A while ago. Chris Stelloh sent me a Death by Audio germanium Filter to trace. The trace went pretty smootly and I sent the schematic to Chris and he made a vero layout out of it. However, this circuit requires some transistor and diode picking to work as intended. See Chris's notes below...
My notes + warnings from my build attempts: PROCEED WITH CAUTION.
1)
The transistors: DBA used MP10b NPN germanium transistors which have a
common hfe range of 20-50. The lowest gain transistors I have on hand
are MP35's around 40 hfe. And none of these worked properly. At lower
settings, the gain was sputtery and sounded misbiased. I was able to get
the circuit to work properly with some NPN Fairchild silicon dots from
Small Bear that measure at 25 hfe.
2)
The unknown silicon diodes on the mosfet gate: 1n4001-4007 worked quite
well. I had my best sounding results with 8.2v Zener diodes. the 9v
supply will cause any Zeners with lower than that rating to heat up and
pop.
And here is layout with trimpots for the collectors solve the gating issue when the Gain control is turned counter clockwise. I also ditched the internal trimpot in favour for an external volume control.
Awesome! I never thought I'd see the day where I could use my MP10b's :-))
ReplyDeleteCheers Chris and Anders, this worked first time for me!! Used a 9v1 zener and still going through a bunch of transistors but it sounds good - love the filter knob
ReplyDeleteGetting better results with AC140's with an hfe of 28, could do with being a little lower I think.
DeleteThat's so awesome, Ian - I was a little nervous that we got something wrong - and I agree on the hfe needing to be as low as you can get it. The filter is basically the modded BMP tone section that DBA uses in the Fuzz War. I love how it has the extra mid bump that only happens right around 11-1 o'clock on the sweep of the pot.
DeleteJust tried with some MP10b's with hfe's of 26 and I'm getting the spluttery miss biased sound. Going to see if I have any lower.
ReplyDeleteMine is pretty gated too, especially on low gain. There must be a magic number for the hfe, I don't have anything lower than 28.
DeleteWhat diodes did you use Ben?
Hi Ian
DeleteI used 9.1v, but I do have some lower so I might try them too. I tried with some 10b’s at hfe’s of 19 and 22. Same result. Also tried some MP 35’s in the mid teens with similar results.
Cheers
Ben
Cracked it Ben, just change the 68K and 82K resistors (connecting to the collectors of both transistors) to 100K trimmers. Set the Gain to Clean and dial in - it goes from real nice clean to full on fuzz. I'm still using my AC140's, hfe 28.
DeleteSo I"ve been thinking about this a lot, Ian - I don't suppose there is any way to measure what you have those trimpots set at, is there? Because I'm guessing that the amount of resistance each one is providing between the 9v supply and the collectors of those transistors is LESS resistance than 68k and 82k. My original intuition was that DBA used really low gain transistors but I'm now wondering if higher, not lower gain BJTs are what is necessary. Do you follow me on this? If i had any schooling whatsoever in electrical engineering I could probably make short work of this problem... it could be that the sputtering/gain issues at low settings is coming from those stock resistor values starving the voltage to the collectors. I'm definitely going to have to build up this circuit again and play with some higher gain transistor values and also try replacing the resistors between the supply and the collectors with trimpots.
DeleteHi Chris, I will measure the values and get back to you. I was wondering if DBA use the same resistor values on all the original pedals or they measure, and use different values. It would be a long winded way to do things but DBA pedals are not always the easiest to get right.
DeleteI must say, with the trimpots, it sounds really good - pretty much spot on with the original. The way it goes from very clean to full fuzz is great, and that filter is really nice.
Did anyone measure the voltages on the trace?
DeleteThanks Ian. Trimpots did the job :-))
DeleteIan - as far as I know, DBA is using the same resistor values on these because I opened the pedal up before I sent it to Anders, and the component values are all printed on the board next to the pads where they are soldered. Of course, I didn't actually break out the magnifying glass to see if those component values matched what was printed on the pcb - the bits are all SMD on this one.
DeleteMakes sense, I can't imagine them messing around measuring resistor values to match transistors. Either way, they are always fun to build and a challenge to get them sounding like the real deal. A huge thanks Chris for sharing this layout (and Anders)
DeleteJust measured where I set the trimmers. 9.5k for q1 and 23.6k for q2 using MP10b's in the 24 hfe range.
