Fuzz face npn and pnp

suspiciousminds

Spent the day building two types of fuzz face on printed pcb 's.
The pnp uses germanium ac128's and the npn uses bc109 silicon trannys.
I've populated the boards using the same components (except the trannys) and orientated the 2 electrlytic caps correctly on both boards.
Dpdt footswitch, no led, and just battery operated.
Now the pnp germanium sounds bloody great - vol pot works fine as does the b1k fuzz pot.
The npn silicon however is not a patch on it.
I've used the same components - pots, resistors, caps (electrolytics fitted the opposite way to the pnp), and the battery wire soldered to the board opposite as stated on the pcb.
Now the vol works ok but not as loud as the npn and the fuzz pot doesnt do much at all - its as if its set lower and doesnt change anything when turned.
Any thougts?

suspiciousminds

This is the info:
"You are biding set of 2 PCB's ( NPN and PNP)
40mm x 60mm

1.

Dallas
Fuzz Face NPN
NPN – Negative ground version of Dallas Fuzz Face using silicone NPN transistors.
Possible silicon transistors are BC108 BC109, BC109C, BC209C

2.
Fuzz Face PNP
PNP – Positive ground version of Dallas Fuzz Face.
This is original design of earliest Fuzz Face version using PNP germanium transistors.
Use germanium PNP AC 128 transistors.



C1 - 2,2 µF
C2 - 20 µF
C3 - 10 nF

R1 - 33 kΩ
R2 - 330 Ω
R3 - 8.2 kΩ
R4 - 100 kΩ

VR1 (fuzz) - B1kΩ
VR2 ( vol. ) - A500kΩ

Q1 - BC108 NPN or AC 128 PNP
Q2 - BC108 NPN or AC 128 PNP

ecc83

There are several schematics for the FF I have found. Most make little sense to me!

I would have thought signal out would come from the second collector? The battery seems to be the wrong side of the decoupling R from "normal" circuitry.

<<<Why......Can't......We.......Attach.......STUFF!???>>>

Two reasons why silicon might sound different..

1) the signal has to get to a higher Vbe before it starts to affect the bias.

2)There is NFB from the 100k to Tr1 base. Silicons have a much higher hfe therefore the loop gain is higher and distortion less.

IMHO!

Dave.

suspiciousminds

So, swapping the battery pos and neg may work?

[Deleted User]

AC128's can vary widely in gain - in my experience (especially with NOS ones) you need to try or measure a good few of them before you find something which works for you. If it's PNP you will need to get the polarity correct as well

imalone

JonathanWesson1 said:

AC128's can vary widely in gain - in my experience (especially with NOS ones) you need to try or measure a good few of them before you find something which works for you. If it's PNP you will need to get the polarity correct as well

It sounds like the NPN he has problems with, not the AC128.

suspiciousminds said:

So, swapping the battery pos and neg may work?

Not if it's already the opposite way round (negative ground) compared to the germanium circuit. If it's the same way around (positive ground) it wont work and may pop the capacitor. If it's working, just not working well, then it's probably one of the issues ecc83 points out.
Assuming you've got something like this:
http://www.beavisaudio.com/schematics/Dallas-fuzzface-npn-schematic.htm

(From values it looks like your labelling for C2 and C3 is different from this diagram, don't worry about that, it's almost certainly just a labelling issue.)
You could try increasing R4 (or put another 50k in series with it to see if it makes a difference) to decrease the negative feedback, but first check the voltages at B on both transistors, between B and E on both and at C on Q2.
The circuit is a bit funny, it's actually got the same topology as a current source http://jeelabs.org/2012/11/24/its-a-current-source/ with R4 in the way and R1 is so big it wouldn't work as a source at 9V (think it would drop more than 9V at the operating current, but depends what VBE is), so check the transistors are actually on is probably the first step.

suspiciousminds

Its strange because the germanium one i've done sounds brilliant - but the silicon is pants - the only thing done different is swapping around the neg/pos battery wires and the orientation of the two electrolytic caps.
I havent housed either yet - could it be a grounding issue?

LastMantra

IMO a schematic will only give you a rough idea. You're best to select and bias transistors by ear (There are loads of different FF schematics!). Also try a variety of input/output caps....all the other components for that matter, there's not many!

GuyBoden

LastMantra said:

IMO a schematic will only give you a rough idea. You're best to select and bias transistors by ear (There are loads of different FF schematics!). Also try a variety of input/output caps....all the other components for that matter, there's not many!

