"Piggybacking" trannies for lower gain

[Jay Doyle]

It also follows that if you used some small unbypassed emitter resistors (seperate one on each transistor), the beta would approach the average beta of the transistors (as it does in a power amp).

I definitely think that you are on to something here, inserting equal value resistors on both emitters would would overwhelm the difference in the Shockley resistance and equalize the b-e junction voltage. Isn't this similar to the equal value resistors normally used on the emitters in differential amplifiers?

This is a great discussion! Though at some point, someone is going to have to draw a picture   .

Jay

[WGTP]

Just to show how stupid I am, how about mixing a 2N3904 with a 2N7000?  

Glad I'm finishing up one, it looks like I'll be going back to the Tone Bender.  

[Jay Doyle]

Just to show how stupid I am, how about mixing a 2N3904 with a 2N7000?

You could probably figure out a way to do something that works with these two but it won't work like we are talking about here. A MOSFET doesn't draw any current through its gate, so if you attached it like we are talking about here it wouldn't have any effect at all, it would be like adding a teraohm resistor from the base to the emitter of the 2N3904, nothing would change.

If I understand the discussion here correctly, we are talking about using a second transistor whose base is connected to the first transistors base, and whose emitter is connected to the first transistors emitter, but whose collector is left floating. Then this two transistor combo is used as a single transistor in a circuit.

[Brett Clark]

Aron - what can you tell us about this "Incredible Stompbox Index"? What goes here vs. elsewhere?

BTW, thanks as always for this wonderful forum...

[aron]

Guys,

This thread is VERY COOL!

I've made a new forum - the iSi (Incredible Stompbox Index) that will house interesting discussions.

This one is the first since it's pretty interesting. It doesn't really affect the thread although it says "moved".

If there are other very interesting discussions you guys want in there, let me know. These would be live discussion, not FAQ type ones.

Thanks!!!!!!!

[Nasse]

there must be some magig in emitter resitors, especially in differential amps. Transient intermodulation distortion was big debate some years ago, when one Finnish bloke discovered it (M. Otala, he was in the team that designed Citation CXX power amp (costed more than a Mercedes Benz car)). After that some finnish electronic mags suggested tips how to make your transistor amps sound better.

By adding those emitter resistors

But that was slightly off topic.

[WGTP]

Thanks Jay, I knew that.     I just plug the 3904's and 7000's into the socket the same direction and it works, so I thought...  

Guess I will try a 3904 and 5088.  

[javacody]

RG, have you had a chance to test this out yet?

I cannot believe how much better your original suggestion sounds. Now I have to try the resistor thing. I'm not able to picture in my mind though how I connect the resistor?

Does the resistor go between the two emitters?

[Jay Doyle]

Does the resistor go between the two emitters?

In my mind the bases would be connected together, each emitter would have it's own resistor and the other ends of the resistors would be connected together forming the "artificial emitter", and one of the collectors would be left floating.

Remember in the end you have a circuit that will have three terminals just like a BJT and be used as such.

[Nasse]

:mrgreen: I just discovered few cool names for pedals that can made if this works :mrgreen:

I think you can name a fuzzface "Miss Piggy" and paint it pink

If I make a Rangemaster I think it can be R.G. Master

:oops: I need a holiday

[Elektrojänis]

I think you can name a fuzzface "Miss Piggy" and paint it pink

You could even buy one of those rubber pignose things from a store that sells masks and halloween costumes etc. and glue it on your pedal.

[brett]

Hi.  Concerning javacody's question about connecting the resistor: Yes, I simply connected it between the two emitters.  Values of between 1k and 10k gave interesting results (1k=lower gain, more piggyback, 10k = higher gain, less piggyback).

Here's a layout for a "Miss Piggy" (thanks Nasse for the cute name), a  Fuzzface with the usual trimmers plus piggybacked trannies for Q1 and Q2:


Also, Javacody wrote:

I cannot believe how much better your original suggestion sounds. Now I have to try the resistor thing.

Thanks for the compliment.  It's good to know that someone else has come to the same conclusion (ie that it sounds very good).

Last night I built a Fuzzface (as per the Miss Piggy, above), and piggybacked (no resistor) 2 x BC547s for Q1,Q2 (synthetic hFE=10) and 2 x BC549Bs for Q3,4 (with no resistor, hFE = 4, with 1k hFE = 7 and with 3.3k hFE = 17).  With zero resistance between Q3 and Q4 the gain was too low, and the fuzz was very soft, but with either 1k or 3.3k between Q3 and Q4 it sounded excellent.  I definately think that a low-gain Si fuzzface is going to find a place on guitarists' pedalboards.

