Saturday 2 June 2012

NPN Silicon Fuzz Face - Germanium Emulation

From an older thread on FSB and DIYSB about using two pairs of piggybacked silicon transistors to have control over the gain.  To paraphrase, these were the main points:

1. Take two identical random silicon transistors (for more mojo: your fav fuzz silicons)
2. Tie the BASES together,
3. Cut off one COLLECTOR
4. connect a 3k to 6k resistor between the EMITTER of the collectorless device and the EMITTER of the other transistor.

Someone mentioned using a 200K trimmer but I suspect that would be too high and that a 20K trimmer would probably be fine with most common transistors, allowing better resolution in fine tuning the gain.

The basic idea is that altering the resistance between the two emitters using a trimmer allows you to control the gain of the transistor pair, which lets you fine tune an effect to get the sweet spot of gain every time.

This could be used on any of the classic germanium effects, but I thought I'd try it first with a Fuzz Face to see what the results would be like.

Redhouse on FSB wrote:

How does it sound? ...IMHO with the right resistor it sounds much better than Ge's, stays tight in the low end where Ge's start to mush-out, much less noise (hiss) and you get away from that temperature problem the Ge's have.

To find the right resistor for your two transistors temporarily install a 200k trimmer where the resistor goes, plug in your unit and play a playing volume (not bedroom levels) and dial in the trimmer untill you arrive at where you bet the best sound, remove the trimmers and install resistors.

I still have a small stash of AC128 and AC125 I got from a guy in the UK a few years ago (not selling any) I save for Ge builds that "must have" them, but IMHO the Bret-Piggyback thing sounds way better.

Give it a try and see what you think:



And if you want to include a filter cap, which may be a good idea considering how silicon Fuzz Faces can be prone to oscillation:


36 comments:

  1. Seems like a nice effort for making the FF better. Personally i just don't get it. Why build FFs over and over again - there must at least 100 better silicon fuzzes out there that offer the same smooth, light fuzz.

    I'd be building this though, if i just had the trimmers in stock :)
    +m

    ReplyDelete
  2. Like what? A properly functioning Fuzzface has a feel and tone I haven't found anywhere else, IMHO.
    The piggyback idea is great, but if you source some low-gain Si trannys, it is a bit of a non-issue.
    +100 for avoiding Ge when possible. Love the sounds, but DAMN are they a pain in the butt in AZ where the temperature swings are 40 degrees F a day!

    ReplyDelete
    Replies
    1. That's true, the 2N3903's a great for this kind of thing, but no matter what transistors you get you're still stuck with a single current gain using a single transistor. The point of this is that for the sake of adding a couple of extra transistors and a couple of trimmers, it will allow you to fine tune the gain exactly to get the response you want. You could even have it as external Gain 1 and Gain 2 pots which would be interesting.

      We don't have that heat problem in the UK, our main problem is our hands turning blue because we're cold :o)

      Delete
    2. Skreddy's Lunar Module for example - i simply fell in love with it. And once i tried my latest Ge FF and that after each other... Bosstone sounds like viable alternative too. This is good example in question of taste though. :)

      Lunar module uses low gain (60-350 hfe) BC109Cs. Likewise, 2N2222's datasheet says just minimum of 75 hfe. Both are fairly common and they do not cost that much when compared to their quality and sound.
      +m

      Delete
  3. At the risk of sounding stupid (a risk which I gladly accept), how do you make this Frankenstein transistor without it looking like an unholy mess on your board? And does the above FF project have 4 of these transistor combos or is 2 pairs? Thanks

    ReplyDelete
  4. The way i read the layout is that you don't have to cut or bind anything. The transistors are palced so, that unused collectors don't connect to anything. So just 4 transistors are needed. And the trimmers are supposed to be added permanently, instead of what the forum thread says.
    +m

    ReplyDelete
  5. mirosol is correct. You could cut the collectors if you want, but I've done it so that even if you solder all the pins in, the collector is isolated from everything. And the links sort out the required connections.

    ReplyDelete
  6. Great. Thank you both. I think I will try this one this week

    ReplyDelete
  7. This one sounds really great. Ive always found fuzz faces to be dark and the low end loses all calrity when cranked, but that doesnt happen with this one. The bias knob allows for a huge range of classic fuzz tones too. Dig it.

    ReplyDelete
    Replies
    1. Nice one, thanks for verifying. How did you set the trimmers and do they seem to give you good control of the gain?

      Delete
  8. One more question and I swear I'll leave you alone... until next time. But, I notice that the trim pots are 3 in line as opposed to the triangle mount type. Can I use the ones I have (triangle) or should I just go get some inline ones? Is it possible to add another column between 14 and 15 to accommodate them or am I better off just getting the correct part? As always, thanks.

    ReplyDelete
    Replies
    1. It's no problem, and yes using those would be fine. I bought a batch of these inline trimpots purely because I knew they would come in handy sometimes with vero layouts because they can fit in a single column. But there's no problem in using the offset taper type providing you have the room for the additional column(s) that may be needed for it. Just make sure you add an appropriate number of columns to accomodate the pitch between the wiper and outer lugs of whatever you use.

      Delete
  9. Heh. Just thought of it. I'm going to build this using MP38As!
    +m

    ReplyDelete
    Replies
    1. It's pretty sick, but when compared to my AC128 FF (hfe's ~95 and ~115), four MP38As really do sound better. I guess i should try it with the classics like 2N2222 and BC109 too...

