Biasing a Silicon Tonebender

Started by Joe Hart, October 28, 2008, 08:19:46 PM

Previous topic - Next topic

Joe Hart

I built RDV's Tonebender MKII NPN Silicon version (can't find the schematic on line right now -- I will search some more), but I am having some issues. I have built a couple of Fuzz Faces and could never really get them to bias, but they sounded fine.

If you look at the schematic:
I changed the cap to ground at the input to a non-polarized film cap.
I approximated (pretty close) the 2.7K and 5.6K resistors with two resistors in series.
I also added Joe Gagan's "Easy Face input blend cap knob trick" thingy where the polarized 4.7uf cap is using a 1uf and a 10uf cap.

Are any of these changes fatal to the circuit?

Well, it sounds okay, but I cannot get it to bias correctly and it sounds a little gated and sputterey. Here is what I have for transistors:

Battery is at 9.66V

Q1 (hfe 69)
C 7.76
B 1.00
E 0.41

Q2 (hfe 110)
C 1.87
B 0.63
E 0.00

Q3 (hfe 217)
C 3.07
B 1.87
E 1.23

I know that the voltages for (at least) Q3 are way out of whack, but I cannot get it to bias even close to 4.5V. I am using the 5.6K resistor going to the collector. Is that correct? I even put a 50K pot there and still couldn't get it to bias. Could I need 100K or more?

Also, is the 100K resistor to ground at the base of Q1 that transistor's bias resistor? I did some research and I think it is, but it doesn't make much sense.

And is the 47K resistor going to the collector of Q2 the correct bias resistor? I think it is.

In addition to my biasing woes, the volume knob really just seems to slightly change the fuzz amount, the attack knob does nothing, and the input blend knob does practically nothing. Is this just a function of my unbiased Tonebender?

I know I am asking a lot, but I have been researching this and cannot seem to figure it all out. Thanks!!
-Joe Hart

Joe Hart


petemoore

I also added Joe Gagan's "Easy Face input blend cap knob trick" thingy where the polarized 4.7uf cap is using a 1uf and a 10uf cap.

Are any of these changes fatal to the circuit?

  That can be done right, gotta watch polarities.
  I guess you put them in ok, BTBack +<->+,  or 'parallel opposite' making a non-polarized capacitence out of the two capacitors.
  The voltages, especially Q3C seems off. I would go through the FF portion of the circuit, R by R, connect by connect to see if there isn't an errant or trace in the circuit, around below the Q3 is tricky with the gain pot etc.
  The circuit works as shown, the inability to bias and gating point toward possible wiring problem.
Convention creates following, following creates convention.

Gus

#3
Q1s collector voltage is a problem

The bias is set by all 4 resistors.  The base to ground voltage across the 100K is the the Vbe drop of Q1 and the voltage drop across the 2.7K (emitter current x the emitter resistor)

Q1s collector should be about the base to ground voltage across the 100K  PLUS the voltage drop across the 470K  OR about 5.7 x the base to ground voltage (the 5.7 is from 470K plus 100K for 570K)  I would try a higher Hfe for Q1 at least 200.  This approximations only work with higher Hfe transistors maybe >200 other wise you need to find the other current paths things depend more with lower hfe.  If you start to think about this circuit fragment you will start to understand how the parts interact.  OR to get an idea of what I am posting, try a 220K then a 680K  1 meg etc. for the 470K feedback part of the bias resistor and see what happens with the collector voltage

OR look at the hot Si first stage that bias setup and values used "forces" the circuit to bias in a more controlled manner, things like available current in the bias string to bias the transistor vs the transistor hfe and emitter resistor.

The voltage at the TOP of the gain pot is part of setting Q2 and Q3 bias the voltage drop across the 100K feedback sets the current(voltage / 100K) being used for Q2 bias

For people that understand a bit more how much is the 100 ohm emitter resistor doing with a 47K collector resistor?  Note I use a 10K in my FF builds.

