Building a Univox Super Fuzz

[twabelljr]

"Here is a chart of voltages taken at the transistor pins. Use these voltages as a guideline. You won't get these exact readings, but they should be somewhere in the area."
Not excact, but close.

[duck_arse]

excellent. I was going to post that yr transistors were wrong connected for 2N or BC, the more common types encountered, but the 2SC types are different base position. and yes, the collectors of those 2 are touching (shorted together//connected//commoned), as are the emitters of those 2. and the base connections are what makes it 'differential'.

I BB'd Q1 and Q2 tonight, with both low hFE's and way high hFE's, and there was not enough difference in the voltages to tell you about. some, not a lot. and Q1 relies on the DC feedback from Q2's E to bias properly.

bias resistors. they apply a bias voltage to something, that which is biased. biasing provides a starting point, a quiescent state for something: a transistor, a fet, an opamp. so a resistor string from + to - producing a bias voltage are bias resistors. the [edit : 47k the 47k is part of the AC loop, not the DC bias] 470k and 100k between Q2 E and Q1 base are also biasing resistors. sometimes the bias is fixed, and the operating point is shifted (because of ohm's law) by a collector pot, sometimes an emitter pot. see the fuzzface biasing methods. and sometimes the transistor biases itself, with a resistor between collector and base. this depends on device characteristics, so can vary part to part.

and we bias to an operating point, because we sometimes want max output swing without clipping, and sometimes we want odd clipping, so we move more towards either 0V or supply.

[chuckfalcon]

Hi again!  OK I took my BB down, and rebuilt it.  My dying battery sound got solved, but it didn't sound quite right and my collector voltages on Q4 and Q5 were 3.5 and they should have been closer to 2.1.  So I took down the BB again, this time not completely, I left Q1-Q3.  I rebuilt it and I got even closer.  The Q4 and Q5 collectors are 2.5.  I haven't tested out the sounds yet because its kinda late, but here is what I've got so far:

Q1 B=0.67 C=5.32 E=0.12

Q2 B=5.33 C=8.49 E=4.72

Q3 B=3.35 C=5.71 E=2.78

Q4 B=1.5 C=2.5 E=0.91

Q5 B=1.5 C=2.5 E=0.91

Q6 B=1.07 C=3.6 E=0.46



I connected the Q4 and Q5 collectors with an orange jumper.  Does that seem correct?


Thank you guys for all of your help so far.  I, like you guys are obsessed with gear and it was bound to happen that I was gonna start building the stuff, haha. 

[chuckfalcon]

I spoke too soon.  I have been messing around with the super fuzz again, and the collector voltages of Q4 and 5 are back up to 3.3.  I am not sure what to do to tame this thing.  Any ideas?

[duck_arse]

have you had a listen to it yet?

yr breadboard. one thing I've always done is allocate the closer upper rail to V+, and the lower closer rail to 0V, and then strung between. the 2 outer rails I use for really long links, or for V/2, sometimes. so, all your collector resistors would be on the upper half board, all yr emitter resistors on the lower half. the transistor goes either/both sides, depending on other connections, and then use bare wire links/staples to connect across the gap, where needed.

the voltages. what appears at the emitter is detemined by what is applied to the base (the bias voltage). the emitter resistor then determines, via Ohm's law, the current through the transistor. emitter current (I'm simplyfing/leaving stuff out/generalising) is the same as the collector current, and Mr Ohm tells us the collector voltage will be Vsupply-(collector current x collector resistor). is your bottom resistor for Q4/5 a 1k8 or 1k5? running the numbers here says 1k5, but they look right enough.

if you search this here forum, you'll find more accurate explanations of how, why and Ohms, and with diagrams, from people who do know what they talking about.

[chuckfalcon]

have you had a listen to it yet?

yr breadboard. one thing I've always done is allocate the closer upper rail to V+, and the lower closer rail to 0V, and then strung between. the 2 outer rails I use for really long links, or for V/2, sometimes. so, all your collector resistors would be on the upper half board, all yr emitter resistors on the lower half. the transistor goes either/both sides, depending on other connections, and then use bare wire links/staples to connect across the gap, where needed.

the voltages. what appears at the emitter is detemined by what is applied to the base (the bias voltage). the emitter resistor then determines, via Ohm's law, the current through the transistor. emitter current (I'm simplyfing/leaving stuff out/generalising) is the same as the collector current, and Mr Ohm tells us the collector voltage will be Vsupply-(collector current x collector resistor). is your bottom resistor for Q4/5 a 1k8 or 1k5? running the numbers here says 1k5, but they look right enough.

After preparing a house show over the weekend, I am now back to the pedal grind.  The schematic says I should have a 1.8K resistor attached to the emitters of both Q4/5.  I traded that resistor for a 1.5K resistor and THAT changed the collector voltages to 2.3.  I way more accurate reading opposed to the 3.0-3.4.  When the collector voltages were high on the Q4/5 the sound was more distorted than what I think its supposed to be.  With the new resistor and collector voltages more accurate, it sounds closer to what I consider the "super fuzz sound.

