Big Muff Rams Head build report and some questions on DC offset and biasing

[iamok]

Hi everyone! My first post here and I love how valuable this forum has been in helping me build again after a long time. All your questions and answers are really helping me understand concepts in a new light (sound? ).

I have built a Big Muff - Rams Head, based on these schematics from Effects Layouts https://www.dropbox.com/s/gg62jvk86i5a5y2/BMP%20Project%20File.pdf?dl=0(David Gehring - is he here?) that works very well. I have used BC107B because those are available.

At first the output was quite limited and felt like the output waveform was being clipped excessively so I was losing lows and highs when playing chords and also on some notes. At time notes would start out low volume and gradually rise - as if there was an attack. I built an audio probe (so happy to discover this technique to debug!) and checked different stages. I figured that the signal sounded great until Q4 (gain recovery stage) where when I probed the Q4 collector I could hear the harsh clipping. For a day or two I played through the pedal by completely bypassing Q4 (as it was socketed) - I removed Q4 and figured I could just connect a jumper between pins of the socket where CB of Q4 would go. I actually quite liked the sound of the pedal without the recovery stage but I also felt I might be losing some aspects of the "big muff tone and gain" (not sure about this because it still sounded great - in fact it got quite subtle and i could control fuzz from my guitar volume alone - i might just make a mod for this!). Regardless, I set out to figure what the problem with Q4 really was and so I put it back and started reading up to understand the clipping that was occurring on Q4.
Thanks to all the previous and this particular Q&As (http://www.diystompboxes.com/smfforum/index.php?topic=62869.msg494850#msg494850) on this forum I could debug all the common issues listed by R.G on that thread. I found out that the voltage at Q4 collector was only 0.87V. I followed his listed recommendations to rule out various reasons for that low voltage but I found none. Other than the low voltage on Q4 I found that some caps were showing the same DC voltage on both ends (for example ~1.2v just before and after C13)  - as I understand Caps should be removing all DC and one side of the cap should be 0 - is that correct? But the audio sounded fine on both ends of the caps when checked with audio probe. So I arrived at the conclusion that Q4 biasing is where the problem was. Referring to the linked schematic I changed R22 to 36k and found that Q4 collector was now at around 4.2V (yippy!) and the audio signal was clear with all the headroom.
I am happy with how this has turned out but I want to ask if I have a correct understanding of the cap issue I mentioned above. Does seeing the same voltage (more than 0) on either ends of the cap mean that there is a DC offset? If so could this actually be the reason for low voltage on the Collector of Q4 and the resulting clipping? Have I really solved the problem? Can I make the sound better by removing this offset - if it really is one?

One other thing I have noticed after increasing R22 to 36k is that I am getting feedback (a loud whistle) on certain combination of settings on the guitar, pedal and amp. Changing some of these removes feedback - for example lowering sustain on the pedal does, and at times changing (both, increasing or decreasing!) my guitar volume helps. I can't understand what is going on. Sometimes there is feedback even on a low sustain setting but then it goes away if I twiddle knobs on the guitar and vice versa. Any thoughts on this could be useful to further understand the concepts too.

Many thanks~!

[thermionix]

Caps are frequently used to block DC, and should not have DC on both sides.  But in some instances they could, it depends where they are in what circuit.  You may very well have a leaky or shorted cap causing your problems.  Can you provide a schematic and point out which cap you're referring to?

[thermionix]

Oh, and welcome! 

[antonis]

Referring to the linked schematic I changed R22 to 36k and found that Q4 collector was now at around 4.2V (yippy!) and the audio signal was clear with all the headroom.

If I trace correctly, R22 is Q4 Emitter resistor which should have a value of 2k7..
(with R23 Collector resistor value of 12k, result in 13db gain/recovery..)

Changing its value to 36k and having good result means that your actual R23 value is about 13 times higher..
(probably 120k instead of 12k - very common building fault..)

The above assumption is underpinned by the heavy saturated Collector voltage (870mV)..
With 390k/100k voltage divider resistor values you should  have an Emitter voltage of about 1.1V, hence a Collector current about 400μA resulting in a Collector quiescent voltage of about 4.2V (yippy!)

IMHO, you better re-check Q4 bias circuit (R20-21-22-23) and set original values 'cause if the above case stands you have a very "fragile" Q4 amp..

P.S.
I second thermionix second post..

[iamok]

Thank you thermionix and antonis for the explanations.

According to the schematic ( link -> https://www.dropbox.com/s/gg62jvk86i5a5y2/BMP%20Project%20File.pdf?dl=0 ) the original values of voltage divider for Rams Head are (you can find them on second page BOM table)
R20 390k
R21 100k
R22 10k
R23 100k

Now these R22 and R23 values don't match most other schematics I have seen. Is it possible that they are wrong? I only changed R22 to 36k and that made Q4 collector voltage to be ~4.2v. I have limited understanding of a voltage divider so I need to look into that.

Changing its value to 36k and having good result means that your actual R23 value is about 13 times higher..
(probably 120k instead of 12k - very common building fault..)

Note that the R23 value as suggested in the schematic is indeed 100k close enough to what you calculated but R22 in schematic is 10k, which I changed to 36k.

Caps are frequently used to block DC, and should not have DC on both sides.  But in some instances they could, it depends where they are in what circuit.  You may very well have a leaky or shorted cap causing your problems.  Can you provide a schematic and point out which cap you're referring to?

