Tonbender MKII Discussion on input High Pass filter and transistor performance!

[timbo_93631]

As continued from this thread: http://www.madbeanpedals.com/forum/index.php?topic=4595.15

So how does that affect transistor performance?  For instance, what would you gain by having a 1k with the 10uf cap for a corner frequecy of ~16hz?

[JakeFuzz]

I know it changes the input impedance into the first transistor. Probably effects the matching to your guitar pickups. I am not sure about specific values though because I think you have to take into account the impedance of the emitter junction as well.

[mgwhit]

So how does that affect transistor performance?

My concern -- and I should warn you in advance that I'm talking out of my butt here -- is that the transistor could amplify low-frequency or even subsonic noise that sneaks into the input signal and essentially waste some of its amplification power on unwanted or useless frequencies.  I've heard of sub- and/or ultrasonic frequencies causing oscillations in tube amps.  Don't know how transistors would be affected by this.  I'd love to hear someone who actually knows what they're talking about comment on this so that I can stop making stuff up.

I should add that I know that people have been building and playing through Tone Benders for decades and that there's nothing wrong with the stock high-pass filter.  I originally chimed into the previous conversation because I just thought that was the wrong place to add bass response to the circuit.

For instance, what would you gain by having a 1k with the 10uf cap for a corner frequecy of ~16hz?

I've got an OC75 MkII on my breadboard right now (essentially the GuitarPCB Pumped Up Tone Bender), and it's noisy as @#$%.  I wanted to get the electrolytic cap out of that early position and replace it with a (less noisy?) film cap, so I've been using 1uF/100K, which should have the same corner frequency as the SupaFuzz 10uF/10K (1.6 Hz).  I was hoping that with corner frequencies that low I wouldn't be able to tell a difference, but I honestly felt I could hear a slight bass roll-off, which is why I'm wondering if anything above and beyond the RC values could be affecting the performance of that high-pass filter (the 10n to ground?  input impedance?).  I didn't exactly do a scientific double-blind experiment on this, so I'll try a few more values tonight and see what happens.

[timbo_93631]

Hey, thanks for the info.  I am wrapping up summer school this week, 7 units in 6 weeks has really put a damper on my pedal building.  Maybe this weekend I'll breadboard the Supafuzz and look at some parts sub interactions.  I am gonna use OC139 in Q1 and Q2 in the higher gain position (~80hfe) and OC 141 in the lower gain position (~100hfe). 

[mgwhit]

Okay, I just tried the following input cap values (all with a 100K resistor to ground) on my lunch break: 10uF, 4.7uF, 1uF, 0.1uF (0.2Hz-15.9Hz).  Verdict: I wasn't sure if I could tell a difference.  Nothing was too woofy, nothing was too shrill.  And this was using my telecaster through a 5F2A Tweed Princeton clone, so the potential for shrill and/or woofy was there.  Q1 obviously performs just fine with a ridiculously low corner frequency on the input high-pass filter.  I'm gonna stick to the 1uF/100K values just to keep as close to stock as I can without resorting to an electrolytic cap.

Hey, thanks for the info.  I am wrapping up summer school this week, 7 units in 6 weeks has really put a damper on my pedal building.  Maybe this weekend I'll breadboard the Supafuzz and look at some parts sub interactions.  I am gonna use OC139 in Q1 and Q2 in the higher gain position (~80hfe) and OC 141 in the lower gain position (~100hfe). 

Yeah, go nuts.  All of these fuzz circuits have such a low parts count that they're perfect for breadboarding.  I'm going to monkey with the Attack knob cap next.

[timbo_93631]

Summer session is over and I have been playing with this on the breadboard all weekend!  I have found that the combination of 10k/10uF 100k/5uF doesn't make nearly as much difference as changing the value of the other 5uF/10uF cap or the transistors.  I'm still working on getting things biased right too.  It seems like there is a high pitched whine with the attack control maxed, and the noise/hiss is incredible even though I used 1% metal film after thinking it was too crazy with carbon comps.  So far I can't see much difference between this and a MKIII aside from it being worse/noisier/lacking in tone shaping controls and adjustable bias (pastyface).  Maybe I am just more of a MKIII guy...  Will keep at it today and see where I end up.

[timbo_93631]

I made a few tweaks and things are sounding good, gonna build it tonight, got the PCB done this afternoon.  I replaced the 100k (MKII)/47k (SupaFuzz) with a 100k trimpot for adjustable bias, and dropped the resistor between the Base of Q2 and Emitter of Q3 from 100k to 56k.  Everything is sounding really groovy now and using OC139's in all positions, low gain orientation on Q1 and Q2, high gain on Q3, this are good.  Will post sound clip when complete.  FWIW, changing the HPF corner frequency a bunch of ways didn't make a whole lot of difference in tone, but it did seem to affect overall fuzz levels.  When I tried 100k and 5uf it was fuzzier, 10uf and 10k was more defined at full fuzz and what I stuck with, so I have a SupaFuzz.

