Question regarding how transistors are biased, ie clean boost vs fuzz

[nightendday]

Seems like a noob like question, and it is. 
but basically what I'm asking is, what changes what and how can I learn to bias a transistor to be clean, for say a clean boost, or output stage. Or, how could I bias one for fuzz, or overdrive? I've learnt basic things from building my own circuits, and see the differences in the biasing, but my question is basically what changes what?

[darron]

i can't give you the best answer like many others here.

it will depend on the topology. a fuzz face is so simple that every single resistor will effect the biasing. the 8k2 is the favourite to adjust though.

which topology are you talking about? which circuit are you modifying?

if you've played around then you'll know that missbiasing can cause clipping/break-up/fuzz. making it sound pleasant is the trick.

[nightendday]

i can't give you the best answer like many others here.

it will depend on the topology. a fuzz face is so simple that every single resistor will effect the biasing. the 8k2 is the favourite to adjust though.

which topology are you talking about? which circuit are you modifying?

if you've played around then you'll know that missbiasing can cause clipping/break-up/fuzz. making it sound pleasant is the trick.

Two good examples are something like an LPB1 and something like a muff i suppose! I know that the resistor coming off the emitter will effect the gain levels, but i mean will limiting the voltage going to the transistor make it fuzz more? is there a magic voltage that all transistors should have on collector?

[WaveshapeIllusions]

I know RG has a good series of posts somewhere on biasing, but I don't recall exactly where

I know a little about it. The gain of a common emitter is generally the collector resistor divided by the emitter resistor. If you ground the emitter or bypass it with a cap, the gain goes up quite a bit. The emitter has a small resistance in itself, but it varies based on current. So a resistor makes things more stable.

The emitter is generally a diode drop below the base (above for PNP) so usually you set up a bias voltage on the base to match with the voltage on the emitter. Messing with the base bias can lead to a more gated, splatty sound. In a lot of the stages I've seen, the bias string for the base is usually a similar ratio to the collector and emitter resistors. For example, say you have 33k on the collector and 10k on the emitter. On the base you'd probably want 100k to ground and 300k to V+.

I think that's about right. Hopefully it helps out a bit.

[darron]

Lpb1 is a good start. Here's a schematic from Breavis audio. Hope he doesn't mind:



The 1m and 100k form a voltage diver, which should put near 1vdc on the base. The 390r on the emitter should set gain. Getting closer to zero should increase gain. Getting the most gain out of a transistor I guess would be a good step to getting more fuzz, but this circuit is already pushing the transistor.

You're right. Adjusting the 10k on the collector will also effect bias. Increasing it might make the output hotter and make it fuzzier. Try a 47k, 100k, see how far you can take it before the circuit really stops working.

Typing on my phone. Edited spilling mistake.

[nightendday]

Lpb1 is a good start. Here's a schematic from Breavis audio. Hope he doesn't mind:



The 1m and 100k form a voltage diver, which should put near 1vdc on the base. The 390r on the emitter should set gain. Getting closer to zero should increase gain. Getting the most gain out of a transistor I guess would be a good step to getting more fuzz, but this circuit is already pushing the transistor.

You're right. Adjusting the 10k on the collector will also effect bias. Increasing it might make the output hotter and make it fuzzier. Try a 47k, 100k, see how far you can take it before the circuit really stops working.

Typing on my phone. Edited spilling mistake.

Big help! and that makes sense! do you mean increasing in the voltage to collector will increase output or that increasing the resistance, and thus lowering the voltage will?

[nocentelli]

OP - The difference in cleanliness or fuzzyness between the examples you've given (LPB- and BMP) is much less to do with the bias of the transistors, and much more to do with the topology of the circuit: The LPB-1 is a single transitor stage, about as simple as it gets to boost the signal. The big muff is four stages, the first is a single transistor booster, much the LPB-1. However, this stage is followed by two transistors with clipping diodes that actually produce the fuzz. The fourth stage is similar to the first, and boosts the level again, after the clipping has distorted the waveshape, but reduced the overall peak-to-peak amplitude. There are ways to get dirt out of a single transistor stage, but it usually requires a pair of clipping diodes to ground after the boost, or overloading the transistor input so it runs out of headroom and starts "hitting the rails", or by deliberately misbiasing, as described above, which also tends to gate the signal.

