Matti Otala - An Amplifier Milestone. Dead or Alive

Jump to Latest
#465
Lumba Ogir said:
I'm sorry to inform you that low open-loop gain and wide bandwidth are unconditionally essential to lower dynamic (and other) distortions. Inaccuracy in time and transients cause GNF to passingly (but frequently) lose control over the close-loop gain and the output signal.
I doubt such a rapid course of events can be measured reliably

Look, I`m not asking you (or anyone else) to agree with me, just to understand what I`m trying to say.
Click to expand...

I'm sorry to inform you that there's nothing to agree and to understand in your statements. The FB loop as a chicken-and-egg problem, or as a dog chasing it's tail problem, denotes a lack of basic understanding of the very basic causality principle. The only limitation of a feedback loop to lose control over the close loop gain is the speed of light. Beyond that, it's only bad design.
 
#468
john curl said:
Some people just don't get it. Mitch Cotter says that hi open loop bandwidth is essential for low differential phase. Prove him wrong, if you can. By the way, he is more experienced than you, Syn08, if not better educated.
Click to expand...

If mr. Cotter is so well experienced and/or educated (I don't have the foggiest idea who the guy is), then he will for sure tell you that passing an (AM) modulated signal through a non-linear, time invariant circuit will always convert the AM signal into a phase modulated signal. This is known for more than 40 years and you can look at this also as a way to convert amplitude intermodulation to PIM. This is what essentially happens in the entire Otala story, and while negative feedback may have a role in this process, there's NO necessary causality relationship involved, from negative feedback to PIM. The variation of the AC transfer function with the signal is just another type of nonlinearity that can aggravate the situation, as described by others.

BTW, am I really your living nightmare? :rofl:
 
#469
john curl said:
Some people just don't get it. Mitch Cotter says that hi open loop bandwidth is essential for low differential phase. Prove him wrong, if you can. By the way, he is more experienced than you, Syn08, if not better educated.
Click to expand...

You make the assertion so why don't you prove it instead of hiding behind someone else's opinion. Then we can see how well educated you are 😉

regards
trev
 
#471
john curl said:
Yes, Bob, as far as MY circuits go, that are NOT class A. Still, with preamps, I have found that no Global feedback is best. That might hold for power amps, but I personally have taken a commercial design far enough to make a successful power amp completely open loop. Charles Hansen HAS done so, and he has a Class A rating, just like me, AND one former JC-1 customer turned in his super JC-1's (Bob Crump's personal amps) for the new AYRE power amp. This tells me that I have to try a zero feedback, just to be sure.
No, I will NOT be boxed into a compromised or conflicted position on this, by anyone.
Click to expand...

Halcro make a class A rated amp and they use lots of feedback in their designs. Care to explain the discrepancy there ??

regards
trev
 
#472
I believe that part of the cause for these repeated back and forths about this subject is partly due to the lack of detail and gross generalization. And it is why I tend to stay out of them.

For example, if an amp has a low open loop bandwidth (OLB) due to older semis with high internal (NON-LINEAR this is important) Miller capacitance and a relatively small or no Cdom cap then it will have high open loop PIM due to the large non-linear Miller capacitance.

On the other hand, if low capacitance semis are used, or methods (cascoding) to minimize their impact, and the low bandwidth is a result of a linear Cdom cap then the PIM will be lower all else being equal.

Given these two low OLB cases, we need to know about the cause of the low OLB to estimate the level of PIM.

When we do not know a lot about an amp, if there are devices with high internal junction capacitance - or not for example, higher open loop bandwidth tends to suggest lower PIM since any capacitances that are setting the OLB have minimal effect in the audio band and even if they are non-linear the phase modulation will be mostly above the audio band.

This suggests that high OLB can be viewed as good, when we do not know much about an amp. It does not suggest that it is a good design goal when we know that the low open loop bandwidth is being set by a linear Cdom cap rather than Miller capacitance for example. If we are designing a new amp, or know the details of the amp then we can analyze/measure the PIM and low OLB it not necessarily bad as others have pointed out.

I don't like to generalize, however the above is the best I can do viewing the amp as a black box where we do not know exactly what is going on inside. I believe that the suggestion for high OLB was simply that - a suggestion in a time when better semis were not available and perhaps the emphasis was on TIM rather than PIM. I've not reread the papers to be absolutely sure when PIM was emphasized.

Pete B.
 
#476
syn08 said:


The only limitation of a feedback loop to lose control over the close loop gain is the speed of light. Beyond that, it's only bad design.
Click to expand...

Isn't that a very simplified explanation?

Trevor White said:


Halcro make a class A rated amp and they use lots of feedback in their designs. Care to explain the discrepancy there ??

regards
trev
Click to expand...

Trevor
Why are you always referring to Halcro?
 
#480
syn08,
I was secretly hoping for your unreserved support in this matter(!)
Seemingly, you are happily unaware of what can happen in a feedback loop and grossly misjudging the time aspect. The limitations are there; in those dreadful moments the open-loop gain is in charge. To you, Otala and some other people have labored in vain.
 
You must log in or register to reply here.
Top Bottom