One question, how do you know when your circuit is absolutely 100% correct and accurate to the original?  Is it a combination of sound/functionality VS. transistor voltages?

The  way that I am hearing the pedal now is like this.  There are 2 different fuzz sounds, selectable with the switch.  One is more big and bassy, and one is more trebly and articulate.  I can dial the treble about 3/4 of the way up on my amp and the "more bassy" fuzz sounds great but then when I switch to the more trebly fuzz it an overkill and kinda obnoxious.  If I dial back the treble on my amp and turn on the "trebly fuzz" it sounds great but then when I switch to the more bassy fuzz tone, it still sounds big but kinda dull and not articulate.  What do you think about that?  Does that seem right?  Maybe I need to buy a vintage one and A/B them.

I still have THIS BB set up.  Im probably going to tare it down again and utilize the positive and negative rail method that you last mentioned and see where I get.  I know the way that I have set up is more prone to shorts and confusion.

Here are some more questions that I have been wondering:

1.) what are some of the software programs that people are using for circuit layouts.  I see people posting diagrams and digital vero boards all the time.

2.) which would be better to use for guitar pedal circuits, vero boards with copper strips and holes in them, OR vero boards that have circular copper pads that aren't connected.

[duck_arse]

hey chuck! why don't you open a few cans of worms while you're at it?

look up ohm's law. look up biasing transistors. doesn't matter if you don't grasp it all at once, but make a start. it will illuminate your resistor//collector volts and their changings. it may help you understand why the transistors do less bossing in this circuit than in a bigg muff, as well.

100%? when it works. I'm not the person to ask a question like that. I've had about 2 "original" pedals pass through my hands, ask a gear-hound like the inestimable (Mr.) Hammer, for inst.

as for softies, diy layout creator for the veros and perfs (use different colours for each track so you can see what is; and it drives some people nuts as a bonus). eagle for general circuit drawings, and it does pcb layout as well. you can't go wrong with either of those, and there are threads for problems//help hereabouts.

vero takes a certain mindset to work with. I've used it and hated using it for about 35 years now. I've only just started using pad-per-hole type perf board, and it is a revellation. it works a bit more like a pcb layout, with tracks can go N-S AND E-W, sometimes at the same time! try that with yr vero. but it's personal preference, it won't hurt you to try both, there always seems to be another something needing building.

[chuckfalcon]

OK everyone, I would like to post what I have so far on my build.  I am happy with my successes so far, and I am super grateful for all of the contributors on the forum helping me out so far.  So, I would like to show everyone my progress of my Univox Superfuzz layout and build.  I haven't done the enclosure yet.  Still waiting on some punching tools so I can do some accurate drilling in the enclosure.

Here is the layout that I did in DIYLC:

Behold:


Behold again:


Behold the back (there was a solder bridge on the first cap in this pic.  I have fixed it since then):


Here is my current wired version:


I A/B'd this one with my Wattson Superfuzz and they sounded identical. However, the Wattson Super Fuzz has an internal trimmer on the base of Q4 and Q5.  The schematic I was using didn't have the trimmer, so currently it doesn't have one.

Progress is good so I decided to show some today!

[digi2t]

Smart looking build. Daddy likes.

The trimmer between Q4/Q5 is to balance out any possible mismatch between the two transistors. As such, you can tune the octave effect to your liking. Otherwise a carefully matched pair should do fine.

[bool]

You can use that trimmer in two ways: to balance the octave generator or to un-balance it (aka. to add some "flavor" to the tone).

You can balance it "by ear" - simply turn the trimmer and look for the spot where there is minimum of feedtrough (when there is no signal). Ie where the noise seems minimal. If carefully balanced (and if transistors and other components match sufficiently to start with), it will seem a tiny bit "gated" but often in a good way.

Of course, you can deliberately mismatch and unbalance it, if that goes with the tone you're after.

[chuckfalcon]

I am reviving this thread a bit just to show some progress that I have made with my super fuzz builds.  Since I last posted, I decided to make all of the transistors socketed.  I also found out what HFE was.  From my research, Superfuzz Tranny's measure in the 100-200hfe range.  The 2SC828's that I have been using are 250-330 range.  SOOO, I went with some 2N2222A's.  I also ordered some lower gain 2SC828 (Q) NOS trannys that I haven't tried out yet but they are more accurate to the original from my studies.





Enjoy.  Let me know what ya think.

[antonis]

We think that there are many small drops of solder on unused pads which are potentionally vibe-helped shorts...

I know that resoldering them it's hard and time consuming so you can cover the solder side of the board with a hard laquer or similar - and forget anything concerning future troubleshooting measurements...