TB MkIII - what were collector voltages like on vintage units?

[Bret608]

Hi everyone,

I recently built MkIII on Madbean's "Pastyface" board. Following the Fulltone Soulbender, it incorporates a 500k trim and a 10k resistor between the 1n270 diode and the collector of Q3. I didn't want to do the voltage starve thing, so I jumpered pads one and two of the trimmer. I should've bumped up that 10k resistor though, because my collector voltage on Q3 is too low (0.36v). I guess all the 10k was really doing was setting the minimum resistance for the trimmer.

Taking a step back before I dig in and fix this, what were the collector voltages like on vintage units? On the schematic at Fuzz Central, I don't see anything between the diode and the 18k resistor. If it helps, I used 2n1305s in all positions with pretty low leakage (none over 80uA). I have about 3.8 volts on the Q1 and Q2 collectors.

Thanks!

[Davelectro]

Q3 biases at 1.39V in my build. I don't know about right voltages but I can tell you it sounds really good.

[Bret608]

Thanks! I bet I could hit that if I replace that 10k with about 500k. That way it will more or less bias as if the trimmer was there and at max resistance. I heard over at the Madbean forum that 1.2-1.4v sounds great.

[Electric Warrior]

The vintage units I have at hand right now measure 2.01 and 2.8V.
Low leakage is good for Q1 and Q2. Q3 likes it leaky in my experience. I guess the soulbender uses the trim pot to increase the leakage a bit.

[Bret608]

I am pretty sure that's what it's doing. I should probably shoot for that amount of resistance since the bulk of my germaniums are low leakage. I actually tried out the leakier ones I had in Q3 and the biased up even lower than the 2n1305--as low as .01v in some cases!

Do your original units have any kind of resistor or trimmer between the diode and the 18k resistor (i.e. like the Soulbender)?

[Electric Warrior]

No, vintage MKIII/IVs never had it. The transistors were leaky enough.

Here's a stock schematic:



Should be pretty close to the Soulbender. I believe the copied the version with the 47k/220k bias string?

[Bret608]

They did indeed. I actually used the parts values on the schematic you showed, with the exceptions of leaving off the 220k resistor at the end for bit more volume and having some resistance between the diode and the 18k resistor. The Soulbender is nice but a bit smoother than I wanted.

I'm gathering that 2.0v or less will sound best for low-leakage transistors in Q3. I'll report back after I get a chance to swap out that 10k resistor tomorrow. Thanks again for cracking open one of your vintage units to help me understand what is going on in some of the modern-day derivatives, including my own builds!

[Electric Warrior]

Since you're simulating leakage with that trim pot/resistor it shouldn't behave all that differently. Have you put in the 500k yet?

[Bret608]

Not just yet; I'm taking the pedal to the electronics lab where I work today to make the swap. I'll plug it in this evening to see how it went. In the interests of sharing though, here's how it looks so far; the resistor just below the empty trimmer space is the one I'll be bumping up:

[Electric Warrior]

You're probably simulating way too much leakage then.

[Bret608]

I just swapped the 10k out for 560k and that did the trick! I'm getting about 1.36v on the Q3 collector, very similar to what Davelectro reported above. Can't wait to rock it tonight! It already sounded pretty good, just a touch more starved/glitch-y than I wanted.

[Electric Warrior]

I guess with over 500k you could have simply removed the 10k and gotten the same or a similar result 

[Bret608]

Tell me more about that!  Someone told me recently that having no resistor there would actually be like having a lot of resistance. By that logic, I could have just jumpered it, you're right. I'm just curious what's going on electrically that makes this possible.

[Electric Warrior]

no, jumpering them is putting zero resistance there. Having them not connected means the resistance between parts is close to infinity. Nothing is coming through.

Edit: Now that I think about it, you still have lots of leakage. To get the voltage up to 2V you're gonna need a bigger resistor. Or try to leave it away. The leakage your transistor has without any mods may be just right for your build.
But then again you might prefer it as it is now..

[Bret608]

I'm guessing it would have taken about 1m to get to two volts. As it is though, it sounds fantastic with 560k. I may fool around with it more later. Actually I'd like to try one with the 680k/100k bias string. Sounds like socketing the position where that resistor goes will really be the key to nailing it down in the future.

[Electric Warrior]

One of the vintage units I have here has 680k/100k. Doesn't sound much different compared to the one with 220k/47k. I built my clone with 680k/150k and it's more @#$%ed up sounding with the fuzz turned up all the way - in a good way. Here's the schematic I used:



Measures around 2V on Q3's collector as well. 

[Bret608]

Thanks, I did not know the Vox version had 1m between tone and volume. The 680k/150k sounds worth trying out for sure. I think if I did it on Madbean's board again I may keep the trimmer but go with 1m in order to allow for more different transistors for Q3. In my current build I'll try 750k or 1m when I have time to see if I can get closer to 2v. Again, it's a great-sounding fuzz; just curious about how different it will be.

[Electric Warrior]

It's not "the Vox version". It's the version they started out with. When Sola Sound tweaked the circuit, all OEM versions were affected by the changes. They made the Vox MKIII the longest, so it can have all variants of the circuit (including the 220k/47k one).

They started out with a 1M between the pots. I can't be sure what value the volume pot was. A slightly later Rotosound unit has a 470k resistor and 250k pot.

The 1000pF the early units had in the tone stack makes quite a difference as well. Give it a try!

[Bret608]

Nice! When Sola Sound sneezed, Park, Vox, Carlsbro, etc. caught cold... 

I wonder if volume pot was 500k with the 1m resistor? Not sure it matters, but that seems to be the "ratio" they had going for lack of a better word (470k resistor, 250k pot; 220k resistor, 100k pot). Anyway, this circuit is a lot of fun and I have to rein myself in a bit to not go nuts and build a couple more straight away.

[Electric Warrior]

500k would make sense, indeed. I've been thinking through this before. I'm not sure if all units with 470k limiting resistors had the 250k, but if they had, they probably only changed the values of components while keeping the output of the voltage divider the same.