What is "Bias", what is "Gain" (in electrical engineering terms)?

I hope I get this right:

I mostly understand bias in terms of vacuum tubes - I haven't stuck my nose into solid state yet. The signal comes through as a sine wave, which means that you have a point that's zero and the signal will swing or move positive and negative from the zero point. For a crude example, you can say that the peak positive will be 100V and the peak negative will be -100V.

Now, a tube cannot go negative. You cannot have the voltage on a tube at 0 and have it go down 100V - it isn't possible.

So what you do is set the bias of that tube at, say, 300V, or you can say that you set the zero point at 300V. When this hypothetical tube swings 100V in either direction, the voltage on the tube will vary from 200V to 400V, which is in the tube's operating envelope.

Also within the meaning of bias is keeping the tube with a zero point at 300V. As tubes age, their output changes. Some amps autobias themselves, but others have a potentiometer where you can vary the amount of B+ (the main voltage) going in. So after maybe a year of use, the zero point will change to, say, 295V. You can hook a meter to the amp (some have meters built in or indicator LEDs or other schemes) and turn the potentiometer until the tube is operating at 300V again.

You should also know about tube curves. Maybe you've seen them - a Google search will turn up the curve for any popular tube. You'll find a graph with a bunch of lines on it. This shows you how linear the tube is given the bias voltage. I'm still learning about this and can't offer much detail, but you generally want to pick the straightest curve so the tube will have the most linear response. There are other considerations here, but that should give you an idea of what the designer is doing. Tubes will sound different depending on how you bias them. The same tube biased at 300V in one amp will sound different if biased at 250V in a different amp. There's a range of play here - often there is no perfect voltage, and you might want to vary it for other considerstions. Also, tubes have a maximum voltage. You don't want to set the bias where the maximum swing exceeds what the tube can handle.

Gain. This is the amplification factor, or how much the input signal can be amplified. You could go for twice or nine times the original input, or possibly more. It might seem like you'd want to max it out, but there are drawbacks. If you have 100 times the amplification, then you might overpower your headphones. You can also end up with a very narrow range on the volume control - it'll blow you ears out if you slightly turn it up. You can also run into problems with distortion with high gain. Lower gain tends to be more linear. So you have to balance the amount of gain used. You want some amplification of course, but you also want to keep the gain as low as possible to keep the amp linear. You'll sacrifice volume, so you have to be careful where you set it.

Anyhow, I hope that answers your question. I also hope that if I got something wrong or missed an important point that of of the guys who really knows this stuff drops by.

Actually, they're not ohm meters, they're Wheatstone bridge meters with a DC Op amp driving a nice big meter.  They are used to measure GSR (Galvanic Skin Response).  You apply the probes, null the meter, adjust the gain as needed, and start asking questions, noting the response of the meter correlated with the subject's response to the question or stimulus. 

 

I built one back in the 60's, inspired by an article I read in Argosy magazine on the Baxter effect which was about a plant's (a philodendron) apparent awareness of its environment.  Anyway, it intrigued me so I designed a psych study as a freshman, for a high school class based around verbal stimulus / response and correlation to the subject's background/past.

 

It was great fun and I would probably not have been allowed to do anything like that these days because of invasion of privacy issues.  Over the course of time, I learned to completely control the response of the meter.  I could make the needle go any direction I wanted, regardless of other circumstances.  This was before biofeedback was much understood or written about. Many years later I learned that Scientology used these gizmos to enslave people.  Hocus pocus...glitches be gone!  Fundamentally, nothing more than biofeedback and conditioned response training.

 

Anyway, in the context of the E-meter, gain adjusted the sensitivity of the op-amp and bias zeroed (nulled) the meter

 

The bias balances the bridge to match the subject conditions currently under test and the gain adjusts the sensitivity of the op amp to amplify the changes in galvanic skin resistance.  Yup, you've got it.  So you would typically set gain to minimum, zero the meter, and start increasing the gain while continually re zeroing the meter as gain is increased.  During this calibration process, you ask the subject a few questions to allow you to set gain appropriately for the baseline of the session.

 

Cheers!