[mgwhit]

Cool.  How's the noise level?  I can't get excited about boxing mine until I get that figured out.  I'll play  with your tweaks on my breadboard tomorrow.

[timbo_93631]

     Oh the noise...  Yeah, that seems to be the issue.  Going with the OC139's I am running at fairly low gain in Q1 and Q2, around 40 hfe, Q3 is reversed and I'd say maybe 80 hfe.  I randomly picked through my latest stash for the ones that made the least noise in Q1 and Q2 so gain ratings are guesses based off random sampling done by the seller which showed them to be accurate to the ranges listed in the datasheet.  I tried 140's and 141's which are higher and much higher gain in all positions but noise increased with hfe's.  I think one thing is that you want Q1 and Q2 to be fairly close in terms of gain, like in a Fuzz Face.  MKIII's are way more forgiving in that respect and can sound good with any number of wacky combinations. 
     With this guy I wouldn't suggest too many carbon comps, if any.  The only one I used was a new stock Xicon for the 56k swap mentioned above because it was in my loose parts tray in front of me at the bench.  Sounded better than with the 100k so it is staying mojo or not.  When I loaded everything onto the PCB and just hooked up the I/O's and controls thigs were also a bit quieter than on the breadboard.  I am hoping that getting it into an enclosure with do even more for it. 
     I tried going to a 2KB Attack pot as metioned in the Fuzz Central article on all things MKII, but I found that though things were smoother and quieter it lost some character and insanity in the max fuzz setting and the range was all bunched up at the end of the knob travel, maybe by playin with some resistors across the lugs you could get a better result with a ~1.5k pot, but I am just gonna keep the 1KB.  I'm interested to see what your experiences are.  Have fun breadboarding!

[timbo_93631]

     With the setup above ^ I was still getting too much noise when using my Gator G-Bus-8 power supply.  I think it would have been acceptable with an isolated power supply like a Voodoolabs unit, but I had to dig deeper to get rid of some more noise.  This involved much disassembly and re assembly, but it was so noisy on the breadboard I really think it was better to do it in the enclosure, just much more time consuming. 
     I bootstrapped a 33uf dry tant I had on hand to the 47uf filter cap so it is now ~82uf on my LCR meter, added shielded cable to the board input and output off of the level pot, grounded at only one end on each cable, and put a 47 ohm resistor from+9v on the dc jack to to board 9v+.  I replaced the 56k mentioned above with the stock 100k value and replaced the 470 ohm resistor with a 1k as well.  The best combination of transistors I could find was an OC139 in Q1 and Q2 in high gain position which are about 80hfe, and an OC141 in low gain position in Q3 ~120hfe.  I also replaced the 10uf cap in the input high pass ilter with a 10uf dry tant.  Low noise, big tone.  Really digging it now that the noise is under control.

[Scruffie]

You really need to play about with the trannies in this one, you need high leakage in Q1, and swapping things around can make huge differences. Some Ge trans just are noisy too so a good selection is handy. Bias wise I find having Q2 as the 'main' bias and Q3 for tweaking/smoothing helps. Amp settings can be a reason for noise too, it's pretty high treble so you may need to compensate. Mine took a good 2 hours of swapping to get it to a low noise level (it shouldn't be hissy or whine when things are right) and get the 'fullest' tone, somewhere on the D*A*M forum there's the gain and leakage measurements from an original 81D model and they're not what you'd think (as a point of interest I actually didn't really like 81Ds in the circuit and these were proper mullard originals, not the non brand ones that float about, I ended up with one leaky 81D and two non D 81's) so yeah, don't write it off but it ain't easy at times. Worth it when it works though, kicks the nads off a fuzz face and is fuzzier rather than overdrivey like a MkIII, which it may be you prefer, if you want trouble and noise try a MkI lol.

[9Lives]

Hey, not trying to hi-jack, but what layout are yall using? I've been wanting to do a bender. Do I need similar gains to the FF? Isn't q1 supposed to be a leaky or low gain? Thanks I'll shut up

[timbo_93631]

You can use the one at fuzzcentral and there are all of the variant schematics you'll want to refer to:
http://fuzzcentral.ssguitar.com/mkII.php

There is this one, minus the .01uf input cap to ground:
http://pigeonfx.com/layoutsf.html
I have an extra board from that seller I have not populated yet, but it has to be trimmed to fit a 1590b, otherwise 1590bb is the way to go with it.
Here is the gutshot of mine in a 1590b:


And the epic must read thread at FSB if you are gonna build one:
http://www.freestompboxes.org/viewtopic.php?f=19&t=920