Silicon Tone Bender MKI

[polifemo]

Any updates on this?

[mac]

Any updates on this?

mac

[cmdrfun]

Hi, hopefully this thread is not too dusty now... I just found it recently when looking for info about piggybacking. Wow! Awesome stuff! 

I did my own science experiment starting from the last schematic that mac posted.  Instead of using low-gain Si transistors, which I don't have any of, I built it using only 2N3904's, but piggybacked them for Q2 and Q3. In the buffer stage I don't think that the actual part for Q1 is very important, in fact if I remember my textbook correctly, the higher gain the better to reduce the output impedance of the buffer stage. Anyway, I just used a single 2N3904 there.

In addition to the piggybacking, I also used the B-E junctions of some OC44's to provide leakage. I tried some 1N34A's and 1N60s that I have, but they didn't leak much at all, and really had no effect. The OC44's both had hFE of about 40 (using R.G.'s measuring technique).

I found that piggybacking two 2N3904s, using an OC44 for leakage, and a 15K resistor between the emitters of the 2N3904s gave me an equivalent hFE of about 95 - 105. This was measured just plugging straight into the multimeter.

The only adjustment I made to the overall circuit was to change R6 to 24K, because the collector voltage on "Q2" was over 9V when the Fuzz pot was at a minimum. This swap dropped the collector voltage on "Q2" to 8.56V.

I really like the sound a lot! Not having ever heard an actual MkI Tonebender, I have no real reference to compare to, but it sounds like I'd expect a late 60s fuzz to sound. Thanks for all your work mac!

[mac]

In addition to the piggybacking, I also used the B-E junctions of some OC44's to provide leakage. I tried some 1N34A's and 1N60s that I have, but they didn't leak much at all, and really had no effect. The OC44's both had hFE of about 40 (using R.G.'s measuring technique).

If you have germs, build the real thing! 
Interesting what you did, RG suggested to use a reversed biased germ diode from C to B to simulate leakage. I experimented a bit and it works fine, but the problem is to find a diode with the exact amount of leakage for a given transistor gain.

mac

[cmdrfun]

If you have germs, build the real thing!

Ha, yes, I got the irony of using a Ge transistor to simulate a Ge transistor.  But in this case, the hFE of the OC44's is too low for them to work well in a fuzz circuit, and I don't have any leaky Ge diodes. (Unsure where to procure such a thing since I don't have an old stash of parts lying around.)

[moterrac]

Apologies for bringing this thread back from the dead.. Has anyone got any tips with this, i've been trying this circuit with slightly higher gain transistors but with no luck what so ever. I have ordered some BD175s so hopefully i'll have better luck with that. The circuit seems to work - just haven't really got a distorted sound. I've been looking at the FZ1 diagram mac posted also.

[moterrac]

I've spent the last few days fiddling with this, just recieved some low gain silicon transistors what work great in this project! can't seem to get much control with the attack knob but it sounds massive. Very good sound!

Edit: So I attempted to make a vero layout of this, not sure if it works but here it is if anyone is interested.

[mac]

I just read your last 2 posts.
Glad to read you made it work.

Just for the record, can you post transistors and hfe?

mac

[moterrac]

When testing I used 2n2369a's for q1 - 3 but when I soldered it to tagboard I accidentally used a bc108 for q2 (no idea of the hfe) but the hfe for the 2n2369a's was around 90. I'm using a 50k pot in place of the 25k and getting the effect cutting out on when the pot is at one extreme and then a high frequency roll off on the other extreme, in the middle the pot works well, adjusting the bias of q2. I will continue breadboarding the circuit, got some bd175 and bd237s so will try them in the circuit also, I was thinking about incorporating an additional fuzz control on the base - ground of q3 like you did on your fz1 silicon conversion, have been trying to think of ways to ads additional controls, perhaps controlling q3 collector? In all honesty I'm pretty new to electronics and pedal building so been trying to absorb as much information as possible. 

[mac]

You can control Q3 collector with a 25k pot from Base to Gnd.
A fixed 100k to 220k from Vcc to Base can do it, IIRC.

