Fuzz Bender DIY does no fuzz, or sound

[matopotato]

Hi,

EDIT: Updated IC readings to fit standard listing and added Q1-3 readings May 25 15:40 CET

I have built another stripboard clone found here https://dirtboxlayouts.blogspot.com/2020/03/keeley-fuzzbender.html.
And it wont play.
It has 5 pots: Level/Vol, Fuzz, Treble, Bass and a dualgang Bias/Gate.
The Volume, Treble and Bass clearly have effect on the sort of white noise I get when engaged.
The Bias/Gate is trickier to tell if it does anything, mostly since the Fuzz is quiet and does not react. I have managed to hear something that could be described as extreme fuzz like dying battery plus all the fuzz you could get. But it is not useful and just appears random and seldom.
Bypass is ok.
Some pics. You can see below how I have "beep-traced" each leg, and all checks out between the solders and schema
I also like to add that for each leg or wire in the perf board I have beep-tested the leg before cutting it against its own strip (beep) and its neighbors (no beep), to at least minimize any bridges. Same with the cuts, verified each with beeper to ensure it was cut enough.
(Painstaking perhaps but felt like a good method. And worth it since I found a few possible mistakes. Need to be aware of any short jumper links made on the other side, or could be confusing)






This is the Layout I built after


This is the schematic that comes close. A few discrepancies I feel are ok: The D2 1N5817 was moved from +9V to ground, to instead sit directly after 9V. The designer confirmed this is intentional, and the layout obviously works from others who built it.
C14 and R17 are swapped, and I guess it does have no impact.
The Gate A in the schematics, leg 3 go to "X" which is effectively pin 8 on the three ICs. (In the alternative schem below it goes to VB indicating lug 2 and 3 could be joined on the Gate A, although they are not in the layout.
For C11 I used 470pF instead of the 500pF (I had 470+30 but cut the 30 off during troubleshooting)


Here is another schematic that might also be close, but haven't followed that one. Yet.


2 errors on my part I found and corrected:

• C12, the 100nF at S2-T2 I had as 100pF first. It takes 9V to ground, but there is also a 100u electrolyte doing that, so I hope this mistake was not frying anything
• The cut at L5, I had put at M5 instead. So C10 cap never reached R5 or Q2 emitter. and Q2 base went to R13 as well as where is should. And  Fuzz lug 1. So that was pretty wrong.

I wonder if either or both mistakes could have "fried" anything?
I think the transistors check out, at least in my mini-meter they come out with pinout, hfe and some other sensible values.

I am using 2N2222A for Q1 and Q2 (should be OK Si NPN alternatives from what I have read)
Q3 I have 4 different Ge NPN I have rotated, 3 from a 70s TV set and one AC176 from an old radio way back.
IC 1 and 2 are RC4558s and IC3 is TC1044s
And they measure when power is on:
(Ordered by:
pin1 pin8
pin2 pin7
pin3 pin6
pin4 pin5)


IC1, RC 4558
This warms up a bit, but not so it gets hot...
0.06    8.86
0.00   -1.13
0.00   -1.13
-2.71   0.00

IC2, RC 4558
Stays cold. Makes sense though given the readings.
0.00   8.86
0.01   0.06
0.01   0.05
-2.70  0.05

IC3, TL 1044s
This one quickly gets quite hot
8.86    8.87
4.66    5.28
0.00    3.91
-1.20  -2.71

Q1:

C: 6.65
B: 1.54
E: 0.97

Q2
C: 6.67
B: 0.97
E: 0.40

Q3
C: 0.00
B: 0.01
E: 0.02

I did try to probe with looper playing in, but it quickly got very random and confusing so I did all the other stuff above instead. Will probably go back to probing later on.

Any help is much appreciated.
Let me know if any more pictures or measurements should be made...

Edit: Got some better resolution on the schems I hope

[antonis]

The D2 1N5817 was moved from +9V to ground, to instead sit directly after 9V. The designer confirmed this is intentional, and the layout obviously works from others who built it.

Although D2 can work either in series with power supply (+) (Truth Bender) or in parallel (shunt, Keeley Fuzz Bender) its low forward votage drop (Schottky diode) makes it more indicated for the former usage..