      The subtle uncontrollable gating in the style of muffs is a bit more noticeable than in the PNP germ.
      +m

      Delete
  10. fwiw, i use a hybrid fuzzface live EVERY NITE. the most-used pedal in my collection of a hundred or so.

    the thought of living without it makes me break out in a cold sweat..i use it when i need filth, i use it when i need glassy clean...and all points in between.

    there's nothing else on the planet that comes close to a good fuzzface!!

    ReplyDelete
  11. Thanks, this layout reignited my interest in the Fuzz Face circuit. After breadboarding it, I ended up doing my own layout (https://dl.dropbox.com/u/11996927/Piggyback_Fuzz_face.pdf), which is an identical circuit to this (bar a protection diode and larger filter cap) - no offence to your layout Mark, I just wanted the gain pots on the edge of the board. It sounds really good and if you move the gain pots off-board, it is almost infinitely adjustable (probably too much so actually). I needed to use a 50k bias pot in order to get down to 4.5 V on the Q2 collector, but this may be because I used 100k gain pots and a 10k fuzz pot - which is worth trying. Anyway, it's a nice circuit and definitely gives the Ge Fuzz Face's a run for their money.

    ReplyDelete
    Replies
    1. which brings me to ask, what tranny am i checking voltage on for biasing? seems like i get a close reading on collector of Q3 when i adjust my trim pots... but Sam says Q2, just like a normal FF. the tranny pairings confuse me as to how this works. aside from that this thing sounds killer. been using 2N3904's... any other types you guys are using out there that sound great?

      Delete
    2. Q1 and Q2 are a pair, Q3 and Q4 are the other pair. You're checking voltage on the collector of Q3 which you can control with the Bias pot.

      Delete
  12. Yeah sorry for the confusion. You want to bias the second transistor pair, so that would be Q3 in Mark's layout. I like 2N3904s in this circuit but if you want a bit of a heavier sound, higher gain transistors like BC550Cs are pretty good. That said, I think you get much more variation in sound by varying the gain of the transistor pairs (with the trimpots - you may want to use larger ones though) than by switching transistors.

    ReplyDelete
  13. This comment has been removed by the author.

    ReplyDelete
  14. Built this one but can't get it to sound very good. I can get it to do a dark overdrive but not fuzzy. When I turn the bias and trim pots to extremes it will get real sputtery. Tried some different transistors and component values and double checked connections and cuts but with so fuzzy success. Any thoughts?

    ReplyDelete
    Replies
    1. I haven't built this and so don't have a working version to compare it to, but post the transistor pin voltages and see if that gives any clues.

      Delete
  15. I can get the collector on Q3 to 4.5V with the bias control. It just doesn't have much gain and hardly any high frequencies.

    ReplyDelete
  16. Would this work for pnp? Like to use a couple of 3906 instead of an ac128?

    ReplyDelete
    Replies
    1. You could use PNP transistors but there's not really a lot of point. The reason why people opt for PNP germanium fuzz is because the transistors are much more consistent than the NPN equivalents. If you're using silicon emulating germanium, then why use PNP silicon transistors that will need you to have a positive ground when there is no benefit in doing so when the silicon NPN transistors are just as consistent and easy/cheap to buy in quantity.

      Delete
  17. Thank you for responding so quickly! My reasoning is just that i have a bunch of silicones and i can't find any ac128s for a decent price. 5 dollars a piece isn't my cup of tea. I wanted to try it in the crimson drive. Also in something like the throbak overdrive which calls for germanium without the emitter connected, would there be any tricks along these lines to help? Btw i love this site, thanks for everything you've done to feed this horrid habit!

    ReplyDelete
  18. Could a 22uF cap take the place of the 20uF? What would be the expected change in tone as the value of that cap is increased or decreased?

    ReplyDelete
  19. What is the bias supose to be on each transistor???.So far i only get Q1 to bias the other three do nothing

    ReplyDelete
  20. Built it (sort of) I started with this schematic http://i1270.photobucket.com/albums/jj607/induction1/javelina_schematic_v1_zps87c58074.png changed the output call to 47nf kept the resistors between emitors at 82k and 220k then changed the caps between based and collector to 47pf as per this layout. Finally I used an MPQ3904 chip (this is for 2n3904 transistors in one, so my layout is completely different) I believe the chip should be quieter and have more consistent HFE values and will be great for baffeling your Germanium purist friends. :-)

    ReplyDelete
  21. I’ve been meaning to comment on this one for a while. I built it with 4 random 2N3904’s I had in my parts drawer and it is the best behaved Fuzz Face I’ve built to date. I set the external Bias control to the centre of its sweep, used the trimmers to bias the Q1/Q3 pair to 1.8 volts and the Q2/Q4 pair to 4.5 volts. The pedal is utterly silent as in zero interference and no oscillation when Fuzz is maxed. It’s as though it has a noise gate and the Fuzz is good. Nice job guys!

    ReplyDelete
  22. Hey,
    Heres my build : https://imgur.com/a/m3KjspN
    It is a pretty nice FF, still have to take the time to bias it better though :)

    ReplyDelete
  23. Any circuit diagram for this Fuzz? It should be useful to debug in my case...

    ReplyDelete
  24. Just as a side note of possible interest: Skreddy makes a fuzz using piggyback bc109c's. I was wondering why there was no commercial realisation considering how long ago the effect of piggybacking was discussed online, the I found the Skreddy BC109C.BTW Skreddy comments that the main benefit is low noise, which surprised me.

    ReplyDelete