Gain and distortion questions in BTJs

[midwayfair]

I feel like this is the sort of thing I should have a handle on, but I read something the other day that made me question a few things: That biasing a transistor to 4.5V creates the most headroom. This statement was made by some knowledgeable people and repeated and discussed as fact by even more knowledgeable ones on DIYSB.

But in my practical experience, this hasn't been true: If I bias a transistor to 7.5 or 8v on the collector, despite being ridiculously close to the power rails, it's clean, whereas the same transistor biased down to 4.5V on the collector sounds distorted. And I'm talking about a single transistor stage with no clipping otherwise. Similarly, a Fuzz Face is cleaner with Q2 biased to 6v than biased to 4.5V. Unless AC gain adds a lot of distortion regardless of the DC bias? Or does the added output feed back into the base and hits the GROUND rail since that's always biased a lot lower? EDIT: A Rangemaster biased to a lower voltage on the collector will also distort more easily than one biased higher. In fact, the whole point of biasing it to 7v is to make sure it distorts properly.

What's going on here? I'm so confuzed. :/

[culturejam]

I guess the theory is that with a mono-polar supply of +9v, biasing at +4.5v gives maximum positive/negative voltage swing (but this may be with respect to a given frequency, I'm not really sure).

Without being sure, this seems really similar to VREF with op amps. As in, +4.5 sets the "mid point".

Maybe with guitar frequencies, setting the bias higher makes for cleaner amplification?

But I may be off base here. Op amp biasing seems more logical to me than BJT biasing.

[RobA]

Were they talking about biasing an emitter follower and referring to the base voltage? 'Cause otherwise, the chapter in my electronics book on biasing transistor stages is a lot more complicated than that.

[midwayfair]

Were they talking about biasing an emitter follower and referring to the base voltage? 'Cause otherwise, the chapter in my electronics book on biasing transistor stages is a lot more complicated than that.

I can't tell you for certain, because DIYSB is down today and I can't check. I'm prettysure they were talking about collector followers though ... there's no collector resistance to 9v on an emitter follower.

I realize it's a heck of a lot more complicated, and I'm not helping matters by not providing an example schematic. The easiest one to grab is the BJT stage in my clipper ship: https://docs.google.com/document/d/1dPlB8Lqn_a1A6Xpb3Lg-2cfTxGCLOkMfNUWALyGE3ZQ/edit (schematic page 3)

This is sort of my go-to for a BJT stage with stable biasing. If I use fixed resistors in place of the pot, it biases to ~7.5V with a 2N3904, and it's clean until the emitter resistance drops below about 680R. If I increase the collector resistor in the same circuit to 10K, it drops to about 7V and it's louder and distorts quicker. If I increase it to ~68K (I think ... might have been 82K), it's about half supply and distorted all the time. Also much louder. If I ground the emitter instead of using AC bypass, the same thing happens, and I need a small collector resistor to avoid distortion.

So it seems like something must be happening that involves the voltage at more than just the collector to determine the headroom.

[RobA]

I'm still confused about what they were referring to, but assuming it's a common-emitter stage, there are multiple approaches to biasing. My text's overview section lists at least three. But, for maximum headroom, the goal is to bias the thing so the quiescent state is halfway between cutoff and saturation. There are a bunch of parameters to consider in setting the bias given a particular topology, including the transistor characteristics and the load. Still, I don't think that biasing to 4.5V will get you there.

Giant caveat: I don't sit around biasing transistor stages for fun and/or profit. Every time I do end up needing to bias a transistor, I have to go and get my textbooks out to remember how. In fact, if transistors didn't sound so good, I'd stop using them and just go to op amps for everything.