First post: Question about sustain in Tone Bender circuit

[Wavelength]

All,

Hey Gordon here from Wavelength. I have been making tube audio and guitar amps for decades. A year ago in my off time I started making germanium Rangemasters, Tone Benders and folded cascode (FET + 2N404) dual stage overdrives. Not really a fan of opamps so I do discrete stuff mainly.

Ok so I built a kit from Reverb on a Tone Bender 1.5:
https://reverb.com/item/14362871-guitar-gear-workshop-tone-bender-mk1-5-germanium-diy-pedal-kit

I subbed in some better components, some OSCONS on the 9V and Black Gate BGN HiQ for the bypass on the 1K pot and shinko tantulum resistors and it sounds really nice. I also have the JHS Josh built yellow 1966 Tone Bender 1.5 which sounds completely different and has significantly less sustain than my version.

So I was wondering if anyone knew what effects the sustain in that circuit. Is it the feedback (100K) resistor? Is that way the vox and some of the other variations use 47K?

Something else?

Thanks,
Gordon

[bmsiddall]

Hi Gordon.

What voltages do you see on both versions? If q2 collector is close to the 9v rail it will gate faster. Q1 with lower leakage should raise q1 collector voltage which in turn will lower q2 collector voltage for more sustain i think.  Others here with more experience will correct me if this is wrong info 

Cheers,

[antonis]

Hi Gordon.

I can't recall if the first answer on your first post was a "Hi" without "Welcome"..

[bmsiddall]

Hi Gordon.

I can't recall if the first answer on your first post was a "Hi" without "Welcome"..

Gimme a break- it's morning here and I'm still in bed trying to face the day 

(Welcome Gordon)

[Electric Warrior]

Hi Gordon.

What voltages do you see on both versions? If q2 collector is close to the 9v rail it will gate faster. Q1 with lower leakage should raise q1 collector voltage which in turn will lower q2 collector voltage for more sustain i think.  Others here with more experience will correct me if this is wrong info 

Cheers,

No, that's pretty much it. The MK1.5 was usually biased rather hot.
You can make up for it by tweaking Q1's collector resistor.

[iainpunk]

welcome to the forum

So I was wondering if anyone knew what effects the sustain in that circuit. Is it the feedback (100K) resistor? Is that way the vox and some of the other variations use 47K?

the specific transistors make huge difference in Fuzz Face circuits, not the type, but the specific ones you used, they make a lot of difference. i fyou have 10 of the same type, they will all sound different in the same circuit.

changing the bias changes the gain as well, add a bias trimmer if you want?

cheers

[Electric Warrior]

the specific transistors make huge difference in Fuzz Face circuits, not the type, but the specific ones you used, they make a lot of difference. i fyou have 10 of the same type, they will all sound different in the same circuit.

changing the bias changes the gain as well, add a bias trimmer if you want?

cheers

I found they don't make that all that much of a difference. The tolerance of the attack pot can have much more audible effects.

[Wavelength]

All,

Thanks I have both a Semiconductor Lab and one of those Peak and selected the best leakage and 80 hfe Q1, 120 hfe Q2 out of 200 of the 2N404 I have that are all old National Semi's.

I have 0.6 on the collector of the Q1 and 4.5V on the collector of Q2. I mean it sounds nice, it just has so much sustain that picking on it is just a blurr of notes. I did add bias pots on this to optimize it.

On the JHS he is using the CV7003 I think. Of course it's totally different sounding as we all know almost anything effects these types of pedals. But the JHS has the perfect amount of sustain. I think I have some 2N1309's here somewhere which is a cross to the CV7003. Maybe I open this up and see what's cooking.

I have a bunch of old Mullard PNP and NPN, maybe I try that. I will also change the feedback (which I know will change the bias) and see if that does it.

Thanks,
Gordon

[Electric Warrior]

Oh, it's biased nowhere near where it would start sounding farty.

Maybe try matching the hfes more closely or just swap transistors until you find a combination that you like.

And don't fear the leakage. You can bias it a lot "hotter" than that by using leakier transistors. Most vintage Fuzz Faces are biased hotter.

Here are the voltages from my vintage MK1.5 for comparison:

Battery: -9.99V
Q1 C -0.31V B -0.08V E 0V
Q2 C -7.64V B -0.31V E -0.24V

[Wavelength]

EW,

What transistors are you using? The VBE on the 2N404 is 0.35 typically and my understanding is that typical collector voltage for germanium Q1 was 0.6 and silicon 0.8 or 0.2-0.3v above VBE.