[matopotato]

Thanks for clarifying.
Any ideas of where I made some mistake based on the original post? Or is some info missing?

[GibsonGM]

My 2 cents - go back over EVERY single connection again. 99.9% of the time, it's something we did when we built the circuit. It's frustrating to look and not find anything, I know!   But there is typically a pretty simple explanation for why it won't work.  Shorts between parts are VERY common, so look for 'bridging' too!

Can you post IC and transistor voltages like the following?

pin 1   pin 8
pin 2   pin 7
pin 3   pin 6
pin 4   pin 5

and transistor collector, base and emitter like:

C:
B:
E:

Makes it easier to interpret and to be sure where you took the measurements on the IC as viewed from above!


Do you have an audio probe yet?  There are many variations on the same theme, easy to learn about with a net search.  http://www.doitfuzz.com/forum/diy_tutorials.php?name=Homemade_Audio_Probe_for_debugging_DIY_pedals

This may help you locate where the signal 'in' stops, which can point to an area to look closer!

[antonis]

Q3 bias is related upon transistor leakage current to bring Q3 Base 200mV - 300mV higher than GND for Q3 to be active..
I'd put a low voltage supply(*) on Fuzz pot lug 3 (C5/C11 joint) to see what's happeninbg with the particular transistor..

(*) 9V supply is handy but you have to be sure that Fuzz pot can travell all the way CCW or you'll have a permanently bottomed transistor..
Better make a 10k/1k voltage divider, fed from +9V, connected to Fuzz pot lug 3 for wider trimming margins..



P.S.
ICs measurements make no sense without respective pin numbering..
edit: Sir Mιke dealt with it..

[matopotato]

My 2 cents - go back over EVERY single connection again. 99.9% of the time, it's something we did when we built the circuit. It's frustrating to look and not find anything, I know!   But there is typically a pretty simple explanation for why it won't work.  Shorts between parts are VERY common, so look for 'bridging' too!

Can you post IC and transistor voltages like the following?

pin 1   pin 8
pin 2   pin 7
pin 3   pin 6
pin 4   pin 5

and transistor collector, base and emitter like:

C:
B:
E:

Makes it easier to interpret and to be sure where you took the measurements on the IC as viewed from above!


Do you have an audio probe yet?  There are many variations on the same theme, easy to learn about with a net search.  http://www.doitfuzz.com/forum/diy_tutorials.php?name=Homemade_Audio_Probe_for_debugging_DIY_pedals

This may help you locate where the signal 'in' stops, which can point to an area to look closer!

Thanks
I updated the original post with proper listing format and added transistor readings. Clearly Q3 is not "getting any"...
I will at least go over the strips to verify there are no bridges. I did during soldering, but will try again.
I also went over yesterday each component leg (where possible) with its connected component leg with the DMM-beeper. Some cases I had to take the solder or strip since the leg was not reachable. (Or a pin for a ceramic cap with isolated/painted legs.)
And to my surprise they check out. I was expecting errors.
And I did find once cut I had made wrong, but I corrected it and thought it would improve things, which it didn't.
So I was wondering if that error might have fried anything?
IC2 does not seem to participate much, and IC3 getting hot. Then Q3 getting zero. And the Fuzz pot not doing anything.
I think it was doing something subtle at first, then I re-soldered all the wiring because it sounded like a wire glitch (touching wires made it glitch, but not clear or consistent enough to pinpoint any one specific. Strangely). But after that the Fuzz went even more silent.

I appreciate your experience and tip to go through again and again. It feels tedious but I understand that is what I must do.
So the strips, wires and probe. Then connectivity a second or third? loop...
I have a probe that I use with a looper. I tried a bit half-heartedly at first, but it was sort of random and inconsistent in my limited understanding, but I will try again. I had good success troubleshooting a Spectron Env Filter with it, just need to be more structured.

[matopotato]

Q3 bias is related upon transistor leakage current to bring Q3 Base 200mV - 300mV higher than GND for Q3 to be active..
I'd put a low voltage supply(*) on Fuzz pot lug 3 (C5/C11 joint) to see what's happeninbg with the particular transistor..

