ETI Hyper-Fuzz

[soggybag]

I found this ETI Hyper-Fuzz. Looks like it was in ETI Guide to Making Music 1988. It reminds me of this clipping thing by Bernie Hutchins, maybe it's more closely related to the Crowther Prunes and Custard.



The input sections is really complicated and hs a buffered output. I wanted to simplify this and get rid of the buffered out. I replaced the original input with an LPB-1. Untested so far!

[antonis]

The input sections is really complicated and hs a buffered output. I wanted to simplify this and get rid of the buffered out. I replaced the original input with an LPB-1.

I should consider it as a Fuzz configuration rather than a buffer..
Q1/Q2 Sziklai (Complementary Feedback) pair acts as Q3 Collector load while providing "Fuzzy" essential feedback..
(plus that Q2 output signal is inverted..)

[iainpunk]

Im contesting Antonis, as i believe it is a buffered gain stage.
R1, R2, R3, C2 and Q1 form a stable constant current source, creating a high gain gain stage with Q3. Q2 acts as a buffer, not a Sziklai, as it has a tap point in between them where there shouldn't/can't be one for a proper Sziklai operation.

Its basically a diode ladder wave folder pushed by that gain stage.
This isn't a 'full fold' wave folder tho, its more of a wrinkly clipping threshold, instead of a 'true' fold. I have build this type of topology on my breadboard a long while ago, and compared to some of the other wave folder based circuits i built, this one is super milquetoast (which is probably a good thing for 99% of guitar players)

Cheers

[antonis]

i believe it is a buffered gain stage.

Agree with "gain" stage..
Disagree with "buffered" one..

P.S.
Just sit on Q3 Base at look around from impedance point of view..
(also considering R5 added on signal source output impedance..)

[antonis]

@Iain, come on..!!

I'm pretty sure I'm 50% right, so I'd like an argument for the rest of 50%..

P.S.
Of course, OP reserves the right to have both of us banned from his thread..

[Mark Hammer]

Not that this would stop anyone from drawing up their own (weird hobby, that), but you do realize there was a PCB layout included with the project article?

The project can be found here:  https://worldradiohistory.com/UK/Electronics-Today-UK/80s/Electronics-Today-1987-10.pdf , although it would seem that PCB pattern is in the Canadian issue but not the UK one.  I'll have to dig around to see which issue that is.  In the interim, here is a rather poor scan.

[soggybag]

I can't read Canadian Mark, which explains why I missed that!

In my many years Of reinventing the wheel I always search for the most difficult solution to every problem.

That input buffer (no offense Atonis) bugs me so I thought to replace it with something "simple." (Nothing about this hobby is simple)

[antonis]

That input buffer (no offense Atonis) bugs me so I thought to replace it with something "simple."

You can't get the same outcome with a simple booster neither from gain nor from impedance point of view..
Q3 gain is by far larger than LPB-1 gain while Q2 output impedance is by far lower than LPB-1 impedance..

[soggybag]

How do you figure the input impedance here?

For the LPB-1 is it roughly the parallel resistance of R1, R2, and R4?

For the original roughly R5?

[antonis]

How do you figure the input impedance here?
For the LPB-1 is it roughly the parallel resistance of R1, R2, and R4?

Yes but you omit the most dominating item which is Gain pot unbypassed value X h_FE..
For Gain pot set FCW, it turns into 0.026/I_Collector X h_FE..
Considering LPB-1 Collector biased at about 4.5V, Collector current should be about 450μA hence Base input impedance is about 57R X h_FE..
(less than 10k for a BJT of h_FE = 150..)
You can realize that the above impedance isn't "linear" 'cause Collector current isn't constant hence we have some kind of distortion on LPB-1 output..

For the original roughly R5?

No..
It's more complicated here due to Q2 Emitter NFB action on Q3 Base..
By lifting up R8 upper leg and bringing it to Q3 Collector (or ignoring Q2 V_BE voltage drop for clarity) you effectivelly get Q3 self-bias configuration..
So, R6+R7+R8 "appear" to Base as a resistor of apparent value the sum of them divided by Q3 stage gain plus unity..
(you can consider Collector standing at -(Base voltage X gain) and calculate current flowing through R6+R7+R8 resistors or search for "Miller effect" - http://web.mit.edu/klund/www/papers/jmiller.pdf)

edit: At a second glance, R10/C4 and RV1 decouple half of feedback resistor (junction of R6/R7) so this branch input impedance simply is R6+R10+RV1..
(the way from Base to GND)

So, total Q3 stage input impedance is the parallel combination of (R6+R10+RV1) // (0.026/I_Collector) (grounded CE amp)..
(much more linear here due to roughly constant Collector current..)

To make things even worse, R5 is (or should be)* considered in series with signal source impedance forming a voltage divider with Q3 input impedance and bias configuration one..
(*) You may find in some textbooks anaysis R5 considered in series with stage input inpedance and then contributing to stage voltage gain domination, due to voltage dividing effect..
Personally, I prefer to add R5 on previous stage output impedance and consider it in series only for stage gain calculation..

[soggybag]

Thanks for the explanation Atonis. I'm trying to wrap my head around this. I might have some questions!

[antonis]

I might have some questions!

HOW PECULIAR..!! 

P.S.
Anytime you like..
(but each project at a time..)

[iainpunk]

@Iain, come on..!!

