Class A voltage amplifier circuit w/o caps or inductors

[jcx]

Originally Posted by cuallito

Hi, I'm designing a "minimalist" amp and I need some sort of circuit that will amplify voltage for the input stage without nasty coupling capacitors and operates in class A mode. No current will be flowing out of the circuit.

Tried a simple common emitter circuit but seems like you need the input capacitor for it to work properly.

I know I could use a class AB push-pull configuration but darn it I want it pure class A.

all internal stages of op amps run Class A

all op amps have some bias current in the output Q, so for small enough load all op amps can operate push-pull Class A

typically this may be as small as 100 uA but a few are known to run "hotter"- the AD797 and LT1x28 both use 500 uA output Q bias which gives +/- 1 mA Class A

then there is always the option of single ended Class A output bias with a current source - or even just a resistor to a rail
just load the op amp with several times the desired output I

you can bias 2 paralleled output op amps with many mA deliberate offset current running through the "current sharing" resistors

if you go for CFA op amps developed for A/DSL service you can get >100 mA Class A output this way

[Bigun]

Originally Posted by MiiB

The funny this is that there's is no design that excludes capacitors...If you don't use them as coupling in the input and out put (input cap can sometimes be omitted), then the PSU caps are right in the signal path...as they support the GND end of the signal...

I believe you can avoid capacitors if you use a bridge design, where the two ends of the speaker are attached to single ended amplifiers both amplifiers consisting of an active element fed from the same supply rail via constant current sources. In this case there is a constant current draw through the supply at a constant voltage and theoretically you can avoid any a.c. flowing through a capacitor.

Take a look at Nelson's Balanced Zen line stage for something close to this (replace load resistors with CCS). Perhaps it might even be a starting point to solve the problem raised at the start of this thread ?

[MiiB]

Thats right...Balanced (bridged) single Ended class A designs has the option of not having capacitors in the signal path...

[Mooly]

The reservoir caps in a "normal" amp are not in the signal path by my definition

If they were.... why do you not hear the 50/60hz ripple current component flowing through them

[MiiB]

Because it is smoothed due to their size and the PSSR of the amplifier circuit....They are in the signal path as they serve to stabilize GND...I believe that that is a major reason why the PSU caps contribute so much to the performance of the amplifier.

GND stability is vital....It's like bouncing a ball on the floor...it just works better on something solid than on a soft carpet...

[Mooly]

Make sure you read post #20 and post 27#, #38, and #39

Output capacitor: subjective and objective views?

[MiiB]

I would still argue that they are in the path....without them the GND would modulate with the load currents....The current may flow around them through the transformer via the diodes, but the caps keeps the spread and stabilises the GND..

[DF96]

If the PSU capacitors are in the signal path, then at low frequencies the whole national electricity grid is also in the signal path. The problem is that 'signal path' is not always a helpful concept; it seems to lead to misunderstandings.

I'm not sure what is meant by "GND stability". Provided the input, output and feedback are all referred to the same potential (i.e. a single point) it doesn't really matter what that point is doing relative to some other point. Most ground problems are about reference being to different points, when they should be the same point.

[MiiB]

Then why do we need them....???

[DF96]

Why do we need PSU caps? To provide a reasonably smooth DC supply voltage. The cap has to be big enough to reduce ripple to below the level required by the amps PSRR, and also to prevent voltage droop between charging pulses dropping the supply too low to accomodate the required signal voltage.