(*) 9V supply is handy but you have to be sure that Fuzz pot can travell all the way CCW or you'll have a permanently bottomed transistor..
Better make a 10k/1k voltage divider, fed from +9V, connected to Fuzz pot lug 3 for wider trimming margins..



P.S.
ICs measurements make no sense without respective pin numbering..
edit: Sir Mιke dealt with it..

Thanks!
So the 1k/10k divider is enough to protect the transistor? Or you think that is what happened, I already bottomed it and it is now broken?
Or is the 1/10 divider to get more out of the Fuzz pot?

[matopotato]

On the other hand, if Q3 was broken I wouldn't get my analyzer to accept it... I guess

[duck_arse]

IC2 does not seem to participate much, and IC3 getting hot. Then Q3 getting zero. And the Fuzz pot not doing anything.

there it is, you have found it. that IC is supposed to be providing [I think, I've never used one and it's not marked on the diagram] -9V. you have bugger all. also, Q3 is supplied via the 2 10k resistors dividing +9V down to 4V5, so if one of those resistors is way off value, or something is shorting across one, say to ground, all yer volts will be rong.

so, power off. switch your meter to the ohms range - NOT continuity in this instance. measure the two 10k resistors providing Vb. measure the resistance from the Vb point to ground as you wind the bias pot on Q3, see what readings you get.

[edit :] also, both your schems show IC3 pins 1 and 8 not joined, yet your layout has shorted them together. is this allowed with your particular part fitted? specs sheet allows this?

[antonis]

So the 1k/10k divider is enough to protect the transistor? Or you think that is what happened, I already bottomed it and it is now broken? Or is the 1/10 divider to get more out of the Fuzz pot?

The last one..
It will provide "enough" voltage to Fuzz pot, hence a healthy Q3 will reveal its elegance, being aloof to leakage current..
Of course, as Stephen already pointed out, Q3 Collector MUST be fed with a proper supply so check for +4.5V (give or take - the last dependent on your DMM impedance) on R22/GATE_A wiper joint..

[matopotato]

So the 1k/10k divider is enough to protect the transistor? Or you think that is what happened, I already bottomed it and it is now broken? Or is the 1/10 divider to get more out of the Fuzz pot?

The last one..
It will provide "enough" voltage to Fuzz pot, hence a healthy Q3 will reveal its elegance, being aloof to leakage current..
Of course, as Stephen already pointed out, Q3 Collector MUST be fed with a proper supply so check for +4.5V (give or take - the last dependent on your DMM impedance) on R22/GATE_A wiper joint..

Thanks, will try this, but not until next week due to travelling.
I am a bit puzzled though since others have built this and it worked for them. Also the schematics I have come across do not use voltage divider to get more out of Q3.
I would expect your suggestions will make it better as you describe, but I would have thought it would work as intended before improving.
I will also go through each component and solder etc to verify it is done as designed.

[matopotato]

IC2 does not seem to participate much, and IC3 getting hot. Then Q3 getting zero. And the Fuzz pot not doing anything.

there it is, you have found it. that IC is supposed to be providing [I think, I've never used one and it's not marked on the diagram] -9V. you have bugger all.

Just to be clear, you mean the IC2 to provide the -9V or the IC3?

also, Q3 is supplied via the 2 10k resistors dividing +9V down to 4V5, so if one of those resistors is way off value, or something is shorting across one, say to ground, all yer volts will be rong.

so, power off. switch your meter to the ohms range - NOT continuity in this instance. measure the two 10k resistors providing Vb. measure the resistance from the Vb point to ground as you wind the bias pot on Q3, see what readings you get.

Will do. It seems that measuring resistors after they are in a circuit is usually ok, but I find that caps can be more difficult when in the context of the circuit. I mean all components are connected to somewhere and often that context has R and C of its own so I have run in to some "parallelisms" that were tricky to distinguish from the proper metered value. Found out the hard way in another troubleshooting where a cap just didn't make sense, so I took it out to change for another and then it turned out to be just fine and measured up correctly.
For resistors I think this has been less of a problem for some reason, maybe the DMM methods are differently sensitive to circuit context.
But I will check the relevant R's for sure. Thanks

[edit :] also, both your schems show IC3 pins 1 and 8 not joined, yet your layout has shorted them together. is this allowed with your particular part fitted? specs sheet allows this?

