OT: quiescent current for class A amp

[brett]

Hi.
I'm building a Class A amp that uses a MOSFET voltage follower for output.

The output stage is very similar to this;

Image from the excellent Elliott Sound Products website.
My question is how closely the current through the current source needs to match the current through the speaker.
That is, what would happen in the design above if R6 were 2.2 ohms and the current is only 1 amp?  Does it reduce the power output proportionally? (P = V*I)  
I'm sure that the answer is obvious. :oops:

[davebungo]

OK, this is just my initial opinion so it may be flawed:

No is the simple answer to the assertion that it will reduce power pro rata, but the power will be reduced somewhat.

What you have to look at are the AC conditions.  The DC bias is essentially only there to provide enough swing on the amplifier with a low enough source impedance.  If you reduce the quiescent current to 1A then the effect of this will be (amongst other things) to reduce the quiescent VGS (note capital letters for DC condition).  If you do this, then you will have less negative going input headroom before the MOSFET shuts off.  I am assuming that this circuit is driven by line levels as the circuit is after all a source follower and will have essentially unity voltage gain.

If you do increase R6 then you will have to re-adjust the bias resistors R2 and/or R3 to maintain approx 20V at T1 VS (currently 19.8V).

[davebungo]

Yes, my argument was flawed because as it is a voltage follower, VGS is essentially constant (as is the drain current), so the device doesn't cut off as I said.  I will have to have a think about it a bit more unless someone else can contribute...
:oops:

this isn't very professional is it.  Anyway what I think will happen as ID is reduced is that gfs (the forward transconductance of the FET at the operating point i.e. how many amps/volt do you get) will decrease and hence the effective output impedance will increase and you will end up seeing less AC voltage across the load, but I still don't think power will go down in proportion to ID as you originally asked.

(BTW The output impedance of a source follower (or common drain cct) is approx 1/gfs.)

Sorry for the jumbled up response - I'd still be interested in a second or third opinion.

[gez]

http://sound.westhost.com/efficiency.htm#classa

[gez]

'High Power Audio Amplifier Construction'* by R A Penfold is an excellent intoduction to 'rolling your own'.  Covers all the basics like how much gain you need to drive a speaker, power ratings etc.

Dirt cheap new and you'll probably pay more in postage if you can find a 2nd-hand copy on amazon.

* Babani BP 277