If you want the FZ-1 slow attack I posted some months ago try this:
1. No interstage resistors
2. 22uf before T2,3
3. Use whatever booster or clean disto you have instead of the BMP input stage I added to the Maestro.

mac

[Steben]

Just revived the thread because I couldn't play the sound samples... 
Reminds me of my "why does a TB mkI sounds good" stuff. The main conclusion was high pass filtering effect in the third stage, making the unit arguably the first transistor based distortion rather than fuzz.
Any transistor clipping with the fitting filtering would come close.
Silicon transistors usually have higher internal resistance, so any device with the same hfe might have higher input impedance than a germ one.
Said that, a look in this light at the last circuit brings the conclusion the high pass effect before the clipping q3 will be less. Especially with R16 in play. This will result in less treble boosted tone the original has. Lowering the 8k2 resistor will give higher cut off point.


Siliconising germanium circuits is nice. My favourite Fuzz Face is a silicon one.

[Frances Rhodes]

hello all

I've been working on this version of the mkI and I don't know if I'm there or if I still need to tweak the values a bit.
here's a quick demo on my girlfriend's vox AC4, standard E tuning and drop A tuning.
https://youtu.be/FXaZXsIRXJI
what do you people think?
cheers

[Steben]

hello all

I've been working on this version of the mkI and I don't know if I'm there or if I still need to tweak the values a bit.
here's a quick demo on my girlfriend's vox AC4, standard E tuning and drop A tuning.
https://youtu.be/FXaZXsIRXJI
what do you people think?
cheers

Sounds great beyond the gating. Probably a finetuning thing of the bias in each stage. How did you build the stages? Collector - base resistor?

[mdcmdcmdc]

Sorry for the necrobump - just curious if anyone happens to have the schematics for this project? The image links seem to be lost to time.

Thanks!

[m_charles]

I have nothing to do the rest of the year except going to the beach, so I'll try your idea 

Go ahead, rub my face in it. 

Since this circuit calls for leaky germs I think that new germ diodes have little leakage for it, most likely it will work with old ones.
BTW, a reversed diode and its leakage current can be considered like a big mega ohm resistor from B to C?

Taking the question first: no, not quite. A reversed diode acts more like a constant current source (or leak) than it does a resistor. The leakage on a reverse biased semiconductor junction rises over a small-ish few volts of reverse bias, then flattens out a LOT. It's not perfectly flat, but the leakage remains reasonably constant, not rising linearly with reverse voltage like a large resistor would. It was this idea that actually kicked off the Millenium Bypass. It's very, very difficult to get resistors that are consistent to any degree, and much larger than 1M, which is what the Millenium circuit needed.

That being the case, going to an external device to supply "appropriate"  leakage is much simpler. You can choose diodes for leakage, using old unsuitable germanium devices for their collector base or base-emitter junction. You can use two (or more) modern germanium diodes with low leakage for germanium, but still 1000x the leakage of silicon til you get the leakage your circuit likes.

This offers a degree of freedom in using transistors that picking among old germaniums does not.

In combination with piggybacking to lower gain, it may well be possible to trim in a generic silicon transistor to sound like an old, low gain and leaky germanium.

Wow. Who'd have ever thought that there would be a use for deliberately making things worse. 

How much leakage do you like to see RG?

[bmsiddall]

Sorry for the necrobump - just curious if anyone happens to have the schematics for this project? The image links seem to be lost to time.

Thanks!

[bmsiddall]

Btw, I ditched the interstage resistors and left out the C-B caps on Q2/3.  I used Si transitors that had similar hfe to my Ge version.  Sounds great- maybe a touch darker and a little more drive than the Ge.

[mac]

Btw, I ditched the interstage resistors and left out the C-B caps on Q2/3.  I used Si transitors that had similar hfe to my Ge version.  Sounds great- maybe a touch darker and a little more drive than the Ge.

Glad you like it.
And thanks for adding a link to the circuit.

mac

[mdcmdcmdc]

Thank you both!

[Steben]

Btw, I ditched the interstage resistors and left out the C-B caps on Q2/3.  I used Si transitors that had similar hfe to my Ge version.  Sounds great- maybe a touch darker and a little more drive than the Ge.

The original MkI last stage's circuit determines a lot of the EQ and sound which is far less "flabby" than most vintage fuzzes. It is the most driven stage and has the lowest input impedance which interacts with the 0.1uF cap.
In the germanium one we have a Re of about 25 ohms, which means the impedance at the base is about a couple of kohms with a hFE of 100 on average.
Any change in hFE (esp. higher) or swapping to silicon which might have higher internal Re might raise the input impedance enough to lower the corner frequency. The stage is less "hampered" in the lows. This means more low mids get amplified and driven which sounds to me like as described.
Lowering the 8k2 resistor might compensate.