This Week on the Breadboard: The Germanium Filter

Chuck D. Bones

Chuck D. Bones

Circuit Wizard
It's a two-fer this week!
This is the first Death By Audio circuit I've breadboarded. I've always thought their stuff was a little "out there" and this one is no exception. Before I built the breadboard, I listened to a demo on YaoTaob. Sounds pretty good. Definitely worth investigating.

The first schematic is one I found on the web. It will sounds like ass if you build it per that schematic because the biasing will be all wrong. In the production pedals, Q2 and Q3 are installed backwards (emitter & collector swapped) in traditional DBA style. There is a tear-down video on YT and if you can get past the incessant blabbing, the guy flips the board over at 8:09 and the trained eye can spot a trace that confirms E & C Are reversed on at least one of the Ge transistors. I ran some sims, and then built my breadboard per the schematic below, with the correct orientation for Q2 & Q3. I also changed the reverse power protection to a series diode and replaced the Volume Trim with a pot with a knob. I deleted the superfluous 1M resistor between pin 2 of the FILTER pot & GND and subbed in a pair of randomly selected MP38A transistors for Q2 & Q3. The production pedal's audio path is faithfully reproduced on the breadboard. I did not need to tweak the bias.

Feedback thru R3 causes the input impedance to be fairly low, around 25K. This makes the circuit very responsive to the guitar's knobs. D1 & D2 provide overvoltage and ESD protection for Q1.
dba germanium filter schematic (typo).png Germanium Filter cb mod v1.0.png

Knobs (L-R): LEVEL - FILTER - GAIN
Q1 on the far right, Q3 on the left.
Germanium Filter cb mod v1.0 breadboard 02.jpg

The FILTER knob is something else! Large range of tones. This is the end-to-end freq response with GAIN at 5 (noon) and FILTER at 0 (green), 5 (blue) & 10 (red). Below noon, the tone becomes fat and at zero is thick & heavy. Around noon, the tone is balanced and clear. Above noon, it gets bright and thins out at 10. DBA really nailed the EQ circuit.

1760237940890.png
 
The thought hadn't occurred to me. The breadboard is quiet, so I think I'll leave Q1 as-is. Good question though. In this case, I believe DBA did the right thing with Q1.
 
Just watched the youtube video and ayyyy... Another must build as I watch my PTP Chuck D Bones project stack grow.

I really like that style of guitar/player oriented interface, like the corroder.
 
Last edited:
I went back and rebuilt this breadboard with some different transistors. I don't know exactly how this is supposed to sound over the range of knob settings. Yes, there are demos, but they don't run the full gamut of knob combinations. The circuit performance is very sensitive to the two GE trannys, and not in a predictable way. Right now I'm running a pair or randomly selected 104NU71's. Bottom line: if you build this, socket Q2 & Q3.
 
I'm liking this circuit! Getting the transistors right is key. Anywhere around +4.5V is good for Q2-E. Q3, you really need to try a few different ones. The one I have now has Q3-E around +2.8V, but it's not just about bias. It has to sound good, and not be too compressed at low gain settings. Q2 & Q3 are both MP38As, but I had to try a dozen or so before I was satisfied. Please don't ask me about HFE or leakage because knowing that won't help in this circuit. I added a 1nF cap from VOLUME-3 to GND to tame some of the top-end harshness.

Here's the FILTER knob's freq response from Q2-E to Q3-E: For most of the range, there is a mid scoop.
1761703761215.png
 
Last edited:
Chuck D. Bones said:
Q3, you really need to try a few different ones. The one I have now has Q3-C around +2.8V, but it's not just about bias. It has to sound good, and not be too compressed at low gain settings. Q2 & Q3 are both MP38As, but I had to try a dozen or so before I was satisfied. Please don't ask me about HFE or leakage because knowing that won't help in this circuit.
Click to expand...
I’m curious about this. What would make one MP38A sound good but not the others if it isn’t the bias/hFE/leakage? Is it hfe? Collector-base capacitance? Something else?
 
It might be e-b leakage & reverse HFE. The other thing it could be is the shape of the emitter I/V curves. When running transistors upside-down, it's all voodoo. Particularly so with Germanium in my estimation.

My gauge of transistor goodness in this circuit is the most gain & volume while being able to play clean (or nearly so) with GAIN between 0 and 1 and FILTER at 5. The other deciding factor is the tone when FILTER is at zero and GAIN is at 5 or higher. I want a nice fat tone with not too much dissonance (nasty harmonics).
 
Chuck D. Bones said:
Anywhere around +4.5V is good for Q2-C. Q3, you really need to try a few different ones. The one I have now has Q3-C around +2.8V
View attachment 105414
Click to expand...
do you mean collector or emitter?
apologies, the references get a bit confusing when the transistor is flipped.

for context,
i’ve currently got your schematic breadboarded with some MP11A NPN Ge.
with the collectors hooked up to resistors (390R) to ground, im seeing about 0.03V at Q2C and Q3C.
with the emitters hooked up to 68K and 82K (to +9V supply), i’m seeing about 3.7V and 3.3V at each emitter.
(this is just with the first couple i grabbed out the bag)
IMG_0754.jpeg
 
comradehoser said:
Wow, I was not expecting such a preponderance of SMD and such a clean board. I always thought their stuff would be grimy and hand soldered with rat fur. But, maybe that's just the image they like to project.
Click to expand...
Maybe with a little beer and cigarette smoke build up like mixing boards from dive bars in the 90s.
 
owlexifry said:
do you mean collector or emitter?
apologies, the references get a bit confusing when the transistor is flipped.
Click to expand...
I'm the one who should be apologizing.
I meant E, not C. I confused myself with the E-C swapping. Post #10 has been corrected.

owlexifry said:
with the emitters hooked up to 68K and 82K (to +9V supply), i’m seeing about 3.7V and 3.3V at each emitter.
Click to expand...
That's a good start, give it a listen. Then look for a transistor that makes Q2-E closer to 4.5V.
 
I had one of these requested a couple months ago. Now I can't remember if I flipped Q2 and Q3 or not. I know my transistors were super low gain soviet. maybe MP10, maybe MP38, I cant remember. It was in the same box going into a fuzz war and an interstellar overdriver. what I DO remember was putting C5 on a switch labeled with an exclamation point for both the Ge Filter and the Fuzz War. they both use that cap to jump the tonestack and MAN does it get shrieky. DBA isn't known for subtlety, but I figured a couple on/off mini toggles to double the number of sounds was a good addition.
 
You must log in or register to reply here.
Back
Top