DeleteI'll check mine out tomorrow, see what the difference is. I must say, I am liking this pedal.
DeleteHi Ian did you get the measurements? Voltages would be especially helpful too.
DeleteSorry Chris, I'll do this tomorrow
DeleteThis comment has been removed by the author.
DeleteI'd love to take a shot at this. Anybody have any recommends for a suitable replacement that I might source stateside? I currently have nothing close to the ideal gain range.
ReplyDeleteHey Tom - from what I've gathered in the discussion so far, it's that my original idea that the transistors were super low gain might be incorrect. It looks like you might be able to get the proper results with NPN germanium in a higher gain range - like 80-90 maybe? OR you can cut down the resistor values going to the collectors of those transistors. I've got a pile of MP38s that are in the 40-70 range and I'm thinking that halving the 68k to around 33k and cutting 82k down to 47k might work. Let me know if you don't have any transistors that fit the bill. I'll be happy to send you a few (I'm in CA) - you can hit me up at capitalistdeathculteffects at the gmales dot coms
DeleteThanks Chris, I really appreciate that offer. I'm going to try Anders' updated layout with what I have first. Thanks for making this circuit available. Appreciate the work. And thanks for the new layout Anders!
Deletesure thing, Tom! the layout Anders just posted with the trimmers should do the trick!
DeleteI have a layout with trimpots for the collectors. I will upload it tonight
ReplyDeleteI'm currently mid build in one. I did alter the layout myself to have trim pots for the Q2 and Q3 collectors. I'm also going to piggyback transistors together too to get the lower hfe.
ReplyDeleteLayout with bias trimpots is up. Still unverified though...
ReplyDeleteI can verify that the fuzzhead version works. Sounds like DBA to me. Thanks guys. The biasing is touchy. I rolled off the gain and found a place where it stopped gating by ear. Clean to fuzzy with a pretty dramatic filter.
DeleteThanks Tom!
Delete
ReplyDeleteHi all,
Got a working one like the video demos with piggybacked 2n3904s instead of germanium! By doing that Q2 and Q3 have 25hfe running through them. Biased with trimpots to ear as said previously and all good! Q2 makes affected the gating, Q3 doesn’t seem to do as much but I tested it with no gain, full gain and low/high filter and think I found the sweet spot.
How to make 2n3904s work! I have two 2n3904 trannies in each Q slot, so it uses 4 in total. For each Q, one transistor goes in CBE as normal, the other transistor connects to the first one like; C not connected, B to B, E to a resistor to the others E). The resistor value will change the hfe you actually get so you need to experiment (I breadboarded them with a trimmer and DMM to get my 25hfe). There’s also a 100p cap across the C and B of the fully connected transistor to simulate miller effect.
If that doesn’t make sense google piggybacked transistors and you’ll find forums and stripboard designs to make them!
And voltages…this is what my pedal has as of the moment and I would assume may help others. I’m fairly confident it’s sounds like the demos but as I play it more I may tweak it (we’ll see).
BS170
D 4.55
G 2.28
S 0
Q2
C 2.07 (no gain) 2.25 (full gain)
B 0.8
E 0.16
Q3
C 8.26 (no gain) 5.16 (full gain)
B 0.58 (no gain) 0.43 (full gain)
E 0.02 (no gain) 0.07 (full gain)
Hi there. Is 390 ohm resistor could do the job ? Why do they use 392 ohm ?!! Thanks wonderful team
ReplyDeleteI used a 390ohm and it sounds great
DeleteThanks a lot . I'll try this one
DeleteDamn! more DBA stuff, building this for sure... front end of this thing is basically a Zvex SHO. Those 2 diodes serve the purpose of protecting the MOSFET gate from static discharge. 9.1v Zener is typically used here...
ReplyDeleteOrdered a bunch of MP10b from here, along with a bunch of other bits...
https://nextfive.pro/products/mp10-mp10a-mp10b-germanium-npn-transistors
My guess is screening a bunch of these things for leakage will yield best results, glad to have an Atlas laying around.
Leaky Germs kill your bias, DBA is most likely screening for near "0" leakage to use fixed value resistors in their production models.
DeleteCould always strap a germ diode to the B - E junction of the transistor. Anode end of the diode would go to Base, with cathode to emitter for a NPN BJT... Burns used to do this with there Buzzaround to get them out the door....
Diode "cathode" to base for NPN.... whoops
Delete