Fuzz Face No. 3: Sort of works but need help with new problem

[Rick899]

This is the third fuzz face I have taken from a breadboard to a final soldered version.  This one allows the clean guitar to come through on bypass and when I switch it on I get minimal fuzz effect.  It has a 10k trim pot with a 1k resistor which allowed perfect adjustment to -4.5V. I am using two  2n3906 transistors from Radio shack.  I changed the 470R resistor to a 330R because of the use of the 2n3906 transistors. 

One problem is that the B1K fuzz pot adjusts the volume but so does the A500K volume pot. The other problem as mentioned is that the fuzz is very minimal.   A third problem is that it is noisy but I understand that if I add a 100-200 pf cap across Q1  or Q2 B-C, the noise problem might be fixed. 

Any suggestions on what could be causing the first two problems mentioned above?   

A couple of pics of mine are attached as well as a couple of pics of the fuzz face on which I based the one I built. That one was posted by "RouseJeremy" on another site and it looked so artfully crafted  that I abandoned the other design I was using and built that one by counting holes from the numbers and letters on his pics. It worked (with the problems described) on the first try.  (I hope it is not improper to post his pics here as long as I give him credit for his work as I have done).

[LucifersTrip]

This is the third fuzz face I have taken from a breadboard to a final soldered version.  This one allows the clean guitar to come through on bypass and when I switch it on I get minimal fuzz effect.  It has a 10k trim pot with a 1k resistor which allowed perfect adjustment to -4.5V. I am using two  2n3906 transistors from Radio shack.  I changed the 470R resistor to a 330R because of the use of the 2n3906 transistors. 

One problem is that the B1K fuzz pot adjusts the volume but so does the A500K volume pot. The other problem as mentioned is that the fuzz is very minimal.   A third problem is that it is noisy but I understand that if I add a 100-200 pf cap across Q1  or Q2 B-C, the noise problem might be fixed. 

Any suggestions on what could be causing the first two problems mentioned above?   

Did this one work how it should on the breadboard?  What are all (EBC) the voltage readings for the 2 transistors?

[Yazoo]

I have had several goes at Fuzz Faces using AC128 transistors. As you did, I used a trim pot to adjust the voltage to the second transistor to -4.5 volts and it did not sound good. I then just tried adjusting the voltage by listening to the effect until I got what I thought was a good sound and in the end the voltage was around - 7 volts so I think this is the best way to go. The -4.5 volts thing can be misleading.

[Rick899]

Luciferstrip:

Thanks for your continued help with this fuzz problem.  I have been working from the same breadboard "model" as the other ones that didn't work.  The breadboard model worked from the first try and worked much better  after you explained how to add a pot to adjust the bias voltage except for the noise which you suggested could be fixed with the 100-200  pf cap.  Because of the way breadboards are designed I didn't try to lay it out on the final board the way it is laid out on the breadboard.

Because I had such a problem with the other layout I found on turretboard.com I abandoned that layout and used this other one that I found. I did change it a bit by adding the trim pot to adjust the bias voltage. I am still using the connections contained on the turretboard.com layout. Maybe that is where the problem  is.  Please confirm whether they are correct. They are as follows:

B1K #2   to  negative side of 22 uf cap
B1k #3 and Q2E  to  100k resistor
B1K  #1  to  Board ground
A500K #3  to .01 uf cap
A500K #2  to out  (#9  of switch)
A500K  #1  to  board ground
33k resistor  to  Q2B  and  Q1C
Positive side of 22uf cap to Q1E and board ground
33K and 330R resistors  to  negative 9V (pin of DC Jack)
330R and .01 uf cap and 8K2 resistor(replaced by trim pot/1k resistor combo)   to  each other and nothing else
8k2 resistor(etc.)  to  Q2C
100k resistor and negative side of 2.2 uf cap  to   Q1B
Positive side of 2.2 uf cap to  In ( #8 of switch)
Positive side of 22uf cap  to board ground       (All the other connections, i.e., to switch, jacks, etc.) are from the GGG wiring diagram (the one with 3 wires to the pin of the DC Jack)  for PNP circuits)

Obviously this layout is not so problematic because I got sound on bypass and with effect switched on (on the other one I never got this far) on the first try.  But now this new problem: the fuzz pot changes volume but not fuzz. I still don't understand why I could not get the other one to work.  Maybe I had a bad component. On this one I used almost all different components, i.e., caps, resistors, transistors, only the .01 cap was reused. 