[Electron Tornado]

These will help:

http://www.beavisaudio.com/techpages/HIW/HIW.png

http://www.geofex.com/article_folders/fuzzface/fffram.htm

Another good source is "Handbook of Simplified Solid-State Circuit Design" by John D. Lenk

First, concentrate on understanding how a clean booster works, then move on to the fuzz. They are both essentially amplifier circuits, and I think of a fuzz face as a mis-biased amplifier.

[R.G.]

The output of a single NPN (for the simplest example) stage can hit the positive power supply by the transistor turning completely off and the collector resistor pulling the output up to the positive power supply. Likewise, the other extreme is the transistor being effectively a short circuit, and pulling the collector and emitter resistors together like a switch.

That's all the output swing that's possible with only resistive loads.  If the emitter resistor is much smaller than the collector resistor, as it usually is, we can simplify this to the transistor pulling the output nearly to ground. So the maximum possible swing on the output is the power supply to ground. It's like what height can a tennis ball be at when inside a room? It can be at the floor (0V) or you can throw it up to a maximum height of the celing. That's all there is.

If we want the biggest possible output swing, we know that audio signals generally have about equal positive-going and negative going signal peaks. If they go positive the same amount they go negative, and the only available swing before distortion is the power supply and ground, then for maximum positive and negative-going swings, we have to bias the collector right in the middle of the available power supply voltage. Otherwise, one side of the signal, positive or negative, bangs into the power supply or ground before the opposite side, and we get distortion sooner than we otherwise would have.

The emitter resistor, if used, subtracts a little from this range, because the output can now only go as low as the collector and emitter resistors as a divider will let it. Generally the emitter resistor is less than 1/10 of the collector resistor, so even with an emitter resistor, setting the collector voltage with no signal in the middle of the power supply is a good approximate starting point. A few percent higher than dead middle would be more perfect, but it's generally not that fussy.

So middle of the power supply on the collector is where you want to be for biggest undistorted output signal. How do you get it there?

The trick is that the base and emitter are always tied together by a forward biased diode inside the transistor. The base is never more than about 0.7V higher than the emitter, and is never less than about 0.45V higher than the emitter. If this is not true, the transistor is not amplifying.  To get the collector at any particular DC voltage you want, you pick the voltage, and then compute the current that has to flow through the collector resistor to get the collector voltage to be there. If you have a 9V power supply, and want the collector at 4.5V, and you have picked a 10K collector resistor, then the voltage across the 10K resistor is 4.5V. That means that the current must be, by Ohm's Law:
I = (9V - 4.5V) / 10k = 450uA

Let's say you have a 1K emitter resistor. The same 450uA plus the base current flows in the emitter resistor. The base current is less than 1% of the collector current if the gain of the transistor is over 100. So ... ignore it! Your answer will be more correct than the tolerance on the resistors will let you get!

So the voltage at the emitter will be 450uA times 1K, or 0.45V.

All that remains is to figure out what the base voltage must be to make our assumptions about the collector and emitter come true. We know the base will have to be between 0.45V and 0.7V higher than the emitter. We could guess anywhere in that range and get close. But for high gain ( hfe>100 ) devices at low current, it's more likely to be in the low end of that range, since the base current will be so tiny ( in this case, less than 450uA/100, or 4.5uA). So guess - about 0.5V.

The base voltage must be at 0.45V, plus 0.5V, or 0.95V. All that remains is to pick resistors that set the base to 0.95V when there is no signal and the collector and emitter voltages are as stated.

That's where we start for the case of biggest undistorted signal swing.  But how do we make it bang into either the power supply or lowest possible voltage if we WANT distortion?

Easy. Move the collector voltage nearer either cutoff (that is, the transistor is off and the collector is nearer the power supply with no signal ) or saturation (collector nearer emitter) as you choose. Figure out the current in the collector and emitters from the voltage you want the collector at, then figure out the resistors to put the base where it forces the emitter to be the right place.

[nightendday]

RG to the rescue! That made perfect sense RG, thank you so much!