In the link to dirtboxlayouts in my first post there is a discussion, and the schematic and layout are not 1-1 on all parts. About the IC3p1 and IC3p8 I got the reply:

The schematic at FSB is a bit confusing at a first glance with all those unessecary "x's"...
Pin 6 and 7 should not connect to anything. Pin 1 and 8 connects to 9v through a polarity protection diode(D2).
<snip>
For pin 1 & 8. I almost always use the more common (and cheapest from reliable sellers) charge pumps avaiable like the one written in the notes under the layout. For these, both pin 1 and 8 must be joined together, otherwise you will dissconnect the Boost pin (pin 1) and will end up with squealing and noise. Slightly less common charge pumps like LT1054 and TC1044 almost always has pin 1 hanging free.

To which I replied

Many thanks! I thought the schema and layout were from the same, but now I understand. And the ref to Xes led me to that as well.
I plan on getting TC1044S, and might as well keep it connected anyway as I interpret you as it is ok.
Again thanks!

Anyway, I was not expecting pin1 being or not being joined would have created the problems I experience.
I see now that I could have been wrong and pin1 should perhaps be left alone?
It is fairly easy to cut and if needed reconnect so I can try next week.

Any thoughts on my faulty cut at L5? I had first cut at M5 instead. So C10 cap never reached R5 or Q2 emitter. and Q2 base went to R13 as well as where is should. And  Fuzz lug 1. So that was pretty wrong. (Adjusted now) Could that have fried anything?
As well as misreading C12 as 100pF instead of 100nF. (also fixed now)

[antonis]

I am a bit puzzled though since others have built this and it worked for them. Also the schematics I have come across do not use voltage divider to get more out of Q3.
I would expect your suggestions will make it better as you describe, but I would have thought it would work as intended before improving.

You didn't get me..

Q3 B-C reverse biased junction needs to exhibit an about 2.5μA leakage current for proper Base bias (at 200 - 300mV)
In case of that current is smaller, Q3 can't work in its active region..
So, by injecting a proper voltage on Fuzz pot upper lug you could tell about Q3 "self-bias" non plenitude..
If it works, you have two options: Heat it up substuntially or replace it..

P.S.
The above test should be implemented AFTER +4.5V on R22/GATE_A wiper joint verification..

[matopotato]

I am a bit puzzled though since others have built this and it worked for them. Also the schematics I have come across do not use voltage divider to get more out of Q3.
I would expect your suggestions will make it better as you describe, but I would have thought it would work as intended before improving.

You didn't get me..

Q3 B-C reverse biased junction needs to exhibit an about 2.5μA leakage current for proper Base bias (at 200 - 300mV)
In case of that current is smaller, Q3 can't work in its active region..
So, by injecting a proper voltage on Fuzz pot upper lug you could tell about Q3 "self-bias" non plenitude..
If it works, you have two options: Heat it up substuntially or replace it..

P.S.
The above test should be implemented AFTER +4.5V on R22/GATE_A wiper joint verification..

No big surprise, I am not seasoned in this diy pedal stuff.

Will try after weekend. Again, thanks!

[duck_arse]

Just to be clear, you mean the IC2 to provide the -9V or the IC3?

no sir. I'm saying IC2 is a dual opamp, supposed to be doing audio workings. I'm saying IC3 is the inverter IC, is supposed to be producing a negative rail for both audio IC's.

as for measuring the resistors in circuit, in this instance we don't care about the absolute value showing on the meter. we expect it to be lower than the colour bands, but I can't see well enough in your photo, so you are checking that the 10k marked resistors are not instead reading 100k or 1M or indeed 120R. you just need to confirm the 10-ish k-ishness of them so that we know they are not the cause of the bad low voltage.

as for what they say at other sites - we pay no mind.

as to your linking pins 1 and 8 - as I say, I have no idea about those charge pumping IC;s, just that there are very many variations and permutations. and you don't have a circuit diagram to exactly match the layout - so it's on you when you do mods and substitutions, as with transistors, you CHECK THE DATASHEET for the exact part in your hand, WITH all the prefix and suffices. only then decide what to do w/ pins 1 & 8.

it's far more likely a build like this will fail due to a charge pump IC misconnection providing a dud voltage than it is likely to fail from a cap being off value, even by 100 times. the caps will affect the audio, but only once you get the right static DC conditions around and about.

