VAS buffer

[GEirin]

Carlos
To see thread:
Modify my old 80 amp of ECBLN
Saludos
Guillermo

[GEirin]

Hola Carlos.
Here link.
modify my old 80s amp
Saludos.
Guillermo

[wahab]

hi,..buffering the vas bring vast amount of open loop gain, so
the negative feedback is more efficient to reduce distorsion..
sadly, there s lot more phase shift, making the circuit prone to
oscillation...the solution is to increase the dominant pole capacitance,
thus reducing the high frequency open loop gain and so rendering
the buffer useless..the best is to use one high gain transistor
to do the work...

[AKSA]

Yes, Wahab,

Reluctantly, I think I agree with you. Best device for 8-15mA VAS stage current is a very low Cob, fast transistor like the 2SC3423.

Hugh

[Bigun]

I think a more interesting course of investigation would be to have two Cdom capacitors, not a dominant pole but a split pole. Self waxes lyrically about this for low distortion.

[CBS240]

Originally Posted by wahab

hi,..buffering the vas bring vast amount of open loop gain, so
the negative feedback is more efficient to reduce distorsion..
sadly, there s lot more phase shift, making the circuit prone to
oscillation...the solution is to increase the dominant pole capacitance,
thus reducing the high frequency open loop gain and so rendering
the buffer useless..the best is to use one high gain transistor
to do the work...

This is where the cascode (common base) VAS comes into play. You can use a low voltage transistor for the amplifying device. It is much easier to find a fast, high gain, low noise transistor with a low Vce requirement. And besides, cascode has other benefits too, such as negating the Early effect by maintaining a constant Vce on the amplifying transistor.

[AKSA]

I can agree with the low voltage, high speed transistor observations, but these days there are good devices rated to high speed and voltage, and in listening tests between single VAS and cascoded VAS I never found any advantage at all, and a small cost in headroom with the VAS. There is a small advantage in THD, but it's not audible.

The single, common base VAS - here is an advantage I've noticed in listening tests. More nimble and quieter, but compensation is not easy.

[wahab]

yes hugh, i was reluctant too about the one trans solution....
the bjt you re proposing is just perfect for the job..
currently, i m designing a symetrical differential using a cascode
in the vas, as suggested by CBS240 , but there s some serious
ill-effects near the zero crossing point as soon as the load
is reactive, and i m struggling making the damn thing linear..
and as proposed by BIGUN i choosed to not use a classical
lead dom pole capacitance , i.e from base to collector of the vas,
but rather using two capacitances in two different paths
of negative feedback...
i already use this compensation scheme in the same amp
but with a classical vas, with no cascode....it works
very fine...i ll publish schematics in an another thread, as
a discussion about topologies...