I will take the voltage (EBC) readings and post them later.

[Solidhex]

You might need to simplify the way you are looking at the circuit. I can't tell what's going on with the trace side of your board. Think of the fuzz face circuit as having 11 "nodes" or places where the components meet. Here is a list of them and what components are connected in each one. This is for your standard value PNP version:
1) input signal wire, positive end of 2.2uf cap
2) negative end of 2.2uf cap, Q1 Base, 100K resistor
3) Q1 collector, Q2 base, 33K resistor
4) Q2 collector, 8.2K resistor
5) Q2 emitter, 100K resistor, Lug 3 of 1K fuzz pot
6) negative end of 22uf cap, Lug 2 of 1K fuzz pot
7) 470ohm resistor, 8.2K resistor, .01uf cap
.01uf cap, Lug 3 of 500K volume pot
9) negative battery terminal, 33K resistor, 470ohm resistor
10) input jack ground lug, input signal ground, output signal ground, Q1 emitter, Lug 1 of 1K fuzz pot, positive end of 22uf cap, Lug 1 of 500K volume pot
11) input jack switched lug, positive battery terminal
  This is sort of the simplest way of looking at it and helps to cut down on redundancies. Make sure all that stuff is connected to each other in that way and it should work. If you're seeing normal voltages but don't have enough fuzz its probably your fuzz knob that is wired wrong or a bad 22uf cap

[Rick899]

Solidhex:

Thanks  for posting the connections for the 11 "nodes". They were very helpful. I compared my actual connections  one by one to your 11 nodes. They were all okay through number 9,  but I  have several questions about the terminology you use to make sure I got the connections in Nos. 10 and 11 correct.

1.   Where do you put lug #2  of the A500K pot ?  I have it going to lug no. 9 of the switch. and Lug no. 6 of the switch going to the output jack tip.
2.  In your number 10, By "input jack ground lug"  do you mean  the input jack sleeve?
3. In your number 10,  what do you mean by "input signal ground"?
4.  In your number 10, what do you mean by "output signal ground"?
5. In your number 11, what do you mean by  "input jack switched lug"?      I have the positive battery terminal connected to the DC jack sleeve shunt and the DC jack sleeve connected to the input jack ring.
I am going to replace the 22 uf electrolytic cap just to be sure.
I am using the GGG wiring diagram posted below:

   

[Solidhex]

This node thing is meant to be the simplest way of listing all the parts that need to be connected for the circuit to work. I left out the bypass switching. it would just confuse the issue. Ignore the switching and just assume that the effect is engaged.
  Both the input jack and output jack have two conductors. The input signal and its ground, and the output signal and its ground. Tip for signal, sleeve for ground. The in and out signals have to be connected to the circuit and both their respective grounds have to be connected to both the circuit's ground and each other. The circuit won't work if either of the input or output grounds aren't connected that way.
  I should have considered a 12th node for Lug 2 of the volume control and the output signal wire!

[LucifersTrip]

note: i typed this as Solidhex replied, so there may be some redundancy

I  have several questions about the terminology you use to make sure I got the connections in Nos. 10 and 11 correct.

1.   Where do you put lug #2  of the A500K pot ?

to the tip of the output jack

2.  In your number 10, By "input jack ground lug"  do you mean  the input jack sleeve?
3. In your number 10,  what do you mean by "input signal ground"?
4.  In your number 10, what do you mean by "output signal ground"?
5. In your number 11, what do you mean by  "input jack switched lug"?      

I can see how that would be a by confusing. I don't want to use a different terminology so I'll let Solidhex explain.

Just note that the input jack has sleeve, ring & tip and the output jack has sleeve & tip. Both sleeves are grounded. The
ring & sleeve of input are connected when the guitar jack is plugged into the input jack (a switch). The DC jack also has a switch which connects the ring to + battery when the DC jack is removed.