I'm pretty sure I'm 50% right, so I'd like an argument for the rest of 50%..

P.S.
Of course, OP reserves the right to have both of us banned from his thread..

sorry for the late reply, i had a busy weekend and first half of the week, haha.
i believe you are exactly 50% right, indeed.

looking over the function of the Q1 and Q2 together i realised you were right in that they were working together as a Sziklai pair, but its functioning is partially inhibited by the ''tap off'' current being conducted by Q1. so its a really clever circuit doing 2 things as once.

what i meant with 'buffered' gain stage it that it can be percieved as Q1 and Q3 forming a gain stage, and Q2 is an emitter follower (A.K.A. a buffer as it doesn't add gain) to lower its output impedance, especially as it works with a current source load.

respectfully yours,
Iain

[PRR]

> looking over the function of the Q1 and Q2 together.....

You mean the plan in post #1?

Q1 is a plain current source, hi-Z load. Q2 emitter follower preserves the hi-Z. Then Q3 is a simple grounded-emitter gain stage with an "infinite" load giving "infinite" gain (500-1000). The resistor NFB limits gain to like 450 to 6. The input impedance is worse than figuring US income tax, but indeed R5 330k sets a lower limit and we probably don't need to figure any closer.

[Chris Dancer]

Hello guys, I designed the ETI "Hyper-fuzz" (cringeworthy name, I know). It came about when I was pondering what would happen if you could apply a sine function to a musical waveform - what would that sound like? This is the closest I could get to it with the limited set of components available to the average electronics hobbyist at the time. It was all done with paper and pencil and a lot of design tweaking on a breadboard. The PCB design was done old-school: crepe tape and rings on acetate film. I was very pleased with the result, not to mention the £135 I got paid for it by the magazine. Kind of pissed off that they got my name wrong though!
I remember that one problem with it was that the gain around the zero-crossing point had to be really high so it was very noisy when the guitar wasn't being played. The circuit would be better suited for use in a modular synth where you have a cleaner signal. In fact I've seen "wave-folder" modules on sale that do a very similar thing.

[antonis]

Hi & Welcome..

Finally, your name is Chris or Charles..??

[pinkjimiphoton]

Im contesting Antonis, as i believe it is a buffered gain stage.
R1, R2, R3, C2 and Q1 form a stable constant current source, creating a high gain gain stage with Q3. Q2 acts as a buffer, not a Sziklai, as it has a tap point in between them where there shouldn't/can't be one for a proper Sziklai operation.

Its basically a diode ladder wave folder pushed by that gain stage.
This isn't a 'full fold' wave folder tho, its more of a wrinkly clipping threshold, instead of a 'true' fold. I have build this type of topology on my breadboard a long while ago, and compared to some of the other wave folder based circuits i built, this one is super milquetoast (which is probably a good thing for 99% of guitar players)

Cheers

ummmmm.... i'm usually wrong, BUT

a "buffer" cannot be a "gain stage" as a "buffer" is a "unity gain amplifier" unity gain being slightly BELOW actual unity gain, innit?

[soggybag]

Hello guys, I designed the ETI "Hyper-fuzz" (cringeworthy name, I know). It came about when I was pondering what would happen if you could apply a sine function to a musical waveform - what would that sound like? This is the closest I could get to it with the limited set of components available to the average electronics hobbyist at the time. It was all done with paper and pencil and a lot of design tweaking on a breadboard. The PCB design was done old-school: crepe tape and rings on acetate film. I was very pleased with the result, not to mention the £135 I got paid for it by the magazine. Kind of pissed off that they got my name wrong though!
I remember that one problem with it was that the gain around the zero-crossing point had to be really high so it was very noisy when the guitar wasn't being played. The circuit would be better suited for use in a modular synth where you have a cleaner signal. In fact I've seen "wave-folder" modules on sale that do a very similar thing.

Wow that's amazing Chris! It's really great to hear the backstory on this! It's in my queue to build some time soon!

There are some things I'm still contemplating. The bias voltage is not mid point of power supply?

I'm also still trying to wrap my head around the input stage.

[Mark Hammer]

Hello guys, I designed the ETI "Hyper-fuzz" (cringeworthy name, I know). It came about when I was pondering what would happen if you could apply a sine function to a musical waveform - what would that sound like? This is the closest I could get to it with the limited set of components available to the average electronics hobbyist at the time. It was all done with paper and pencil and a lot of design tweaking on a breadboard. The PCB design was done old-school: crepe tape and rings on acetate film. I was very pleased with the result, not to mention the £135 I got paid for it by the magazine. Kind of pissed off that they got my name wrong though!
I remember that one problem with it was that the gain around the zero-crossing point had to be really high so it was very noisy when the guitar wasn't being played. The circuit would be better suited for use in a modular synth where you have a cleaner signal. In fact I've seen "wave-folder" modules on sale that do a very similar thing.

What a delight to hear from the source!
But then, on a site devoted to how-does-it-work, how-can-I-make-it-with-what-I-have, and how-can-I-make-it-sound-the-way-I-want,  I suppose it should not surprise me.

While we have you here, any other published circuits we should know about that you're connected to?

[Chris Dancer]

While we have you here, any other published circuits we should know about that you're connected to?

I designed a switchable waa / volume pedal which was published in Practical Electronics. I'll see if I can scan that and post it here.