I agree, I'd keep the best sounding one and list the parts/schematic/info/voltages clearly for future builds projects.

imalone

Transistors are funny things, they have a minimum 'on' voltage across the BE junction which determines a lot about how a circuit like this (which is dependent on it) operates. Apparently Vbe for the BC108 is somewhere 0.55-0.7V. Can't find hard numbers for the AC128, but seems to be much less (maybe 0.3V). In a current mirror (without R4), you'd estimate VR1 would be dropping the Vbe voltage and therefore carrying a particular current as the operating point self-stabilises. That'd be around 0.3mA for the PNP device and 0.6mA for the NPN (except R4 is quite big, by a ratio to VR1 similar to the Hfe, so the voltage difference is going to be increased). Maybe if there was another 1k between VR and earth it'd put you into the same operating area.
When you say the silicon is pants, is it actually passing sound at a sensible volume and just not doing much to it, or is it obliterating the signal?

randomhandclaps

For BC109s, 8.2K is too high for R3 regardless off what the instructions say.  For 108 I would suggest you would be starting at around 4.7K and bearing in mind 109s could be higher gain still (though not necessarily) you are probably going to end up down nearer around 3.3K.

Do you have a multimeter to see what voltage you are getting at the collector of Q2?  This will not only tell you is there is a biasing issue, but as the gain pot seems non-responsive, if you are getting full 9 volts at the collector then you'll know you've got a ground problem.

randomhandclaps

LastMantra said:

IMO a schematic will only give you a rough idea. You're best to select and bias transistors by ear (There are loads of different FF schematics!). Also try a variety of input/output caps....all the other components for that matter, there's not many!

The thing is though if he is trying to build a Fuzz Face then the schematic isn't a rough idea, it's the plans and there really isn't loads of different Fuzz Face schematics.  Swapping the input and output caps will make no difference to his current issue.

With regards to matching transistors this is far more of an issue with germaniums than silicons due to the wide variance.  The actual FF circuit is incredibly tolerable to a multitude of silicon trannys and the difference in sound may only be minor.  The issue comes with the fact that silicons tend to have a much higher Hfe (the BC109 more than most) than germaniums, so as standard the biasing resistor on Q2 needs lowering to get the voltage down to a usable level.

LastMantra

randomhandclaps said:

LastMantra said:

IMO a schematic will only give you a rough idea. You're best to select and bias transistors by ear (There are loads of different FF schematics!). Also try a variety of input/output caps....all the other components for that matter, there's not many!

The thing is though if he is trying to build a Fuzz Face then the schematic isn't a rough idea, it's the plans and there really isn't loads of different Fuzz Face schematics.  Swapping the input and output caps will make no difference to his current issue.

With regards to matching transistors this is far more of an issue with germaniums than silicons due to the wide variance.  The actual FF circuit is incredibly tolerable to a multitude of silicon trannys and the difference in sound may only be minor.  The issue comes with the fact that silicons tend to have a much higher Hfe (the BC109 more than most) than germaniums, so as standard the biasing resistor on Q2 needs lowering to get the voltage down to a usable level.

I mean the values of the components not how they are arranged. Arbiter made more than one FF all with different component values.http://fuzzcentral.ssguitar.com/fuzzface.php

I make my FF's by replacing all the resistors with variable ones (on a breadboard) and tweaking until I get the best sound, repeating with different combinations of transistors. I also keep a bunch of caps for fine tuning the tone. The one I use at the moment has two AC128's of slightly different spec (from different manufacturers). It has different component values again.

And sounds awesome!

suspiciousminds

imalone;278262" said:

Transistors are funny things, they have a minimum 'on' voltage across the BE junction which determines a lot about how a circuit like this (which is dependent on it) operates. Apparently Vbe for the BC108 is somewhere 0.55-0.7V. Can't find hard numbers for the AC128, but seems to be much less (maybe 0.3V). In a current mirror (without R4), you'd estimate VR1 would be dropping the Vbe voltage and therefore carrying a particular current as the operating point self-stabilises. That'd be around 0.3mA for the PNP device and 0.6mA for the NPN (except R4 is quite big, by a ratio to VR1 similar to the Hfe, so the voltage difference is going to be increased). Maybe if there was another 1k between VR and earth it'd put you into the same operating area.

