About BIAS

[sonolink]

I apologize if this is the wrong section to ask about this.
Since I have been working in studios, I always thought BIAS was a high amplitude high freq signal used in tape recorders to "stimulate" tape domains. I also knew it related to tube amps calibration somehow but never went too much further into BIAS... Recently I've been reading AllAboutCircuits.Com and BIAS is mentioned several times relating to transistors, amps, and all sorts of components. I tried Googling and reading some definitions but they all seem too generic or just create a sensation of vertigo in my stomach, not allowing me to grasp exactly what BIAS is or just how the concept works. So, in my present state of confusion (or might I say dizziness) I would be very thankful if someone is kind enough to try to explain this concept to me.
Thanks in advance

 

[abbey road d enfer]

Bias is another word for "operating point".
It may be the Vg/Va/Ia parameter set of a tube, it may be the Ic/Vce of a transistor, it may also be the DC voltage applied to a condenser microphone diaphragm, or the ultrasonic current superimposed to the audio current in a record head.
The bias in a recorder is set as a compromise between THD and HF frequency response, bias in a tube maybe a compromise between output level and THD or noise vs. THD depending on the application.

 

[JohnRoberts]

"Bias" can mean different things in different contexts.

In tape recording it is commonly a HF sinewave superimposed on the audio to get the magnetic domains moving, but with no net effect from the bias, so the audio sensitivity is better. FWIW in the very early days of magnetic recording I think a DC bias may have been used.

For most other examples bias is a DC term added to or subtracted from the changing signal to manage (Abby's) operating point for best headroom or other performance metric.

Sometimes the bias is not something intentionally added but a consequence of the circuit behavior. For example a LTP (differential input pair) of bipolar transistors will have a DC base current flowing into or out of both input leads of the differential amplifier (more of a problem with old bipolar opamps, less so in new designs). This is commonly known as input bias current.

JR

 

[tv]

Welcome to BIAS!

http://www.bias-inc.com/

 

[sonolink]

Thanks for all your prompt answers!

So Bias is related to powering transistors and tubes, right? It does not relate to audio signal does it?

 

[pstamler]

Well, it can; the audio signal coming off the plate of an RC-coupled tube has a DC component. Instead of varying above and below zero volts, it varies above and below, say, 180V. You could call that a DC bias. Of course, that DC component gets filtered out by the coupling capacitor.

Peace,
Paul

 

[sahib]

Thanks for all your prompt answers!

So Bias is related to powering transistors and tubes, right? It does not relate to audio signal does it?

Well, not really the correct terminology. Unfortunately you have to know a bit more on the semiconductor principals to be able to understand it fully. But within the context of a biasing  transistor and/or valve, it means to place them at a specific operating point.

I'll give an example of a bipolar junction transistor (bjt) which is easier for you to understand. And let's have an npn one. You connect the collector to +V and the emitter to ground. But nothing is connected to base. You already powered the transistor but it is not conducting. For it to conduct you have to apply current to the base. This current could come from your audio signal.  In other words the signal current biases the transistor and causes it to conduct. But at this point you have to know of a particular characteristic of the transistor. For it to start conducting the base voltage has to be above a certain level, and this is roughly 0.7V for the silicon transistor. So, if your audio signal is below 0.7V then the transistor is not going to be biased, hence no conduction will occur. When the audio signal goes above 0.7V then the transistor will be biased and it will start conducting. But as soon as the signal goes below 0.7V it will stop conducting. What this also means is that the first 0.7V of your audio signal will be wasted because it will never be amplified, and the amplified portion is going to be only what is above 0.7V but highly distorted.

So, to overcome this problem you look into biasing the transistor in another means. And that is applying a continuous DC to the base. Easiest is obviously to derive it from +V supply. Now the transistor is biased by DC and conducting.

Now, when the auido signal arrives at the base of the transistor the path is already paved for it to walk through. No more 0.7V minimum limit. So it rides at the back of DC and appears at the output of the transistor. However, as mentioned in the previous post it now has a DC content.  To seperate it from that DC you place a capacitor in the signal path. Now the DC is blocked on one side of the capacitor and the AC signal appears on the other side.

 

[PRR]

> thought BIAS was a high amplitude high freq signal used in tape recorders to "stimulate" tape domains

That's actually bad usage (but that IS what we say).

Bias is your particular slant. The "fair and unbiased news network" slants to conservative opinions; Woody Guthrie slanted to liberal opinions. You can make clothing with the threads running up/down left/right, when you make clothing with threads at 45 degree you say "on the bias". An experimenter who discards "bad" data has biased his results.

http://en.wiktionary.org/wiki/bias

If you just poke at a tube/transistor, mostly it is either ON or OFF. And in most radio and audio, this is useless. You carefully "bias" the device "part-way" between ON and OFF, so that small variations give a useful change.

Small-signal audio stages are often biased so the device output pin (plate, collector, etc) is "about" half-way between the supply rails. A Fender preamp has zeroV and +300V rails, and the 1.5K and 100K resistors around the tube set the plate to 200V.

LARGE-signal audio stages at FULL power will suck whatever current they want to drive the load. However we must pick a "bias" for them to set at when idle or when passing very small signals. If we decide to set half-way between peak min and peak max load current, that's Class A. SE amps must work this way. With push-pull, we can do the same thing (which is simpler) or we can bias for the lowest current which still allows small signals to pass nicely.

Except in politics and dress-making, we usually want to remove bias before we use the result; hence output coupling caps or transformers.

Tubes only go one way. Magnetics (tape) can go both ways, on+ and on- as well as zero. However the curve through zero has a severe kink. The *simplest* way to "fix" this is to add a little DC to the audio signal into the record head, to "bias" the no-signal point about halfway between zero and saturation. Most wire-recorders, and a few super-cheap tape recorders, did it this way. Since wire and steel-tape recorders existed before the Germans came up with the supersonic trick, we say "bias". However the now-standard supersonic oscillator is not a "bias" in any of the above sense. It smears audio all up and down the S-kinked curve of magnetism, averaging-out the kink and using the full range from + to -. There ought to be a better word, but we say "bias".

 

[sonolink]

Thank you, that's the kind of explanation I was looking for, especially sahib's

 

[barclaycon]

Bias was always an important consideration when lining up the multitrack for a recording session.
As has explained been above, it is a high frequency oscillation to assist the transfer of audio onto the tape. Bias is different for each type of tape used and it can even be different for different batches of the same make.
Generally the rule for adjusting bias was 3dB over at 10k.
You send a 10k sinewave from the desk - usually at a lower level than normal so it's away from any tape compression, then observing the output meter on the tape machine you back off the bias adjustment and then slowly increase it until the signal 'peaks'. Noting this level, you continue to increase the bias until the level drops 3dB from the peak.
There's a tradeoff in bias adjustment between HF performance (too much) and the susceptability to dropouts (too little). The 3dB at 10k adjustment was always a good balance.
However, during the 70's tape manufacturers started suggesting less bias (1db over at 10k) claiming that this would give a cleaner top end. Unfortunately, this did result in an increase in the phenomenon known as 'bias bubble' which resulted in clicks and pops with bass frequencies - especially bass guitar.
I'm currently restoring a bass guitar track at the moment which has just this problem.
I remember some engineers in the early 80's tweaking the bias on the multitrack with the bass guitarist playing and listening 'off-tape' to get minimum bias bubble.
For my money, too much rather than too little bias was always better.

 

[tv]

A cool story!

 

[sonolink]

Definitely...back in the tape days we used to struggle to loose as less top as possible. Our concern was to preserve the highs, whereas nowadays it's the other way around, especially with the loudness madness!!