Maximum current in VAS stage

[rtarbell]

I have a VAS stage with ~5mA of DC bias current through it. I'd like to know, what can I expect for the maximum current (audio signal current + DC bias current) through each transistor of the stage? Is there a general rule of thumb to use?

==> When I simulate it in TopSpice, I get maximum currents of about 60mA. This seems high to me, but then again, it could be correct.

[jcx]

typically the vas stage is a single ended ce or cs stage operating into a large impedance at the base/gate of the output device (pre)driver stage(s) to give a large voltage gain, as such it is usually class A and bias i must exceed anticiptaed i_max required for AC and DC current load max swing/low load Z and slew rate conditions

[AndrewT]

Hi,
the VAS stage is a single ended ClassA stage.
The maximum symmetrical current output from the stage is two times Iq.

I suspect if the VAS was limited to exactly two times Iq, it may have a deleterious effect on sound quality.

I have arrived at, but not yet proved, that a sensible upper limit for the maximum output current is three times Iq and that hopefully, this will never be audible in any circumstance, short of damaging the amplifier.

I recommend a VAS protection as detailed in the Leach site (Lo tim 4.5) but with the monitoring resistor increased slightly to achieve 3times rather than Leach's much higher peak current which will overload many delicate VAS transistors. Using Leach's protection topology the resistor needed for 15mApk is about 43r.

[traderbam]

Doesn't a single ended stage have +-Iq as its symmetrical current output as there is a CCS of value Iq supplying one swing?

[AndrewT]

Hi,
if the Vas is on the top (+ve) side and the CCS is on the -ve side then the maximum sink current is with the CCS fully full on (=Iq) and the VAS full off, resulting in max sink for the output=Iq.

The maximum symetrical source current is also =Iq when the CCS is still pulling Iq and the VAS has turned on to allow 2 times Iq. The difference between the source and sink is supplied to the output.

However the VAS can turn on harder and supply more than 2 times Iq and may usefully provide upto many times Iq when driven by very non symmetrical signals.

There must be some sensible limit to what the unsymmetrical signal can demand of the current sourcing VAS.

I am suggesting that setting a protected limit for the VAS source at 3 times Iq will sound as if any limiting was not present for all normal music and sound effect signals.
There may be some test signal that could force triggering of the VAS limiting, but that would not be called entertainment in my definition.

Do you see where I am coming from?

Any disagreement?

[darkfenriz]

It all depends on output stage.

For cascade of 3 emitter follower stages I would recommend 2-5mA.
For 2 EF stages ('darlington') 7-20mA
For mosfet source follower 20-40mA
For EF driving SF around 5mA.

For VAS possible to switch to class B these currents may be lower.
It is all about compromises. Higher current allows sometimes higher slew rate, higher transconductance and therefore gain, lower output stage loading distortion, but on the other hand higher dissipation (need more powerful transistor?) and more early effect distortion.

regards
Adam