Small mod, what is the effect?

[kroto]

Hi all,

I have seen some amp that unusual.
I'm curious with it, this is some of it :

what is C4 doing in there? connected between Vas Collector load and inv-input. (somebody say it was a phase-lead feedback cap)

is it increase stability?
is it the same if C4 placed in parallel with R3?

I simulate that and found that the signal is shifting, is it good?.

* green = without C4 ; red = with C4 installed
* input signal is sinewave 50kHz@1Vp.


Oh, and also, there is now R11, connected in parallel with the NFB resistor.

what is the bennefeit of such that feedback?
with R11 installed in there, the output signal is about 1V higher.
is it good? or bad?


thanks folks,,,

[Mooly]

Well this looks a lot like a rehash of Doug Selfs "blameless class B" design.

R11... should have no effect (it is not in parallel with the feedback resistor). The only reason it may do is at frequencies where the reactance of C3 is becoming high (in other words at LF). If you connect the end of R11 to the other end of C3 the effect should be zero.

C4 introduces hf feedback, here that signal is taken from the VAS stage rather than the output as in D Selfs original. It a valid approach.

Edit... I also see ground lift resistor R18. That also causes undesirable effects with R11 feeding into the input ground. Short out R18 and run your simulation again.

[kroto]

Thanks Mooly,
I will simulate again tomorrow, time to sleep now...

by the way, is C4 gives better stability?

[Bigun]

C4 may allow you to use a smaller C6. Some people prefer that.

[wahab]

Originally Posted by Mooly

Well this looks a lot like a rehash of Doug Selfs "blameless class B" design.

.

It s a LIN topology with a current mirror.
It makes me always jump when i read that these
kind of topology is from douglas self, since he
invented none of the topologies that are in use
to this day...

[kroto]

Originally Posted by Bigun

C4 may allow you to use a smaller C6. Some people prefer that.

I see,,,
my friend have been built such amp with "C4", he use the same value Miller cap, just the tail current is bigger (maybe twice), make it have higher slew rate isn't it? he also use a video transistor for the Vas (very high fT and low Cob), its still not oscillate though.

[Mooly]

Originally Posted by wahab

It s a LIN topology with a current mirror.
It makes me always jump when i read that these
kind of topology is from douglas self, since he
invented none of the topologies that are in use
to this day...

I'm not suggesting he did invent it... what he did do was analyse each aspect of the design and push the performance to pretty impressive limits.
Whether that delivers what folk want sonically is open to discussion, but you can't argue with his method, or the results obtained... well I don't think so anyway.

[kroto]

this is the R11 simulation graph:

* Vin = 20Hz@ 1Vp
* red = with R18 installed
* green = R18 shorted
* blue = without R11 and shorted R18

with or without R18 its make no big diffrences.
just the starting has lower amplitude with R11 installed.

Am I really need R11 in there?
or the R11 value need to be lowered? like 100R?
or with similiar value with the Xc of C3 @ 20Hz?

and sorry for my bad english, I hope you know what I'm talking about.

[kroto]

oh, sorry, I must go now, seya

[Mooly]

Think about it
R11 is a 1k resistor.
The effect of that on loading the output stage is absolutely minimal.

So the only way that R11 can affect the amplifier is by the fact that it's feeding signal back into the input stage... which in a way is what you have with it connected to C3 as shown in the original diagram.

First... if you change the frequency to 20000 hz instead of 20 then the effect of C3 is reduced. Try that...

Secondly, connect the end of R11 to ground instead of to C3.

If it still alters the output then your simulation is wrong somehow