Lastly, I also tried Gus' suggestion of using a 1N4148 diode in place of the piggybacked b-e junction.  In the cases where I tried it, the diode sucked all of the current and the hFE was 0.  Like RG, I figure that having a similar junction in parallel probably increases the chances of the two junctions sharing current in a useful manner.

[troubledtom]

as soon as i saw the dave barber op amp stacking the same day i was was doing it w/ trannies. i used them in the double bazz fuss and got very pleasing results. my bass plaer loves it. i tried so many combo's that the possibilites were endless. i don't if all this stuff is sweat and voodoo. but i had fun.
            - tom

[Gus]

Very cool Brett!!!

Thanks for trying the diode.  

How is the noise with this type setup?

Gus

[javacody]

What noise?  

By the way, the Miss Piggy is exactly the same setup I used, minus the emitter resistors.  I'm very happy with my "compound" trannies gain of 8 and 10 respectively. Sounds amazingly organic.

[brett]

Concerning noise; I haven't really turned the amp up yet, but there's a couple of things noticeable:
I suspect that noise will be as low or lower than a standard fuzz face, maybe not because the circuit is less noisey, but because with lower gain, there's less amplification of pickup noise, etc.
Second, and probably most importantly, if you count harsh harmonics as noise, then I can definately say that there's a difference (I think better) with the piggybacked transistors.  For example, my test signal at 1kHz gives some fairly unpleasant high-pitched harmonics if using 2 regular high-gain Si trannies in a fuzzface.  This disappears with the piggybacked arrangement.  Instead of almost vertical slopes on the trace, I get a recognisable slope.  With hFEs of 10 and 17 I didn't notice any harshness in the sound until the fuzz control was nearly at maximum.  I'm starting to think that much of the magic associated with Germanium might actually be due to low gain.  There's definately a bit of a "clean Germanium (?)" sound to Miss Piggy.

[javacody]

I will second what Brett said. Brett, what is the gain on the trannies you used? The BC109's I used have an average gain of ~490 to 500. There must be a forumula for this? A way to predict what the compound hfe will be?

[brett]

I wish there was a formula.  But these things seem to vary quite a bit with the type of transistor. I've had 2 BC547s (hFE=392 piggied on a 491) give hFE=9, while 2 BC549Cs of 430 gave 4.  I did get a hFE of 20 out of 2 trannies with hFE=410 and 480, but I forgot to write down what type they were.  It's interesting that you were very happy with such low hFEs.  It has reinforced my view that quite low hFEs are not such a bad thing. e.g. I liked 9 and 7 in my Miss Piggy more or less as much as 9 and 17.  

I'm off to the shop this afternoon to buy about 5 transistors of 5 different types.  If there's a pattern to piggied hFEs, I'd love to work it out. (I'm a scientist by occupation AND habit).

[R.G.]

There must be a forumula for this? A way to predict what the compound hfe will be

That's what I'm digging through my thirty-years-old semiconductor physics texts for - a way to predict results.  I'm sure there is a way to do this, but what we need is a way to do it that doesn't involve knowing the n-type and p-type doping diffusion constants ..... aaargh!!

Fortunately, we seem to have a good empirical result - stick a 10K trimmer pot set up as a variable resistor between the two emitters; now tune till you hear germanium. 8-)

I suspect that noise will be as low or lower than a standard fuzz face, maybe not because the circuit is less noisey, but because with lower gain, there's less amplification of pickup noise, etc.

I think there is an even better advantage on noise. The noise generated in a bipolar base-emitter depends on the temperature (same for each), the current density (we're mucking with that - probably dropping it by half) and the net resistance. We're deliberately cutting the resistance quite a bit. There are moving-coil hifi pickups that use paralleled NPN transistors for getting very low effective Rbe. I think something like that may be working here. In any case, we're using glass passivated modern transistors which are much cleaner, and sound it.

Talk about the best of both worlds! Germanium sound and modern transistor availability and low noise.

[b_rogers]

what about a way to effect the amount of "piggyback" or remove the piggyback completely from the circuit. maybe a trimpot? FWIW i also noticed 2 2n4401's with all leads piggybacked in q1 sounds pretty good at least in the multiface.