Historical look at amps

THis is fun. Bombaywalla you need to quit SHOUTING. I was not DEAD WRONG. I was relating what Ayre says. I hold no religious conviction. I live 50 miles from the Ayre factory in Boulder CO and my comments were taken from my discussions with engineering and technical staff members at their factory, not just their marketing hype. I'm not saying your wrong in your opinions which if you read my original post showed the problems of negative feedback. Some reading this post might think this is all a mine-is- bigger-then-yours arguement so I am going to explain the non-circuit design reason for feedback being good and bad..
Feedback is almost always NEGATIVE feedback. Positive feedback is used for oscilators. YOu don't want a amplifier to be an oscillator...its hard on speakers and your ears. Negative feedback takes a bit of the amps output and applies it to the input but inverted (negative sign) This tends to damp oscilations and cancel output errors.
The real problem with feedback is the amp or even the transistor have a non zero propagation time delay. Given the perfect amp a wire with gain the wire itself delays the signal. (Signals do not travel at the speed of light in electronics. The signal travels at just a small fraction of the speed of light, the drift current speed.) All circuitry including the bare wire, take a finit amount of time for a signal presented at its input to reach its output. Feedback circuits regardless of their design therefore are adding an "old" signal that has propagated through the device to the new signal at the device input. For a steady state signal that's not a problem but for music even a few miliseconds of delay mean the feedback signal applied doesnt match the new input signal. Global feedback has this problem in a major way because there is significant delay accumulated through all the amps circuitry. Local feedback would just have to cope with the delay in one FET or transistor. TIM as I mentioned in the first message of this thread is sensitive to this time delay distortion.
By the way there are lots of electronics without feedback. But we are talking about amplifier circuits here so the arguement is whether all or most use negative feedback which is good for stability and bad for music.

Keis,
I was not shouting! I used capital letters to emphasize a word/words. In an all-lower case post, it seemed reasonable to use capital letters to stand out. Sorry that you thought I was shouting.

I re-read you orig. post - certainly appears that you succumbed to Ayre's marketing hype. There was no way to know otherwise. Your latest post suggests otherwise. Wish that you had written the words in your latest post in your original one!

==============================================
Re. feedback in bias ckts - yes, I agree that you do not need to used feedback to design a bias circuit. That's *not* what I meant to say! I re-read my orig post & I cut & paste from there:
"That's because the negative feedback loop allows only so much excursion of the transistor bias point before limiting it".
what I was trying to say was the transistor's bias point's movement along the load line. That's the excursion I'm talking about. Not the static/DC bias point, which is what I think Ar_t & Herman are talking about.
Negative feedback curtails the effective load line movement keeping the transistor in it's linear region of operation. Just trying to clarify my point.
=========================================================

True, today's class-D power amps of tons of negative feedback. I'm wondering (without actually having studied a class-D amp in any details) if today's negative feedback is diff from yester year's negative feedback OR if feedback is feedback is feedback?
In yester year's feedback, as Keis pointed out, the output signal is fedback to some point in along the forward gain path. So, this feedback is directly messing w/ the wanted output signal & botching up the sound.
Today's class-D power amps seem to be using a high-speed switching power supply that makes an attempt to follow the envelope of the music signal while the transistors are merely turned on/off at that same rate. By making the switching speed 64X or 128X or 256X 20KHz, the power supply is able to quantize the music envelope quite well.
There must be an error signal generated to ensure that the power supply is correctly tracking the envelope. It *appears* that the fedback signal is not directly in the music signal path (maybe I'm wrong & it is?!)? Maybe that's why these digital amps (despite the tons of feedback) sound much better than amps of the prev many generations? (I have a friend who seems to love his Rowland 201 mono blocks more than his prev Rowland Model 2).

Technically, you guys are way over my head. I am just a consumer. It seems you techies agree on most of everything. I'm glad that class D has finally made it into this amp history exploration thread.

I do not agree Class D amp's performance value depends on digital power supplies. The best I've heard have analogue power supplies.

Ayre does not use any loop feedback. You don't have to ask them............

As for the sound of amps without global feedback: yes.

The amps we made in the mid-90s (where I got rid of the feedback loop in the output stage) really opened things up. It did sound much more lifelike, although it did lack that certain "punch" that is needed to make a product marketable to a wider market. Fortunately, we were able to find the small segment of the market that wanted something else in their amps.

As for the "closed-in" sound. Yes, that seems to be a direct function of lood feedback. I am not going to try to claim that I know why, but I know that is is true for conventional amps. I can verfiy this by taking the average amps, and lowering the loop gain. This can easily be done by placing a resistor from the collector to the base in the VAS (voltage amplification stage). Yes, the same place that you will find a Miller compensation cap. As the loop gain is lowered, the soundstage opens up, things become more lifelike.

And the bottom end and impact drops off............

(Modding amps this way lead to me to think "Why am I working on the other guy's stuff, and trying to fix the obvious problems? I can do better from scratch." So, I did. 20 years later I wonder what the hell I was thinking.)

BTW........I know from my discussions from those 2 guys who design C-J gear, that they spend a lot of time carefully changing loop gain, to where things just fall into place. Too much, or too little, things don't sound the way they like.

Now.......the obvious question:

Class D had lots of feedback. Yep. Bottom end and punch, right? Yep.

Closed in soundstage?

Nope. Don't ask. I will be the first to admit I have no clue why.

Hi Ar_t
reading your post suggests very clearly that you are a manufacturer of audio equipment. However, you did not make this disclaimer. We love to hear from you audio industry insiders but we also would appreciate if you would be open about it & let us know that you are directly involved.
AudioAsylum forces this in a certain way by putting "(M)" after the moniker (sure one could lie but it wouldn't go over well when one is caught). We don't have that system here on Audiogon & so we rely on the honour system. So, if could please honour the honour system, we'd appreciate it. Thanks!