Some of us try to give repeatable designs and people seem to keep missing it and building things that are not really paint by number builds or the circuit have build issues like needing to select parts.  To build a standard value FF and get it right you need to understand more that the simple looking circuit seems to be.

mac

This mods work for high gain transistors.

- Since you are using a low gain one at q1, which is similar to a Ge device, q1 bias should be closer to the original design. That is a 10k from C to Vcc, a 100k from B to gnd, emiter grounded, and about 1M - 1M2 from B to C to get near 8.0v - 8.5v at q1 C.
You could try bypassing the 2k7 resistor with a 1) big electro cap. because gain is very low 10k/2k7 < 4 compared to the original; or 2) with a wire, and replacing the 470k with a 1M - 1M2 or so. I'd start with 1)

- To increase Q3 C voltage to 4.5v reduce the 5k6. Q2 voltages are ok.

- Originally there is a 100k at q2 C instead of 47k. I prefer the 100k because it sounds more gated, but this is my personal taste.

- The atttack pot never gets clean, and the input pot in mine does a lot, especially when switching to HB mics.

I suggest you using low gain Si like 2N2369A, MPSA42, 2N3903, BD139-10, 2N3440, etc with gains near 60 - 100, and the original schem. All you need to do is to add the feedback resistor at q1 to make it bias.

mac
mac@mac-pc:~$ sudo apt-get install ECC83 EL84

Joe Hart

Thanks!! I went more with Mac's first suggestion because I wanted to try to stay with the transistors that I had (for mojo reasons -- they are reclaimed and they look cooler than the everyday plastic ones). So, I still haven't gotten it biased, but it sounds good (possibly great, I still have to give it a proper test run through my main rig), it cleans up very nicely with the guitar's volume control, and has tons of sustain. Thanks again.

Now the problem is the input cap blend knob. I've done it before and it's quite simple, so I know it's wired up correctly, but it makes no difference to the tone! I even tried just using an input cap like in the schematic without the blend pot, then swapping different cap values. The schematic calls for 4.7uf, and I tried values from 1uf to 220uf (!!) and there is no real perceptible change in tone. Ideas? In the past it has worked beautifully to go from cutting to thick tones.

Is there another place to put a tone control? I like the input cap blend because it changes the character of the fuzz rather than just shaping the tone after the fact. And it doesn't seem to steal any gain (probably because it's pre-gain and doesn't send anything to ground.

Thank you.
-Joe Hart

Gus

Like I posted look at the hot silicon.

Did you try adjusting the 470K? 

RDV

#7
Using the suggested transistors and values you can build them all day and they'll all sound very similar.

I would use the same value tranny for Q1 & Q2, otherwise just skip my little borrowed mod there and build the 1st stage just like the original TB Mk.II.

The "Hot Silicon" is a terrific circuit and a lot more of an original idea IMO. It does not however sound to my ears much like a TB Mk.II, rather a bit more modern & clear.

I still play through my original Si TB Mk.II I built years ago almost everyday. The sound is much like the rhythm guitar tone on Led Zep's "You Shook Me" if not a touch clearer and more distinct.

The one change that made the most difference to me from the original PNP version of the Mk.II was changing the output cap from .01uF to .1uF, as the ice-pickish fizziness went away and the sound was more balanced.

Hope it works out for you Joe!

Regards

RDV

Joe Hart

RDV, is the one that sounds like "You Shook Me" an NPN Si Tonebender or an original PNP Ge one?
-Joe Hart

Joe Hart

Gus, I did adjust the 470K. I did so many swaps on the breadboard, that I lost track of what I ended up using! But it ended up sounding good, so I stopped swapping stuff around! Thank you!!

I still can't get a good tone control happening, though. I guess I'll just go with something like a BMP at the end of the circuit. Any thoughts?
-Joe Hart

RDV

Quote from: Joe Hart on October 29, 2008, 05:08:48 PM
RDV, is the one that sounds like "You Shook Me" an NPN Si Tonebender or an original PNP Ge one?
-Joe Hart
I was referring to the Silicon version I 1st built. Works great for the "Wolfmother"(band) tone as well.

RDV