Big Muff with Germanium Transistors - hFE and bias voltage

[Passaloutre]

So I'm going to be building a Big Muff based on the Hoof, with germanium transistors in the second and third stages. I've built one before, but I'm building one again.

Here is the schematic, for reference:



I've got a stash of NPN germaniums (2N1304), and I built RG's germanium transistor tester to sort out the gains. I picked out two with high gain (hfe = 105 and 120) and low leakage (~100 ua).

Then I built a mock up gain stage on the breadboard to make sure these would bias correctly. So looking at the schematic above I've got:

- 15k from collector to 9V,

- 100R from emitter to GND

- 100k from base to GND

- 470k from base to collector

I put one of these transistors in the circuit, measure the voltage at the collector = 2.9VDC. I was expecting 4.5ish. I put a 2n3904 in the circuit just to make sure I haven't messed up building it, and it biases around 4.5VDC. I put one of the germanium transistors that failed the gain test (hfe = 71), but still with relatively low leakage and it biases around 4VDC.

Now, luckily I had socketed the transistors in my older Hoof build, so I decided to (A) see how my new transistors perform in a known-good circuit, and (B) measure the gain on the transistors I had originally put in that circuit.

A - the new high-gain germaniums (2N1304) in the working pedal sound fine. They still bias at less than 3VDC, but they sound every bit as good as the transistors that biased around 4.5VDC. I think my answer is right here: it doesn't effing matter if they sound good. But I'm trying to learn something from this, so hear me out.

B - the trannies (NTE101) I had put in the Hoof four years ago, in RG's tester, measured hfe around 20-30. But they sound great. Could the gain of the transistors possibly matter so little in the Big Muff circuit?

So I guess the final question is, when building the new circuit, should I favor the transistors that have the highest gain, or should I go for the ones that bias correctly?

[pinkjimiphoton]

yes, the gain of the transistors can matter that little in a big muff circuit.
to me, the lower gain, the better the sound. it gets much more liquid and like a singing cross between a dirty compressor and a smooth, truly violin like sustain. you don't need to crank the sustain as much, and its way more liquid... and reacts to the volume knob, too, on the guitar.
hfe's under 100 sound great.... much closer to the original ones i got out of creem by mail order than the modern variants i hear, big muffs are supposed to sing, not necessarily be brutal.

got a clip we can hear?

[antonis]

when building the new circuit, should I favor the transistors that have the highest gain, or should I go for the ones that bias correctly?

BOTH...!! 

High gain transistors help to form stable bias circuit with also high(er) input impedance..

On the other hand, what exactly is "correct bias"..?? 
(for faithful amplification it might be designated as: Vc = Vcc+Ve/2 - but for a heavy distortion circuit..??)

[thermionix]

to me, the lower gain, the better the sound. it gets much more liquid and like a singing cross between a dirty compressor and a smooth, truly violin like sustain. you don't need to crank the sustain as much, and its way more liquid... and reacts to the volume knob, too, on the guitar.
hfe's under 100 sound great.... much closer to the original ones i got out of creem by mail order than the modern variants i hear, big muffs are supposed to sing, not necessarily be brutal.

So now I gotta build another BMP?  Great.

What modern Si NPN do you recommend for hFEs under 100?

[Passaloutre]

when building the new circuit, should I favor the transistors that have the highest gain, or should I go for the ones that bias correctly?

BOTH...!! 

High gain transistors help to form stable bias circuit with also high(er) input impedance..

On the other hand, what exactly is "correct bias"..?? 
(for faithful amplification it might be designated as: Vc = Vcc+Ve/2 - but for a heavy distortion circuit..??)

Well, with a silicon transistor in there, it biases right at 4.5. I guess what I'm trying to figure out is why do my low leakage (high gain) germaniums drop the bias voltage more than the high leakage gemaniums?

[Passaloutre]

to me, the lower gain, the better the sound. it gets much more liquid and like a singing cross between a dirty compressor and a smooth, truly violin like sustain. you don't need to crank the sustain as much, and its way more liquid... and reacts to the volume knob, too, on the guitar.
hfe's under 100 sound great.... much closer to the original ones i got out of creem by mail order than the modern variants i hear, big muffs are supposed to sing, not necessarily be brutal.

So now I gotta build another BMP?  Great.

What modern Si NPN do you recommend for hFEs under 100?

I've got some 2N3904 that are measuring right around 100-120. They sound very good in this circuit.

[pinkjimiphoton]

to me, the lower gain, the better the sound. it gets much more liquid and like a singing cross between a dirty compressor and a smooth, truly violin like sustain. you don't need to crank the sustain as much, and its way more liquid... and reacts to the volume knob, too, on the guitar.
hfe's under 100 sound great.... much closer to the original ones i got out of creem by mail order than the modern variants i hear, big muffs are supposed to sing, not necessarily be brutal.

So now I gotta build another BMP?  Great.

What modern Si NPN do you recommend for hFEs under 100?

