Sandspur Fuzz with Germanium Transistors?

[greypilgrim76]

Hi All,

As you may have seen in this thread, I recently ordered some MP38A transistors off eBay to build a Muffin Factory. After testing them out, I found that two of them were ideal for the MF, but the other four tested pretty well, too, and I'm wondering about using a couple of them to do an old-school style fuzz. In sorting through the fuzz options available, I noticed the Sandspur Fuzz already calls for NPN transistors, so I'm curious how much (and what kind of) tweaking would be necessary if I replaced the BC108s with MP38As, or if there's a better PCB to work with using these transistors. If it is something that would work, I'm also wondering which of the remaining four transistors would work best. Here's what I've got:

#1: .13 mA leakage, 62.3 hFE
#2: .15 mA leakage, 60.4 hFE
#3: .25 mA leakage, 71.5 hFE
#4: .19 mA leakage, 82.2 hFE

I'm not sure if there are other things I should test, or if there are other components I'd need to adjust or swap out, so I figured I'd just check and see if some of you with more experience/knowledge in this department had any advice.

Many thanks!

 

[jwin615]

Just build it stock and try any of them in Q2 position. I wouldn't bother with Ge in Q1. You'll get better performance and likely better noise using Si there. There may not be enough gain from your germs in Q1 to properly pumble Q2.
Check some other build reports and thread on this one as to biasing procedure and pot positioning.
You may find that you'll need a larger bias trimpot or a larger R3, so I would solder it with minimal solder to start, in case you need to remove it. Pins 2 and 3 on the trimpot are bridged. I think you can skip soldering pin 3 for initial testing, again, for easier potential removal.
Alternatively, if you have some test clips, don't poulate R3 and bias trimmer and go from leg 1 of the sundial to a breadboard to the collector of Q2. Work out your needed bias resistance on the breadboard and populate accordingly.
Or just breadboard it first.

 

[Vild]

I definitely agree that you should breadboard, especially since it's a simple circuit/not a lot of parts.
For the fortin fuzz (another silicon fuzz face style pedal), it uses bc550c's and has a 120k resistor going across the two transistors at the bottom. To use mp38as and get a decent amount of gain I had to switch it to 100k and I used the two highest gain ones I had that were both 80 something hfe. The sandspur however already uses a 100k for the og transistors so I'm guessing the circuit would work well already. But to help get you started with the breadboarding; start off with the 70 hfe going into the 80, then try the 60 into 80, try 80 into 60 to see why normally people go low into high, etc...
Oh and biasing seemed to work best around 4.3v if I remember correctly. I'd have to double check to be sure about the number. But what I did was I turned the bias up until there was a noticeable boost and the characteristic changed a lot in a bad way for a fuzz, and then turned it back down to just before all that. That seems to be the way to bias those by ear. And if that puts it too close to the edge of sounding bad when there's a change in temp since it's germanium then you could bias it a little lower to play it safe idk.