Simple biasing questions

[gjcamann]

So i've been looking at some simple BJT and FET booster circuits. I see two ways that biasing is done One has the straight forward voltage divider at the base/gate (like from an LPB1). The other takes a resistor from the collector to base or source to gate (like the COT50). I'll try to draw it in ASCII art below:

Voltage Divider

   |        |
   >        >
   >        >
   |        |------ out
   |        |
in-+-------|<
   |        |
   >        >
   >        >
   |        |

Fancy Bias

            |
            >
            >
   ---WW----+------ out
   |        |
in-+-------|<
            |
            >
            >
            |

I get the concept that they are both voltage dividers in a sense creating a bias the the base/gate.

Is there a name for the second way?
Does anyone have any additional insight on what's the advantage of one over the other?
Any resources on the web about this?
How do I calculate the bias using the second method? (It's straight forward with the first one).
Are they interchangeable?
How would I calculate the resistors values for switching between the two?

Thanks all.

[duck_arse]

before the experts arrive, I'll put in my bit.

as far as I know, the fancy version is called "collector bias". the book I learnt that from says it is less consistent than the voltage divider bias, because the transistor Hfe plays a part in the final base voltage, whereas you set the voltage div where you want it. colector bias means one less resistor, and so is cheaper.

interchangeable, I don't know, but there is a fancier version of the fancy, with a resistor to ground from the base.

I don't know why you would want to switch between the two.

[Johan]

In your"fancy" bias, the bias resistor is also a feedback resistor and combined with the output impedance of whatever comes before it plays a role in the total gain in your transistor stage...things get complicated if your the type who need to know the exact numbers...I practice plug and pray:P

[R.G.]

The name for the second "fancy" way is either "collector feedback biasing" or "voltage feedback biasing", depending on who you talked to back in the transistor dark ages.

The feedback nature of the biasing for the second way has a profound effect on the operation, as noted above. For one, it lowers the input impedance dramatically. It also makes the transistor better able to make use of all of the voltage swing the power supply has available. It has a different DC bias stability characteristic than the four-resistor network, which is universally referred to as the "stabilized bias" circuit. Or was, back when this mattered.

In my mind, the really fancy circuit is the stabilized bias (i.e. voltage divider) circuit but with an added high value resistor between the transistor base and the voltage divider, and with the voltage divider point bypassed to ground by a capacitor. This one went by the name of the "noiseless bias" circuit, although it's not truly noiseless. It does allow you a higher input impedance than the stabilized bias circuit while keeping lower noise than merely using very high resistors in the voltage divider, and in general being a good low noise way to bias things. It's still used in the way we bias opamps, with a Vbias divider and a series resistor to the + input.

[merlinb]

as far as I know, the fancy version is called "collector bias". the book I learnt that from says it is less consistent than the voltage divider bias,

Feedback bias is more consistent, rater than less, since it is inherently self compensating.

[gjcamann]

Oh man, these are great answers. Thanks so much guys. I was following until merlinb jackknifed everything I had laid out in my brain  but I guess I could see his point - need to learn more.

I can imagine the interesting effects the feedback resistor could have. And the noiseless bias scheeme is a cool one.

Is there any guidance on when to use which scheme?
Assuming high input impedance is usually good, would you deliberately lower the input impedance for a reason? 

OK, i'm off to Wikipedia to learn more about "Collector Feedback Biasing"!!!

[R.G.]

Feedback bias is more consistent, rater than less, since it is inherently self compensating.

That depends on the circuit setup. I had to derive the sensitivity factors for bias stability for both the voltage feedback and stabilized bias circuits in my "All About Transistors 101" class.  The relative stability of the two depends on how much of the voltage and current are apportioned to the collector and emitter resistors and in the case of the voltage feedback circuit, the feedback resistor.

I would have to dig out my old textbooks to regen the info on this - it's been about 40 years now - but the bottom line on this is that either can have bias stability (or not) to any necessary degree, if you can live with what the needed stability changes do to the other characteristics of the circuit. In general, about a 10:1 ratio of collector to emitter resistances nails bias stability whatever the base bias setup is doing, as the change in emitter voltage is then generally large enough with drift to minimize the drift.

That's only really a problem where you really need the full swing of the power supply on the output, and giving away 10% of it to an emitter resistor can't be tolerated. When you get into that, you need to lessen the emitter resistor to much less than 10% of the collector (and hence, 10% of the voltage drop across it) and look to the base side for stability. That's nearly always in the form of both AC and DC feedback in a collector-to-base network, the DC part setting the DC stability and the AC part either adding or removing the inherent AC feedback; you have to live with or modify to suit the lessening of the input impedance that voltage feedback gives you as well.

[duck_arse]

depending on who you talked to back in the transistor dark ages

my "that book" is "fundamentals of solid state, by jamieson rowe, 1974". he was the editor of electronics australia, and to prove I haven't been paying attention, he refers to the fancy bias as "self bias", and notes splitting the bias resistor and ac grounding the centre will overcome some of the bias wobble that the signal causes.

I nearly always go for the noiseless bias when I don't know what I'm doing. the bias voltage I think I want, not what the circuit thinks it needs, at least till it reaches the other end of the base resistor.

[greaser_au]

depending on who you talked to back in the transistor dark ages

my "that book" is "fundamentals of solid state, by jamieson rowe, 1974". he was the editor of electronics australia, and to prove I haven't been paying attention, he refers to the fancy bias as "self bias", and notes splitting the bias resistor and ac grounding the centre will overcome some of the bias wobble that the signal causes.

I nearly always go for the noiseless bias when I don't know what I'm doing. the bias voltage I think I want, not what the circuit thinks it needs, at least till it reaches the other end of the base resistor.

Jim Rowe...    Everything from valve amps. through solidstate to hobby computers. Truly a guru...

david

[R.G.]

I'll have to go read Rowe.

One additional note: the noiseless bias circuit becomes a higher input impedance bootstrapped circuit when you feed some of the emitter signal back to the voltage divider end of it with a capacitor. This can be done right at the voltage divider if you leave off the filter cap to ground from the divider, or by splitting the series resistor to the base from the voltage divider (or other Vbias source voltage) and feeding the bootstrapping signal to the middle of this split base resistor.

[Electron Tornado]

my "that book" is "fundamentals of solid state, by jamieson rowe, 1974".

I couldn't find that book anywhere. Here are three that I do have. The first two were recommended to me.

Handbook of Simplified Solid-State Circuit Design, by John D. Lenk

Practical Transistor Circuit Design and Analysis, by Gerald E. Williams

Transistor Circuit Design With Experiments, by Delton T. Horn

Anyone else use any of these?