WELLL.... i was talking about GE, but 3904's. 2222's, bc108s i just socket and plug til i feel a little wiggle somewhere.
i tend to salvage a lot of parts out of old organs n things so there's tonnes of q's to choose from.

but anyways, with the ge, it can sound like ass or sound great. i forget which stages benefit the most cuz its been a while since the last one i built like that.

but low gain ge works great. high gain ge i don't trust whatever it reads at, i've yet to find a truly high gain ge q with no leakage. the leakage is part of how they seem to bias, and why they're so damn temperature and environmental sensitive.

i generally just do sockets til i like it, then solder one of the legs so the suckers don't fall out like they always seem to do at the worst time

[antonis]

I guess what I'm trying to figure out is why do my low leakage (high gain) germaniums drop the bias voltage more than the high leakage gemaniums?

Perhaps due to feedback bias circuit and - maybe - due to different Vbe voltage drops..
(if there is such a Vbe difference between low & high leakage Ge transistors..)

edit: Jim, 2N5088s are my ultimate choice for Big Muff style circuits.. 

[digi2t]

What modern Si NPN do you recommend for hFEs under 100?

2N2218

They generally measure in the 60 to 95 range, with the occasional one just cracking over 100.

If you're looking for a PNP that's in that range, then look at the 2N2904. Same gain deal as the 2N2218. Actually, they would make a nice complimentary pair.

[MaxPower]

What would happen if you used germanium diodes  with germanium transistors in a bmp circuit?

I've built the bmp overdrive (Cornish G2) and really like it......

[Passaloutre]

What would happen if you used germanium diodes  with germanium transistors in a bmp circuit?

I've built the bmp overdrive (Cornish G2) and really like it......

I think it would be super compressed. This is why I like the LEDs of the Hoof, they really "open up" the sound by reducing compression.

[thermionix]

Wow, today I breadboarded a BMP with low-gain transistors.  Silicon PNPs with hFEs in the upper 100s.  I used the same circuit as my completed 2N5088 build (the Kit Rae supposed-Gilmour ram's head), and it is a completely different pedal.  Obviously much lower gain, and the tone control range is cartoonishly huge.  It sounds great with the sustain maxed and the tone at noon.  The guitar volume cleanup is outstanding, far better than I thought possible with a Si fuzz.  I think I'm gonna hunt down some NPNs with slightly higher hFEs, and replace the board in my current one.

[Belanger]

2N3303 among so many others it's not funny, I've always found it weird when people complain there having a hard time finding low gain silicon's. It seems more like their just to lazy to do the research ( not meaning you )   

From the posts I've seen of vintage rams head big muffs and others the hfe tested in the high 200's. I'll try to locate a few of the posts, i really don't think electro harmonix debated about transistors half as much as we do, perhaps that's why so many vintage muffs are lemons while others are like the holy grail in a box. Wild tolerances of the vintage components used of course probably played more of a role in that then transistors    That being said I've found anything between 300-550 sounds great

to me, the lower gain, the better the sound. it gets much more liquid and like a singing cross between a dirty compressor and a smooth, truly violin like sustain. you don't need to crank the sustain as much, and its way more liquid... and reacts to the volume knob, too, on the guitar.
hfe's under 100 sound great.... much closer to the original ones i got out of creem by mail order than the modern variants i hear, big muffs are supposed to sing, not necessarily be brutal.

So now I gotta build another BMP?  Great.

What modern Si NPN do you recommend for hFEs under 100?

[thermionix]

Well I dug around my stash and tried some slightly higher gain NPNs (approximately 190-250, in ascending order).  Some oddball mofos too.  They sounded great on the breadboard, just enough gain to "smooth over" like a BMP should.  The PNPs I tried earlier did sound really cool, especially at max Sustain, but didn't get ALL the way to Muff Land, if you know what I mean.  Also, I don't know just how much difference it makes, but the "Gilmour" version has 100r emitter resistors, many others have 120r, so maybe this version wants lower hFEs than others.  Dunno, I still probably went lower than most folks would though.

Either way, I dug it enough to go for it.  After some debating and hesitating, I decided to just replace the transistors on my existing board rather than build a new one.  Not an inviting task on a crowded pad-per-hole perfboard, but I managed to pull it off.  Sounds great, and works better with my 3-way toggle (stock, tone bypass, flat mids) than the 2N5088 setup did.  Main thing, the mud is gone, but I still get the fuzz and sustain.  Pretty cool!

