, the upper emitter 1.2V higher or 1.6V; this is forced by 9V supply, 0.6V+0.6V base-emitter junctions, and 22K and 1.2K resistors.

At dead-silent idle: The joint-collector point could be _any_ voltage between 0.4V and 1.6V. The DC impedance here is super-high, DC voltage gain is super-high, and nothing to enforce any specific voltage.

I don't think that voltage means much. If diddling it (how??) is changing the sound, you are probably changing something else.

With significant input signal: the audio bangs between 0.4V and 1.6V. For fairly symmetrical signals, the DC voltage observed should settle toward the center, about 1V.

Thank you very much for your insight Paul. Very illuminating, as always. Diddling with the circuit really doesn't change the "sound" per se, just the gating/decay. In reality, like you point out, the E and B voltages don't change much, but in my case the higher the C voltage, the better the performance. That all depended on which transistors I used. Like I said, judicious choice of transistors is the key here if one is looking for smoothness of decay.

Here's an initial vero offer. I'm mounting the LFO speed switch on the outside as a toggle, and I've added a footswitch to turn the LFO off. As always, not verified, until I build it, but I always put them out there for the eyes to spot errors.



Going to do a check trace tonight, but if anyone spots anything, please pipe up.

Cheers,
Dino

[LucifersTrip]

I'm not quite sure why my collectors are pumping out twice the voltage
 

well, it's a cliché, but you're simply not using the right hfe's. i tested the theory with 2 x 50 hfe's and got 1.59V, then thru a couple 800 hfe's in there and got .58V
...this is just a break from the Turkey build, which, it seems, I may need every day for, so don't know if I'll have time to tweak...

edit:
an 800 (npn) / 50 (pnp) combo also gave me ~ .5V

[digi2t]

OK, after trace checking, here are the corrected schematic and vero. The vero is still unverified, but it matches the PCB trace, so I'm fairly confident that it should be alright.




(Should have 11 Nov. 2012 on it, if not, Photobucket is slow to update.)

Going to start putting it together this week, and I'll update when it's done.

Cheers,
Dino