Biasing for different classes

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The one and only
Joined 2001
Paid Member
OK, the basic formulas are W = V^2 / R and W = I^2 * R
and V = I * R

where V = volts, R = ohms, I = amps and W = watts

For a clean sine wave, the peak wattage is twice the
average wattage.

So 100 watts rms into 8 ohms is 40 volts peak and 5 amps
peak, and 28.32 volts average and 3.54 amps average.

Figure that your amplifier supply will require a few more
volts than the peak output voltage.

For an ordinary push-pull amplifier, the Class A bias is
1/2 that of the peak output current, which for the above
example is 2.5 amps bias.

For other types of amplifiers, the relation between bias
current and Class A output varies according to the design,
but realistically, it's always high.

:cool:
 
demons_wing said:
i was wondering if anyone could tell me the rules of thumb for biasing different classes of amplifiers ie A, AB and B...i would like to know how to work out output power (peak and RMS) and Current (peak and RMS) please

There is a lot of touchiness to the classes as I found out in a post I made a while back.

I'm using the definitions from the Sedra/Smith book:

Class B is essentially a mirrored PNP/NPN EF with no base bias and direct driven from the VAS. Inherently it has no bias current which seperates it from class AB.

Class AB is the same mirrored PNP/NPN with some bias circuitry. The Self book has some notes about how to determine the optimum bias for this topology but I can't remember off hand how it's done. Usually it's between 50ma - 300ma with 300ma being fairly high. The leach amp uses about 100ma.
--
Danny
 
sreten said:
As ever confusion reigns.

Class aB is optimal minimum bias, and the majority of amplifiers.

(note that this is what D.Self would like to call class B)

Class AB is purposefully class A up to a few watts.

(also according to D.Self)

:) sreten.

I refernce Sedra Smith as an authority, sreten references Self for validity, equally qualified authors, equally qualified definitions...

The trouble between distinguishing between aB and AB is defining at what point the transition occurs. If you assume 3 watts, then in the same output topology is a difference class for 150ma bias vs 300ma bias (8-ohm load). It becomes hard to determine from the topology what class the output is but determinable by observing the output.

But then, on the other hand using my definitions, the topology is quite different for each class but hard to determine by observing the output. For example, it is hard to determine exactly at what point does the class B like output instead become class C? At output voltage levels, say about 5Vpp -> 3 watts, conduction only occurs for 75% of the cycle. Clearly a class C response...

I guess you just pick your own definitions from the way you like looking at it and stick to it...
--
Danny
 
Re: Re: Biasing for different classes

azira said:


...Class B is essentially a mirrored PNP/NPN EF with no base bias and direct driven from the VAS. Inherently it has no bias current which seperates it from class AB.

Class AB is the same mirrored PNP/NPN with some bias circuitry. The Self book has some notes about how to determine the optimum bias for this topology but I can't remember off hand how it's done. Usually it's between 50ma - 300ma with 300ma being fairly high. The leach amp uses about 100ma.
--
Danny

Let's back up here. These output stage class definitions were made in the tube era and remain unchanged today. Class A means simply that current always flows through an output device all the time during the output waveform. Class B implies a push pull amp of some sort, and the output devices each are on for either the positive or negative output polarity only; thus with a sine wave without offset, each output device would be on 50% of the time. Class AB is a hybrid, with the devices simultaneously on for significantly greater than 50% of that same sine wave. These definitions are true regardless of the amplification device -- transistor, mosfet, tube or whatever.

Self recommends class B highly over class AB because it has the lowest measured distortion. He considers there is a precisely optimal class B bias for any output BJT stage, whether Emitter Follower EF or Complementary Feedback Pair CF, which yields minimum measured distortion. The optimal bias current is a function of the emitter resistors and output configuration, but for a typical EF output with 0.22 ohm emitter resistors, it is 107ma (presumably at 20C), while for the CF stage it is 11.5ma. [Yes, strictly speaking this is not class B since there is some simultaneous conduction of the outputs -- but clearly, not much, and this is called class B by most authorities.] The optimal way to set bias is using a distortion analyzer for minimum distortion, regardless of the output configuration. Overbiasing will yield higher distortion, so it should be avoided.

Self does not consider mosfets acceptable as output devices and makes no recommendations for bias current (his head is clearly in the sand on that topic). However, most designers bias at or above the knee in the transconductance curve of the mosfet, typically 100-200 ma per device.

There may be unique circuit characteristics which call for different bias currents, but those are reasonable starting points.
 
Re: Re: Re: Biasing for different classes

slowhands said:

Class A means simply that current always flows through an output device all the time during the output waveform. Class B implies a push pull amp of some sort, and the output devices each are on for either the positive or negative output polarity only; thus with a sine wave without offset, each output device would be on 50% of the time. Class AB is a hybrid, with the devices simultaneously on for significantly greater than 50% of that same sine wave. These definitions are true regardless of the amplification device -- transistor, mosfet, tube or whatever.

However, you have not explaind two critical behaviors.
Firstly, with NO bias circuitry, a push-pull BJT configuration has 2*Vbe of dead band. As I showed in my 2nd post, this is a significant portion of the wave (and therefor conduction for much less than 50%) even up to 3-watts output power. This topology does not fall into your definition of a class B ever.
Secondly, doesn't your class AB definition fall under the catagory of class A for power developed for current up to 2*bias? Where do you draw the line? At 300mA bias this is almost 3-watts of power. What if you raise the bias a bit more? Is this not a push-pull but also a class A? So what is it?

slowhands said:

Self recommends class B highly over class AB because it has the lowest measured distortion. He considers there is a precisely optimal class B bias for any output BJT stage, whether Emitter Follower EF or Complementary Feedback Pair CF, which yields minimum measured distortion.

But refering back to sreten's and my discussion, Self's class B is not the same as another authorities class B. He considers class B what you are calling class AB.

slowhands said:

[Yes, strictly speaking this is not class B since there is some simultaneous conduction of the outputs -- but clearly, not much, and this is called class B by most authorities.]

Which authorities? These definitions of class A, AB, B are not what Sedra/Smith defines nor it seems what Self defines.

I agree there is a need to have clear definitions because these are tools we all use for discussion but since there are disparities between the authoritative sources, the only thing to do is to state your definition before your conclusion and let the readers translate to their belief system.
--
Danny
 
The one and only
Joined 2001
Paid Member
"Secondly, doesn't your class AB definition fall under the catagory of class A for power developed for current up to 2*bias? Where do you draw the line? At 300mA bias this is almost 3-watts of power. What if you raise the bias a bit more? Is this not a push-pull but also a class A? So what is it?"

It's Class A up to that point, "Class B" beyond. Taken as a whole,
it's Class AB. If an amp runs Class A up to its rated power into
a particular load, it's conventionally referred to as a Class A
amp, with the understanding that beyond that output current
it's likely "B", if anything at all.

I refer to it as "B" with quotes, as someone here will argue
that it's Class AB. Subtle distinction not worth arguing about
as long as we know what we're talking about.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.