[Belanger]

The emitter resistor will set the gain of that particular gain stage ( it's not the only way to alter the gain of any said gain stage however it's one of the most common ways of doing so)

So if you raise the resistor the gain will decrease, You'd be surprised at how much difference it can make.  However 100r and 120r I really wouldn't suspect to hear much of a difference as a lot of resistors they used had a 20% tolerance. Meaning 100r could very well test close to 120r and vice versa


It's funny how often you'll see layouts for vintage based effects where people are constantly recommending the 2N5088 in place of literally everything. A lot of these effects used transistors with gains no where near the 2N5088. Some circuits it makes a very noticeable difference, really not as noticeable with a big muff though

Well I dug around my stash and tried some slightly higher gain NPNs (approximately 190-250, in ascending order).  Some oddball mofos too.  They sounded great on the breadboard, just enough gain to "smooth over" like a BMP should.  The PNPs I tried earlier did sound really cool, especially at max Sustain, but didn't get ALL the way to Muff Land, if you know what I mean.  Also, I don't know just how much difference it makes, but the "Gilmour" version has 100r emitter resistors, many others have 120r, so maybe this version wants lower hFEs than others.  Dunno, I still probably went lower than most folks would though.

Either way, I dug it enough to go for it.  After some debating and hesitating, I decided to just replace the transistors on my existing board rather than build a new one.  Not an inviting task on a crowded pad-per-hole perfboard, but I managed to pull it off.  Sounds great, and works better with my 3-way toggle (stock, tone bypass, flat mids) than the 2N5088 setup did.  Main thing, the mud is gone, but I still get the fuzz and sustain.  Pretty cool!

[antonis]

The emitter resistor will set the gain of that particular gain stage ( it's not the only way to alter the gain of any said gain stage however it's one of the most common ways of doing so.

So if you raise the resistor the gain will decrease

Not exactly..
(that should work in some other bias arrangement..)

Big Muff's individual gain stage (other than last recovery stage) should be considered as the ratio of Collector to Base feedback resistor and Input resistor..
(e.g. 470k/39k for the first stage)

Emitter resistors are of low value and should be considered of no significance other than of "some" NFB stability..

P.S.
Even if specific Emitter resistors could set voltage gain, their low value should be added to intrinsic emitter resistance which has a significantly high value for specific working current (about 80R) compared to 100R Emitter resistors, so a 20% raise (120R from 100R) should be actually a 11% effective Emitter resistance raise (220R from 180R).. 

[MaxPower]

Has anyone tried placing a pot (as a variable resistor) in series with the diodes? Has a greater effect on the third stage than on the second.

[pinkjimiphoton]

What would happen if you used germanium diodes  with germanium transistors in a bmp circuit?

I've built the bmp overdrive (Cornish G2) and really like it......

i built one, with an added mid cut/boost pot, all ge. let "analog" tom try it. his opinion was that tho the bmp was hiding in there, it was a significantly different sounding circuit, and he "couldn't get the real big muff dialed in" which kinda makes me question his ears 

i guess he never heard or played thru any of the good original ones

[pinkjimiphoton]

Wow, today I breadboarded a BMP with low-gain transistors.  Silicon PNPs with hFEs in the upper 100s.  I used the same circuit as my completed 2N5088 build (the Kit Rae supposed-Gilmour ram's head), and it is a completely different pedal.  Obviously much lower gain, and the tone control range is cartoonishly huge.  It sounds great with the sustain maxed and the tone at noon.  The guitar volume cleanup is outstanding, far better than I thought possible with a Si fuzz.  I think I'm gonna hunt down some NPNs with slightly higher hFEs, and replace the board in my current one.

there ya go. thats the  REAL, TRUE classic big muff sound. dead battery. seriously. turn your guitar almost off, and you'll find some really crystaline compressed clean you'd NEVER expect from a bmp. violin like endless sustain. for real. mmmmm mmmm good

congratulations,

you have found the secret message

[pinkjimiphoton]

The emitter resistor will set the gain of that particular gain stage ( it's not the only way to alter the gain of any said gain stage however it's one of the most common ways of doing so)

So if you raise the resistor the gain will decrease, You'd be surprised at how much difference it can make.  However 100r and 120r I really wouldn't suspect to hear much of a difference as a lot of resistors they used had a 20% tolerance. Meaning 100r could very well test close to 120r and vice versa

yep, thats how ya set the gain. the biasing is DIFFERENT than setting the gain, tho a lot of people use them pretty interchangeably. an e resistor does affect the bias, but in a different way than a c resistor would... the e sets the gain of the q, the c sets the operating point of the q.

that said... i find small e resistors can have a HUGE impact on tone and response.. 120 and 100r are noticeable in this case. 47r and 10r are noticeable in side by side comparisons.

my schizoid face project if memory serves was using like, 100r e resistors on q1, and 22 on q2.  i'd have to look, but i played with a lot of values before i settled on those ones giving me the best response for the circuit. huge impact on stability.. i think any semiconductor probably works better between 60-80% of its power range than all the way up sometimes.

i just checked, the schizoid used 220r on q1, and 47r/1k pot/100r to ground on q2. the transistors i was using were mpsa06, all around 215 hfe. these particular resistor values were the closest i could get a/b-ing it with my DA FF. made a big diff in the response of the circuit, very very close to ge.

i found noticeable differences all the way down to about 22r, in the sound and response.