When you say the silicon is pants, is it actually passing sound at a sensible volume and just not doing much to it, or is it obliterating the signal?

Volume is very low - there is fuzz, but the pot has no effect.

randomhandclaps

If you asked numerous professional builders to sketch you out THE FF circuit, the one above is what they would draw. 

Altering the components in a Fuzz Face becomes counter-productive to producing an actual Fuzz Face.  If you alter every value within that same basic circuit you can get virtually any decent randomly selected pair of transistors to work.  The problem is what you end up with is not a Fuzz Face, it's just what you decide to call a Fuzz Face.  The balance of Q1 & Q2 and just as important to the distortion character as important as just being able to get a fuzzy sound out of the circuit.

The circuit above is widely recognised as the basis of the true FF circuit.  If you look at the circuit alterations on the page linked variations include increasing the value of R2 to give more output level which was not designed to cater for those who like to use a FF as an overdrive device, without altering the fundamental circuit sound.  Changing R2 obviously allows you to tune your FF.  However R1 alterations are a bit of deceptive.  In truth Dunlop altered these value merely to allow them to accommodate less specific (or maybe easier to source) transistors.  The reality is that the further you stray from 33K the less like a FF the circuit starts to sound.  If you cannot get a FF to bias where you want it by altering R2 alone then the transistors are not well matched - and believe me there is a lot more tolerance here then seller of 'paired' transistor would have you believe.  A lot of sellers sell matched pairs that they have matched on paper but have never been in a working test circuit.

LastMantra

I don't really agree, I mean the originals chose components for lots of reasons not just what sounded best. The components and values changed over the years because of this. Famously some sounded great and some not so good, or at least they varied a lot.

All the transistors I tested were within a certain range and the value changes I've made aren't massive so I would still call it a FF, just it is tweaked to (my) perfection. 

 

imalone

There is a point where things work or don't work, yes within those ranges you can tweak, but if, for example Q2 is shifting 1mA and upwards, then then output is pinned to a potential divider between R2 and R3 and there is no signal to be got.

randomhandclaps

LastMantra said:

I don't really agree, I mean the originals chose components for lots of reasons not just what sounded best. The components and values changed over the years because of this. Famously some sounded great and some not so good, or at least they varied a lot.

All the transistors I tested were within a certain range and the value changes I've made aren't massive so I would still call it a FF, just it is tweaked to (my) perfection. 

 

You don't have to agree.  You built a fuzz that you like and that's great.  The truth is that as I mentioned before an important part of the 'Fuzz Face' sound comes from maintaining a certain balance in the gain of Q1 & Q2 - R1 particularly is a massive part of this.  Vary any of these value too much and it is no longer a fuzz face circuit.  This doesn't even need to be explained, debated or proved, it's just basic knowledge.  Fuzz Faces, two transistor Tonebenders, FY-2s, half the Devi Ever range and around 50,000 other fuzz pedals marketed sound different to a FF precisely because these parameters have been tweaked or altered. 

Dunlop did change their values for the exact reason I mentioned, not to improve the sound.  They have always been notoriously inconsistent at making good FF which is the reason a large boutique and small builder markets exist charging three to four time the amount Dunlop do.  The basis of all of these expensive boutique FFs is the circuit above.  Why would people pay three to four times as much for a hand built FF when the could have a proper Dunlop one much cheaper?  Because unlike Dunlop the builders put in the effort and the math to get the correct transistors to work in the correct circuit.

A lot of people look at the Fuzz Face circuit and think it only has a few components so it must be simple.  In contrast the simplicity of the Fuzz Face makes it actually quite a complex beast to consistently make well.  Every small variant has a massive impact of the circuit.  This is why so many mass manufacturers stay away from it.  

To say varying parts to get them to fit or to suit your own tastes but to try and say it's still a Fuzz Face is needlessly grasping at straws.  If you like the fuzz you made why the hell would it matter?  It's like making a tasty shepherd's pie you love but using beef mince (because you prefer it) and then trying to argue that it's still a shepherd's pie and not a cottage pie.

LastMantra

randomhandclaps said:

LastMantra said:

I don't really agree, I mean the originals chose components for lots of reasons not just what sounded best. The components and values changed over the years because of this. Famously some sounded great and some not so good, or at least they varied a lot.

All the transistors I tested were within a certain range and the value changes I've made aren't massive so I would still call it a FF, just it is tweaked to (my) perfection. 