I am going to replace the 22 uf electrolytic cap just to be sure.

it is unlikely the problem unless you used a pulled vintage one, but do try it...

I really believe it is easier, you can learn more and troubleshoot easier when using a schematic and doing a perf, instead of a layout.

Yes, I know I built this on the "wrong" side of the board, but notice how I can basically superimpose the schematic on top of the perf build. It makes troubleshooting insanely easy:
http://www.luciferstrip.com/fuzz/colorsound-oneknob-ge-perf.jpg
http://www.luciferstrip.com/fuzz/colorsound-oneknob-schematic.jpg

good luck

[Rick899]

Luciferstrip

I double checked all the connections and I think I got the wiring correct because the circuit works both in by pass and with the effect switched on.  But it seems I got better fuzz on the one I built on the breadboard.  This brings us back to the EBC voltage readings.

I got the voltage reading s for the 2n3906 transistors. I used the Germanium transistor testing circuit you posted in my other thread.  All the transistors had a zero reading before pressing the switch. The readings after pressing the switch are as follows:  2.54; 3.19; 2.86; 3.01; 3.13; and 3.06.   Does this mean that they all had zero leakage and gains of 254; 319; 286; 301; 313; and 306?  I used the 286 and the 254 and put the 254 in Q1. It seems the fuzz was a little better using these transistors. When using Si transistors should Q1 have a substantially lower hfe than Q2 as I have read is done when using Ge transistors? 

Will better quality transistors produce a better fuzz sound?  Will 2n3906s with lower gains produce better fuzz sounds?   Do you know who makes 2n3906 transistors that are better than the Radio Shack ones?

What is the reason for going from 470R to 330R when using Si instead of Ge transistors?

I studied your pic and the schematic you posted. I see how it makes it easier to follow the schematic. Great idea. I am going to do the next one that way. The only thing about the schematic I don't understand is the 2k with the arrow over the 1k. It looks like a 2k pot with #2 to the 82R and the .22uf; #1 to the 1k resistor  and #3 not connected; but I don't see it on the pic. All I see on the pic is the resistor between the big gray resistor and the 2SA221 on the right. Where did you get those 2SA221 Ge transistors?  I have never seen a reference to those on any FF schematic. I'd like to see a pic of that board with the wires attached if you have one.

Thanks for all your help with this. Thank you Solidhex for your help as well.

[smallbearelec]

All the transistors had a zero reading before pressing the switch.

That's expected. The initial reading is leakage current, and silicon does not leak.

The readings after pressing the switch are as follows:  2.54; 3.19; 2.86; 3.01; 3.13; and 3.06.   Does this mean that they all had zero leakage and gains of 254; 319; 286; 301; 313; and 306?

Yes.

When using Si transistors should Q1 have a substantially lower hfe than Q2 as I have read is done when using Ge transistors?

Yes, and for best results both Q1 and Q2 should be way lower in gain. Now that you know you can build something that works, you can build something that sounds less harsh and is more friendly to a pedal board...see following comments.

Will better quality transistors produce a better fuzz sound?

Generic silicon devices are fine if the gains are right. IMO the 2N390x that a lot of people start with are way too "hot." But they are cheap and readily available.

Will 2n3906s with lower gains produce better fuzz sounds? Do you know who makes 2n3906 transistors that are better than the Radio Shack ones?

Check the gain of the 2N2222 or pn2222 that RS sells. I think those are lower gain and probably a better choice for Q2. Or shop mail-order for better choices.

What is the reason for going from 470R to 330R when using Si instead of Ge transistors?

Those resistor values stem from old schems that everyone has used for decades now. In fact, you are better off reworking the values of that voltage divider to give both proper bias and more output volume.

I held off offering this reference:
http://www.smallbearelec.com/HowTos/BreadboardSiFF/BreadboardSiFF.htm
until you had done your first build. Please check it out. The original FF was done with PNP germanium/positive ground, because that was what was cheap and readily available at the time.  Knowing what we do now, IMO, the Fuzz Face is best done with NPN devices, because then it does not need a separate power supply to play gracefully with other pedals on a board. Suitable, good NPN silicon is dirt cheap, and even NPN germanium can be found today at not-impossible prices. My tutorial covers a lot of basics, tweaks and component selection.