[matopotato]

Just to be clear, you mean the IC2 to provide the -9V or the IC3?

no sir. I'm saying IC2 is a dual opamp, supposed to be doing audio workings. I'm saying IC3 is the inverter IC, is supposed to be producing a negative rail for both audio IC's.

as for measuring the resistors in circuit, in this instance we don't care about the absolute value showing on the meter. we expect it to be lower than the colour bands, but I can't see well enough in your photo, so you are checking that the 10k marked resistors are not instead reading 100k or 1M or indeed 120R. you just need to confirm the 10-ish k-ishness of them so that we know they are not the cause of the bad low voltage.

as for what they say at other sites - we pay no mind.

as to your linking pins 1 and 8 - as I say, I have no idea about those charge pumping IC;s, just that there are very many variations and permutations. and you don't have a circuit diagram to exactly match the layout - so it's on you when you do mods and substitutions, as with transistors, you CHECK THE DATASHEET for the exact part in your hand, WITH all the prefix and suffices. only then decide what to do w/ pins 1 & 8.

it's far more likely a build like this will fail due to a charge pump IC misconnection providing a dud voltage than it is likely to fail from a cap being off value, even by 100 times. the caps will affect the audio, but only once you get the right static DC conditions around and about.

Thanks. Clearer now. I also realize I am a bit naïve and gullible about builds. I google for a component plus equivalent or replacement until I find something I already have stocked or can find relatively easy. Hoping it will be similar enough. I understand now that there are no shortcuts that can substitute doing the proper legwork and "knowing your numbers".

[antonis]

also, both your schems show IC3 pins 1 and 8 not joined, yet your layout has shorted them together. is this allowed with your particular part fitted? specs sheet allows this?

Pin 1 (Boost) is used for on-board oscillator operation at increased frequency (45kHz), instead of its nominal one (10kHz), for higher supply current applications (and possible noise issues..)

IMHO, a cut between Pins 1 & 8 should make things more clear about 1044 working state..

[matopotato]

also, both your schems show IC3 pins 1 and 8 not joined, yet your layout has shorted them together. is this allowed with your particular part fitted? specs sheet allows this?

Pin 1 (Boost) is used for on-board oscillator operation at increased frequency (45kHz), instead of its nominal one (10kHz), for higher supply current applications (and possible noise issues..)

IMHO, a cut between Pins 1 & 8 should make things more clear about 1044 working state..

Pin 1 also runs off to the left in the layout to link, 10k, link before stop. Maybe cut it both sides to isolation?

[antonis]

Pin 1 also runs off to the left in the layout to link, 10k, link before stop. Maybe cut it both sides to isolation?

On a second glance, it should be more complicated 'cause by cutting Q-15 track pad, you effectivelly cut the power for almost half of the circuit..
(you'll need pin 1 both sides cut and some kind of D2 cathode / Q-12 jumpering (under 100μF cap)

Before doing any layout mod, verify proper (-)9V output..
Then, check for +4.5V at E - 6/7/8/10 track pads..

[matopotato]

Pin 1 also runs off to the left in the layout to link, 10k, link before stop. Maybe cut it both sides to isolation?

On a second glance, it should be more complicated 'cause by cutting Q-15 track pad, you effectivelly cut the power for almost half of the circuit..
(you'll need pin 1 both sides cut and some kind of D2 cathode / Q-12 jumpering (under 100μF cap)

Before doing any layout mod, verify proper (-)9V output..
Then, check for +4.5V at E - 6/7/8/10 track pads..

Thanks, will do