 

You don't have to agree.  You built a fuzz that you like and that's great.  The truth is that as I mentioned before an important part of the 'Fuzz Face' sound comes from maintaining a certain balance in the gain of Q1 & Q2 - R1 particularly is a massive part of this.  Vary any of these value too much and it is no longer a fuzz face circuit.  This doesn't even need to be explained, debated or proved, it's just basic knowledge.  Fuzz Faces, two transistor Tonebenders, FY-2s, half the Devi Ever range and around 50,000 other fuzz pedals marketed sound different to a FF precisely because these parameters have been tweaked or altered. 

Dunlop did change their values for the exact reason I mentioned, not to improve the sound.  They have always been notoriously inconsistent at making good FF which is the reason a large boutique and small builder markets exist charging three to four time the amount Dunlop do.  The basis of all of these expensive boutique FFs is the circuit above.  Why would people pay three to four times as much for a hand built FF when the could have a proper Dunlop one much cheaper?  Because unlike Dunlop the builders put in the effort and the math to get the correct transistors to work in the correct circuit.

A lot of people look at the Fuzz Face circuit and think it only has a few components so it must be simple.  In contrast the simplicity of the Fuzz Face makes it actually quite a complex beast to consistently make well.  Every small variant has a massive impact of the circuit.  This is why so many mass manufacturers stay away from it.  

To say varying parts to get them to fit or to suit your own tastes but to try and say it's still a Fuzz Face is needlessly grasping at straws.  If you like the fuzz you made why the hell would it matter?  It's like making a tasty shepherd's pie you love but using beef mince (because you prefer it) and then trying to argue that it's still a shepherd's pie and not a cottage pie.

Dude I'm not sure what you are getting at. We can see that there are many incarnations of the FF pedal. Some with different resistor values, some with different cap values, some with different transistors (germanium and silicon!), some with different pot values, different polarity. If you don't like me using the tern "Fuzz Face" to describe my pedal then fair enough. It's probably not going to stop me though, I think it is close enough to the original for it to make sense to use the term loosely. I know it's not actually a Fuzz Face, it's in a square box for a start. 

It's even got extra resistors as I've put a couple in series to get closer to the exact value I wanted. I wasn't so restricted by cost as I'm not mass-producing them.  

randomhandclaps

LastMantra said:

Dude I'm not sure what you are getting at. We can see that there are many incarnations of the FF pedal. Some with different resistor values, some with different cap values, some with different transistors (germanium and silicon!), some with different pot values, different polarity. If you don't like me using the tern "Fuzz Face" to describe my pedal then fair enough. It's probably not going to stop me though, I think it is close enough to the original for it to make sense to use the term loosely. I know it's not actually a Fuzz Face, it's in a square box for a start. 

Is there any point I said you shouldn't call it a Fuzz Face?  As I said if it makes you happy that's great.

However that is massively different to you giving advice to a guy who is trying to build a Fuzz Face, asking for help and then telling him that the circuit is just a rough idea and he can change what he likes.  Both of those statement are completely inaccurate.  He wants a working Fuzz Face, not a variant two transistor fuzz.  He brought PCBs and transistors to build two actual working Fuzz Faces.  The fact that even after my previous explanation about the amount of significant variable in a two transistor fuzz circuit you still seem hung up on the fact that 'you set out to build a fuzz face so it's a fuzz face' only demonstrates that do not understand what is happening in the circuit and why certain ranges are important for certain sounds.  You are not providing him with helpful information you are providing him with your opinion.

I suppose 'what I'm getting at' would have been clearer if rather than trying to politely highlight misinformation bring provided to the OP I had just simply said that you are wrong.

randomhandclaps

LastMantra said:

 It's even got extra resistors as I've put a couple in series to get closer to the exact value I wanted. I wasn't so restricted by cost as I'm not mass-producing them.  

#sigh#

But their not actually extra resistors are they.  They are two resistor coupled to give the value of one.  Mind you that fact that you weren't restricted by cost and spent an extra 12p on resistors goes to show you know what you are talking about!!!

My original comments on this thread were written to help the OP with their questions.  Dave, I know you can't be wrong (even when you patently are) so unless you have amazingly and instantly managed to gain the experience of building more than your own lovely fuzz and can offer them helpful advice based on knowledge rather than opinion then let's just leave this circular debate here as it's my sleepy time.