[Patrick Turner]

Phil wrote:

Eeyore wrote:

Phil wrote:


Phil Allison wrote:

"Phil"
Andre Jute wrote:

We're
not talking about "-30 degrees at 20 KHz," we're talking *dynamic* phase
shifting, the kind that makes a Crown preamp bite your ears off, while
testing at 0.0001% THD.

Let's hear some more about this dynamic phase shift that pours a pint
of vinegar into a Crown preamp. I'm not overimpressed with vanishign
THD but this is an amazing explanation for why so many silicon amps,
and not a few tube amps, sound like ****.

Pretty much what Patrick said, although I need to reply in the original
thread. What Matti Otala PROVED --


** Otala never proved one, single commercial hi-fi amp suffered from TIM in
a way that was audible.

Many others have proved conclusively that TIM ( ad his cousin SID) is a
furphy.

Could you name an article showing that slewing induced distortions do
not exist? I can use an op-amp with 1 V per millisecond maximum slew
rate with no problems with signals at 10 V per millisecond?


Eh ?



Surely
someone as intelligent as yourself (although it pains me to admit it)
wouldn't use that old, tired, debating trick of throwing out a "general
criticism without any supporting examples."


Firstly your slew rate figures are surely V/us ( microsecond )

No, I'm making a point that, contrary to what Phil said, slew induced
distortion can indeed be a problem.

For any sinewave ( see Fourier theory for applicability ) the max slew rate ( at
zero crossing btw ) is 2.pi.f.Vpeak.

For a 20kHz signal of say 2V peak amplitude that means a slew rate of 0.25 V/us
! Even a rubbish 741 or 1458 can manage that !

Given that modern audio op-amps are capable of slew rates of ~ 10V/us - you're
never even remotely close to slew limited anything.

Graham

Well, remember, the maximum slew rate found in audio signals is much
greater than what a theoretical 20 KHz signal is going to supply, and
not all power amps, with their big, slow, output transistors, are going
to be as fast as even a 741.

Not all output bjts are big and slow.
Some do however dislike turning OFF quickly and some display the truly
horrible habit of cross conduction at HF, ie, the two bjts in a typical complementary
pair
are BOTH turned on during a wave cycle during large signal excursions at above 10kHz,
and the power supply has to supply a lot more current that is simply passing from rail
to rail
and its hang onto your hat time for the ride.

Plus, the point of Matti's work is that
problems begin to appear at all levels below the theoretical
"breakthrough" point of TIM/SID. In any case, the topic here is not
whether most amps have sufficient slew rate -- I assume that *most* good
amps do -- but rather about Otala's proof that a feedback amp's
"correction" of an amplitude distortion of the open loop phase shifts
the high frequency components in the closed loop.

Both amplitude distortions and phase distortions of the open loop
response are BOTH corrected by the NFB.
Typical open loop phase lag in open loop at 20kHz is 90 degrees, and the 40dB of
applied global NFB
at 20kHz reduces this typically to less than 5 degrees.

I am currently
discussing this in the "Negative Feedback in Triodes: The Logical and
Experimental Proof" thread from 8/15, so if you're interested, look
there (articles posted on 9/6). Phil Allison had a "response" here --
his usual slams and blams with no supporting evidence -- but I actually
would like to see his simple test that can show whether low frequency
signals in a feedback amp do or do not cause high frequency phase
shifting, as it would be useful test, and I'm having a hard time coming
up with a simple way to test that myself.

Just apply 70Hz and 5kHz signals to the input of an amp
in a 4:1 ratio.
Filter out all below 1kHz from the output signal.
Then you will see what the effect of the 70Hz large signal is upon the
fidelity of the 5kHz signal and whether there is any phase modulation
in addition to the expected intermodulation.
With most well made SS high NFB amps, the IMD is not visible on the CRO
and a careful peak detector must be used to measure amplitude variations in the 5kHz,
or else filter out the IMD products at 4,930Hz and 5,070Hz.




Apparently, Otala incorporated
a lot of ideas/solutions into his Citation XX power amp, and maybe, if I
can find papers by him on that amp, there will be some useful
information and tests there, but if PA can come up with something in the
meantime, hell that's fine by me! He'll probably think of something
really simple and easy, and then refuse to tell me, the ****head ...

But all these investigations have been done many times before.

What exactly do you hope to gain by goading the ungoadables on the group
to find out what you should be willing to find out for yourself?

Do you suspect to find some hitherto unused uninvented techniques of making
amplifiers perform better?



Anyway, what PA was saying/yelling is that feedback amps DO NOT EITHER
CAUSE PHASE SHIFTING OF THE HIGH FREQUENCIES LIKE THAT DUMMY DR. OTALA
SAID! My response was simply to ask whether (1) he knew of references
that would back up his claim, that Otala's analysis was flawed, and (2)
whether he knew of a good, simple test that can be used to test whether
LF signals in a feedback amp cause phase shifting of the HF signals,
like Otala said they do.

Be like me, find out by building one's own test gear and testing.
It took me months to do it all but after reading all the conflicting opinions about
all this in
Electronics World copies from the 1970s to 1980s BEFORE the internet was mainstream,
I decided to look myself at what happened in amps that i should be worried about.

Patrick Turner.



Phil

To email me directly, cut off my head