So on the output end they should mostly read 0? C13 reads 2.76v on Tone pot terminal 2 side, 1.26v Q4 base side

Battery 9.02v ( when not in circuit)

Q4: C 4.53, B 1.30, E .77v

Note: All these readings are with the R22 @ 36k still in there

What should I do next? Thank you very much.

[antonis]

According to the schematic ( link -> https://www.dropbox.com/s/gg62jvk86i5a5y2/BMP%20Project%20File.pdf?dl=0 ) the original values of voltage divider for Rams Head are (you can find them on second page BOM table)
R20 390k
R21 100k
R22 10k
R23 100k

It's probably a typo..

Anyway, it's not wise to bias Q4 with such big value resistors 'cause you may obtain "correct" bias voltages but only for unloaded output..
(high resistor values result in low working current - severe loaded even by light loads..)

[PRR]

> original values of voltage divider for Rams Head are ...
> R20 390k
> R21 100k
> R22 10k
> R23 100k

With a hi-gain transistor, these values will NOT work. Simple math shows "12V drop in R23", which can't happen in a 9V circuit.


The various values may have been tuned to the crappiest (lowest-gain) parts they had at that time.

There may be a more universal set of values but it is time for my nap.

[Slowpoke101]

Had a quick look through my notes of the BMP Rams Head resistor values and found the following;

R22 = 2K7
R23 = 10K

Sometimes R23 may be a 12K resistor. These values are just the recommended values. You may wish to try different values in order to chase "that certain tone".

As antonis said:

If I trace correctly, R22 is Q4 Emitter resistor which should have a value of 2k7..
(with R23 Collector resistor value of 12k, result in 13db gain/recovery..)

This is what the figures for this stage should be. Change these resistors and your BMP should behave itself properly.

[antonis]

@iamok: Just a rough guide for Voltage/Current calculations in case of voltage divider bias (4 resistor) BJT amp..
(if you imply those to your circuit you'll verify P.R.R.'s conclusion..)



(not verified - hopping for no faults..)

P.S.
Paul, I like your picture title as BadBMP..

[iamok]

I changed the following to
R22 2K2 (don't have 2k7)
R23 10k

My Q4 volts are
C 3.88
B 1.59
E .99

My battery is running a little low at 8.7v

@iamok: Just a rough guide for Voltage/Current calculations in case of voltage divider bias (4 resistor) BJT amp..
(if you imply those to your circuit you'll verify P.R.R.'s conclusion..)

> original values of voltage divider for Rams Head are ...
> R20 390k
> R21 100k
> R22 10k
> R23 100k

With a hi-gain transistor, these values will NOT work. Simple math shows "12V drop in R23", which can't happen in a 9V circuit.


The various values may have been tuned to the crappiest (lowest-gain) parts they had at that time.

There may be a more universal set of values but it is time for my nap.

Thanks for this. It makes sense to me now.

I do have a lot of squealing feedback on high pedal settings which I have sorted for now by placing the pedal after a TS9 (which I found is buffered). I don't understand how the squealing is gone but its working.

[iamok]

@iamok: Just a rough guide for Voltage/Current calculations in case of voltage divider bias (4 resistor) BJT amp..
(if you imply those to your circuit you'll verify P.R.R.'s conclusion..)



(not verified - hopping for no faults..)

P.S.
Paul, I like your picture title as BadBMP..

Why is it R1//R2 when these aren't actually in parallel in the circuit? I just looked up what // means and it says

a//b  probably means the equivalent single resistance for 2 resistors in parallel of size a and b, which is ab/a+b

[thermionix]

Why is it R1//R2 when these aren't actually in parallel in the circuit? I just looked up what // means

|| is more proper than //, but Shirley some keyboards don't have the | character.

R1 and R2 are in parallel, from the perspective of the (AC) signal, as Vcc is seen as an audio ground along with regular ground.  Think about where a filter cap might be, if that helps.

[PRR]

R1||R2 is the equivalent resistance of that Voltage Divider.

RE * hFE is the approximate impedance looking into the Base.

The "10 X" part means we want the R1||R2 impedance much-lower than the Base impedance, perhaps 10X lower. (This is often A Problem, since it leads to small audio input impedance.)

[Rob Strand]

Why is it R1//R2 when these aren't actually in parallel in the circuit? I just looked up what // means and it says

You find the impedance looking into where R1 & R2 join by treating the power supply as a short circuit.  When you do that R1 and R2 look like they are in parallel.

https://en.wikipedia.org/wiki/Th%C3%A9venin's_theorem
http://www2.ensc.sfu.ca/~glennc/e220/e220l6b.pdf

\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
  \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
  \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
---------- Wall  (TBD) ----------
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
  \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
  \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/

[iamok]

This is been so helpful. Thank you everyone!

[antonis]

Think about where a filter cap might be, if that helps.

Or a negligible internal resistance in case of battery powered..

I do have a lot of squealing feedback on high pedal settings

Try lower gain transistors (maybe Q2 & Q3 only) and/or raise C9 (Q3 feedback cap) value to 100nF..
You may alternatively tweak diode pairs series caps..

BMP is THE tunable circuit..


@Rob: Can you plz elaborate..?? 

\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
  \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
  \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
---------- Wall  (TBD) ----------
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
  \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
  \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/
\|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|/ \|

[Rob Strand]

@Rob: Can you plz elaborate..??

I was reading about the different symbols for parallel resistors and had an artistic outburst with an abstract representation of Donald Trump's wall hence the TBD.  The "\|/" are little cactuses.

[thermionix]

A wall of Douglas Devastators?