Happy construction, and enjoy your new pedal(s).

[LucifersTrip]

I got the voltage reading s for the 2n3906 transistors. I used the Germanium transistor testing circuit you posted in my other thread.  All the transistors had a zero reading before pressing the switch. The readings after pressing the switch are as follows:  2.54; 3.19; 2.86; 3.01; 3.13; and 3.06.   Does this mean that they all had zero leakage and gains of 254; 319; 286; 301; 313; and 306? 

not exactly...but close enough. silicon have such small leakage, it is insignificant for our purposes. further, I believe that test is set up for germanium so the readings for silicon will be slightly off....again, not enough to worry about

I used the 286 and the 254 and put the 254 in Q1. It seems the fuzz was a little better using these transistors. When using Si transistors should Q1 have a substantially lower hfe than Q2 as I have read is done when using Ge transistors? 

again, there's no definite yes. but as SB pointed out, most people have found that to be the best, though others like them to be equal or even have a higher gain Q1. In the end, it's only your ear that counts.

Will better quality transistors produce a better fuzz sound? 

what is a "better quality" transistor?  "better quality" for what circuit?  ironically, I think I remember reading that the original FF transistors were not high quality at all  and that low quality is actually a little bit of why is sounds good.

Will 2n3906s with lower gains produce better fuzz sounds?   Do you know who makes 2n3906 transistors that are better than the Radio Shack ones?

again, it's up to your ears only. I prefer a less harsh, more vintage sound so I go with lower gain ones. also note that transistors of the same number today many times have higher gains that the vintage ones. the new 3906's I have have gains usually 150-250 and the vintage ones I have can be 100 and lower.

What is the reason for going from 470R to 330R when using Si instead of Ge transistors?

the silicon needs less gain. it would be loud (and more boomy, less soft fuzz if you left it at 470)

I studied your pic and the schematic you posted. I see how it makes it easier to follow the schematic. Great idea. I am going to do the next one that way. The only thing about the schematic I don't understand is the 2k with the arrow over the 1k.

It is nothing more than 2 resistors in series giving me a range of 1-3K (the full range where I thought it sounded good). the 1K is called a cutoff so the resistance cannot drop to 0K. In a FF it is common to have a 3-5K in series with a 10K pot instead of the 8.2K

It looks like a 2k pot with #2 to the 82R and the .22uf; #1 to the 1k resistor  and #3 not connected; but I don't see it on the pic. All I see on the pic is the resistor between the big gray resistor and the 2SA221 on the right.

it is right below the cap in the upper right

Where did you get those 2SA221 Ge transistors?  I have never seen a reference to those on any FF schematic.

firstly, that's not a FF (it is a very similar variant of course). I don't remember where I got those but of the many Japanese transistors I have, those have been among the best for a FF. The fact that they have never been referenced in a FF schematic doesn't mean anything. Don't be a sucker and believe that the ones called for in a given schematic are the best. There are literally thousands of transistors from numerous countries, and the chance you would pick the one in the schematic to be the best is highly unlikely. The one in the schematic is only the most authentic.

I'd like to see a pic of that board with the wires attached if you have one.

ironically, i breadboarded that to help someone else before I switched to pnp germanium and added a cuttoff.
http://www.luciferstrip.com/fuzz/color-one-bread.jpg

I can send you what my final wiring looks like, but I'm not sure if it'd help because some of my wires go over the components and block stuff

[Rick899]

Luciferstrip, Smallbearelec, solidhex:

Thanks for all the help to get this Fuzz face right.  I have learned a lot from just this project.  I  am pretty sure now the problem with the fuzz is the high gain 3906 transistors I am using. I have a couple of 2n4403 which have 201 and 206 gain and they sounded a little better.  So now I believe it is a matter of getting the right transistors. That may end up being the most difficult part of this.  That tutorial is excellent!  Now let me go see what Smallbear has in the way of low gain PNP Si transistors ...    This should be about it for this thread.

