F6 Amplifier

[Antoinel]

Norton Security Suite did not allow me to complete the above. A potential virus problem. Still, please comment on Figure 2 as it relates to the writing of Mr. Pass. Both of the TCAs in Fig 2 are working in the common source or opposed drains configuration. Will Figure 2 work as a klone [ZM's word useage] to the output stage of simplified F6; theoretically speaking.

[Nelson Pass]

It's a little tough to make fig 2 work - both those parts are N channel in
the drawing. I have never had much luck running N channel parts with the
Source attached to V+.

Presumably he meant to have P channel devices on the top. If you can get
good symmetry between P and N channel parts, you're all set.

[generg]

Cool cool again, Finally!

[Zen Mod]

¿qué

[Zen Mod]

are you boyz talking about part from Self's book ?

this (page 395 etc.):

No doubt something could be done with industrial-sized current-mirrors,
but it struck me that the circuit could be rearranged as Figure 2, by making
use of complementary devices. We now need two bias voltages Vb1, Vb2,
and the positioning of the two signal sources V1, V2 on opposite rails looks a little awkward, but at least the current-difference will be mathematically
perfect, if Kirchhoff has anything to say on the matter.
So far so good. We now have a single current output iout. But is this
any use for driving loudspeakers? I am assuming that current-drive of
speakers is not the final goal; I appreciate that this can be made to work,
and promises some tempting advantages in terms of reducing bass-unit
distortion.4 My immediate reaction to Figure 2 was no, it can’t work,
because with a high impedance output, the output stage gain will vary wildly
with load impedance making the amount of NFB applied a highly variable
quantity. It would also appear that any capacitive loading of this high-
impedance node would generate an immediate output pole that would
make stable compensation a waking nightmare.
However, just as I was discarding the notion, it occurred to me that the
structure in Figure 2 looks very much like the bipolar common emitter
(CE) stage in Figure 3. This is widely used in low voltage opamps because
the low saturation voltage allows a close approach to the rails.5 The more
usual emitter follower type of opamp output is usually called a CC or common-collector stage. It is highly probable that the widest application
of these voltage-efficient CE configurations is in the headphone amplifiers
of personal stereos.
At about the same time I encountered a paper by Cherry6 which pointed
out that, so long as NFB is applied, the output impedance of such a stage
can be as low as for the usual voltage follower type output. Cherry’s paper is
dauntingly mathematical, so I will summarise it thus. The vital point about
using NFB to reduce the output impedance of an amplifier is that the
amount of NFB applied must be calculated assuming that the open-loop case
is unloaded. This condition looks unfamiliar, because the average ampli-
fier usually has a fairly low output resistance even when open-loop, due
to its output follower configuration, and so the loaded/unloaded distinc-
tion makes only a negligible difference when calculating the reduction of
output resistance by NFB.
Using this condition, Cherry shows that output impedance of a CE stage
should be exactly equivalent to the usual CC stage, when the global NFB is
applied. I appreciate that this result is counter-intuitive; it looks as though
the current output version must have a higher output impedance, even
with NFB, but it appears not to be so. Doubters who are unafraid of matrix
algebra should consult Cherry’s paper.

with this pics (attached )

it seems there is a mistake in upper mosfet symbol (polarity)

( as always , I'm doing all the hard work with text and graphics ......... )

[Rush]

Gee it's too bad 2SK82/2SJ28 couldn't be adapted here in figure 2.

But the bias supply would be too hard to manage.

Rush

[Zen Mod]

that fig2 is F5 etc.

[Zen Mod]

Originally Posted by Nelson Pass

.......I have never had much luck running N channel parts with the Source attached to V+......

[jacco vermeulen]

You're slow today, Manuel.

[Zen Mod]

careful with that axe Eugene