Plus what do you mean by hotter? Doesn't that typically mean current? If so wouldn't mine be hotter than yours?

Thanks,
Gordon

[Electric Warrior]

By "hotter" I mean literally, as it will behave with rising temperature.

As temperature rises, temperature, leakage increases, the voltage on Q1C will drop and the voltage on Q2C will rise.

These voltages I posted are from a vintage unit with OC75s. These tend to leak quite a bit.

Collector voltages for Q1 depend mostly on leakage and hfe. 0.3V is pretty much as high as it gets for MK1.5s. Under 0.2V is not uncommon at all.
Vintage germanium Fuzz Faces vary a lot, up to 0.3V on Q1C is not uncommon.

In silicon Fuzz Faces with transistors from the C gain group, Q1C is usually between 1.2 and 1.3V. I got a some vintage BC130C of exactly the same type that was used in some Fuzz Faces and they sure will bias in exactly that range.

As you can see from the voltages, the actual VBE in circuit can be quite a bit lower than the reading on a meter.

[Big Monk]

EW,

What transistors are you using? The VBE on the 2N404 is 0.35 typically and my understanding is that typical collector voltage for germanium Q1 was 0.6 and silicon 0.8 or 0.2-0.3v above VBE.

Plus what do you mean by hotter? Doesn't that typically mean current? If so wouldn't mine be hotter than yours?

Thanks,
Gordon

I find the transistors in my stash (Mullard OC75 and General Electric 2N169) have Vbe < 0.2 V.

I think another thing he is touching on is that the Bender 1.5 is not biased like a Fuzz Face and you have room to go up on Q2C.

Hell, many Fuzz Faces don't bias to 4.5 vDC without some work.

[Wavelength]

EW,

So your saying do a "Benson" on this and then the Vbe may change more to the liking. Agreed the Peak and the Semiconductor Workstation both test at current levels much higher than any Fuzz is running at. But really will any of that change the amount of sustain I currently have. Maybe I should check out some schematics with sustain pots and see what they are altering.

I may go ahead and try some different transistors in here and see what happens.

Thanks,
Gordon

[Electric Warrior]

EW,

So your saying do a "Benson" on this and then the Vbe may change more to the liking. Agreed the Peak and the Semiconductor Workstation both test at current levels much higher than any Fuzz is running at. But really will any of that change the amount of sustain I currently have. Maybe I should check out some schematics with sustain pots and see what they are altering.

I may go ahead and try some different transistors in here and see what happens.

Thanks,
Gordon

I usually don't worry about VBE, but then again I haven't had any transistor that gave me weird voltage readings in circuit...

Q2C looks a little low in relation to Q1C. I guess the Attack pot's resistance may be on the low side. That could decrease sustain a little. If trying different transistors doesn't give you the expected results, you could try adding more resistance there.

[Wavelength]

EW,

Thanks, I think I got the bias voltages from Electrosmash's FF page. Plus I was considering adding a tone control and maybe an emitter follower and the 4.5V on Q2 collector seemed like a great place to start.

I am using a PEC 1K linear on Q2 attach control and 33K on the collector of Q1. I have a 5K 20 turn pot between the -9V and the 33K with a 1uF wima to ground to bias Q1. I may fool with that.... but like I said the sound is great, it's just the damn thing is on sustain steroids.

I have some stupid work to do then I will get back on this. Plus AES meeting tonight via zoom.

Thanks,
Gordon

[Wavelength]

All,

Thanks for the help, I think I got it... So I looked at a couple of the Tone Bender units I had bought and sure enough.... the Vbe on Q1 was usually around -0.102 to -0.11 and mine was almost 0.2 and on Q2 the Vbe was between -0.114 and -0.12 which means I was biasing my transistors with a lot more current. Vc Q1 was typically around -0.5V and Vc Q2 was around 5V.

I made some changes including Q1 load to 47K from 33K and Q2 load up to 12.5K and things sounded better, but a lot more noise. So I bumped the values back down to what they were and added a dropping resistor from the -9V rail (3.3K) and bypassed that with a Wima 1uF/50V cap and everything looked pretty good. Vbe on Q1 is -90mV (may increase the feedback from 100K to 68K which would bring this up a little) and Vbe on Q2 is -101mV and it sounds killer. Q1 Vc = -0.5 and Q2 Vc = -4.8V.

Considering making one up now with an output stage and tone control.

Thanks again for all the help.
Gordon