[Solidhex]

 If the problem you were having was a lack of fuzz and the fuzz control only affecting the volume switching to lower gain transistors won't help. I don't think you ever posted the "in circuit" transistor voltages. That's a different measurement than the gain test you did. Its the only way to see if the transistors are biased properly.

[Rick899]

Solidhex:

Okay. Well I am learning a little more. I didn't understand about the in circuit transistor voltages.  I am not sure if I took the voltages correctly but what I did was put on a fresh battery, 9.23V. I biased Q2C to -4.50V and then measured each leg of both Q1 and Q2 with the black  lead of the DMM on ground and touching each leg with the red lead with the circuit switched on. The readings are as follows:

Q1E:  0
Q1B: -0.57
Q1C:  -1.30

Q2E: -0.69
Q2B:  -1.30
Q2C:  -4.50

Are these the correct measurements ?   Thanks.

[petemoore]

  Correct 'enough'..or within' working bias conditions.
  Base should be 1 diode drop or so above the emitter [see data sheet or measure foreward threshold voltage of the transistors diode with the DMM], there's really no 'correct', basically they fall within operational standards when the bias resistances are set up and DC is where it should be, thereabouts, it always varies somewhat.
  Q1E being connected to 0.0V makes it easy to measure 0.0V from ground, the emitters are above the bases by a diode drop or so, the 2 transistor feedback loop amplifier stage [FF] needs the output collector [Q2Collector] to have room to swing _/+ from it's bias point in order to...swing voltage and follow the guitar signal swings [allowing the amplified voltage swings to have room between the output bias and both 'sides' of the power supply. Since the PS is 9v, 4.5v bias allows....4.5v swing in both the negative and positive swing directions..but the amp needs 'elbow room' to operate so some of the + and - swing potential is 'used' to allow the transistors to operate [something like a volt or a little more is wasted...nonetheless and all the more reason to have the output bias 'centered'.
  That said the FF should work with the voltage shown [barring some other mixwire], and for assymetric clipping, moving the output bias voltage up or down from center [say put the idle bias at say 5 or 5.5v..'offsets' the bias and makes the + swings clip sooner and harder than the - swings...experiment for 'best' sound according to your ears.
  Starting off with Si NPN is fine, then with sockets or even a second FF board...mess around with the fuzzfaces until one is 'better' then better-ize the FF#2....

[Rick899]

A Final Note:

Posted below is a pic of the finished pedal. I think it sounds really good. Not a lot of fuzz but a lot of fuzz is why I have a Big Muff. I was looking for a sound like the tones in the Fillmore East version Of Elizabeth Reed. There are three or four particularly beautiful sustained notes, in that song, I don't know who is doing it, I think Duane Allman but maybe Dickey Betts, whose tones I was aiming for. I read in another forum that those tones come from a Fuzz Face and then I researched it and found out that Duane Allman preferred the sound of a Fuzz Face with partially worn batteries. The tones  in that song do not have a lot of fuzz.

This pedal comes close to those tones. And much better than I thought I would get during that long struggle with the transistors.  I lthink I learned enough to have a working knowledge of them.  I ended up using  a BC528/ECG129   TO39 type transistor with hfe 194  in Q2  and  a 2n4402  with hfe 131 in Q1. Those were the lowest gain PNP Si transistors I could find with a substantial difference between Q1 and Q2. . I'd like to find some PNPs with lower hfe numbers to try in it.  everything else is typical, I used a several old carbon resistors I took out of an old Harmon and Kardon stereo.

But it sounded really good alone; and then  with a JH-1 Crybaby modified with a yellow fasel and the "vocal" mod; it sounded really good with a Hall of Fame Reverb; and then adding in an MXR Carbon Copy it gives it yet another dimension. Another good thing is that it wasn't noisy, maybe a little bit of static, when not playing but not noticeable when playing. The noise may be at least partially the amp. I was using a Peavey Valve King 212 with well used tubes .   I used a Gibson Les Paul Traditional PLus with '57 humbuckers. It gave fat, rich not too trebly tones with good sustain alone and with all the combinations mentioned. . Next, I am going to try it  through a Vox Night Train with a JBL D-130  15" speaker cab.

   

[Rick899]

Oh, yeah ... thanks again to everybody that offered feedback and info and Happy Holidays.