The question was "are you aware of feedback math that does not rely on linearized small signal models"? In what way was my answer a "reading" of what you asked.
Your example in particular of back to back diodes (as in a distortion pedal) does NOT yield at all to this kind of analysis. It would yield to a harmonic balance simulator. The circuit you showed can only be analysed in a large signal sense.
No no, Scott, that's not how it works.
It's YOUR job to DISPROVE the UNSUBSTANTIATED claims made by others.
Yesterday, upon the stair,
I met a man who wasn’t there
He wasn’t there again today
I wish, I wish he’d go away...
se
I am just asking that we should attempt to measure it to the sensitivity that it deserves.
I say this because I still have problems with IC designs that are reasonably fast, measure well, quiet, BUT do not perform as well as my discrete designs, NO MATTER WHAT I DO. This is why we, that is (or was) Matti, Walt Jung and me, for example, still pursue this PIM or FIM distortion. IF we could find IC's that worked WITH NEGATIVE FEEDBACK that sounded as good as our discrete designs with some negative feedback, there would be progress. IF we could make both IC's and discrete designs with negative feedback sound as good as designs WITHOUT NEGATIVE FEEDBACK, I could retire, kick back, and enjoy my royalties.
I say this because I still have problems with IC designs that are reasonably fast, measure well, quiet, BUT do not perform as well as my discrete designs, NO MATTER WHAT I DO. This is why we, that is (or was) Matti, Walt Jung and me, for example, still pursue this PIM or FIM distortion. IF we could find IC's that worked WITH NEGATIVE FEEDBACK that sounded as good as our discrete designs with some negative feedback, there would be progress. IF we could make both IC's and discrete designs with negative feedback sound as good as designs WITHOUT NEGATIVE FEEDBACK, I could retire, kick back, and enjoy my royalties.
I was answering "Where have you seen people using maths which implies the opposite?" as we were discussing published papers. The meaning you got was certainly there and is quite valid, but not the intended one.
As to the diodes I used for the distortion model. I really did look carefully but they really don't seem to behave like real diodes in those circuits. In the second set of experiments the voltage across them was .5 volts or less to keep them out of the hard clip mode as I wanted a softer distortion.
I am not interested in analyzing those circuits as they are not practical audio circuits. I will try to construct more typical audio circuits for the hardware tests.
The idea is to produce a simple article that points out there is a need for proper distribution of signal gain and feedback. As I keep pointing out the flamers who insist on no feedback, local feedback only, lots of global feedback etc, don't really get even the basic issues.
Most readers who understand a simple article will probably model their designs and play with where they place feedback to get better results. I do not intend to aim for a rigorous general case, as even the concept of what the residual distortion should look like is an area of OPINION.
Regarding Otala, his articles are extremely controversial. He has had good points, and wrong explanations as well.
His key article "Circuit design modifications for minimizing transient intermodulation distortion in audio amplifiers" starts with this statement
His key article "Circuit design modifications for minimizing transient intermodulation distortion in audio amplifiers" starts with this statement
and he uses the image attached. This statement is completely wrong, because TIM does not depend on OLG corner frequency at all, it depends only and only on slew rate of the circuit after closing of the feedback.
Attachments
But the simple FET stage goes together quickly and (open loop) is predominantly second order. So I'm trying to make your life easier.
Those were tried by Dick Marsh over 30 years ago to make audio products. Where are those products now? What really does work, for much the same reason as the LH0032 is the OPEN LOOP BANDWIDTH of the AD825, kindly sampled to me by Scott Wurcer. They work well, BUT like the LH0032, they are too noisy for many applications. They sound good, however (for an IC).
Last edited:
How are your ears working John?
The LH0032 is sounding way better than a AD825 or any monolithic IC.....
I listen through it every day!
Where are they? They can be bought from Future Electronics and on Ebay.
Answers.com - What is a monolithic IC
Noise? I am not using them for MC phono. Here is 2SK146 king or matched 2SK369s.
Last edited:
You may have 2 circuit solutions, both with 100Hz OLG corner. One would have 0.5V/us and the second would have 100V/us. Both really do exist. So, it is pointless and wrong to state that it is OLG corner that defines TIM. It is neither OLG corner, nor feedback factor. It is only slew rate that defines TIM, and slew rate depends on circuit design and components used. Yes, stability is a must, but again stability depends on circuit design and compensation scheme.
It appeared, at first, simply by observation, that higher open loop bandwidth meant higher slew rate, but it was NOT always so. This is a 'flaw' in Otala's original argument, but it did NOT ultimately settle the argument, because we found that higher open loop bandwidth was ALSO necessary to make circuits sound good, for the most part. Therefore, the question was still open as to WHY? This is where PIM came to become at least as important as TIM, first with Matti, but also with many others, including me, over the decades. We only want the ANSWER to the problem, not harping on the oversights of the pioneers who started the task.
Hindsight is always 20/20.
I think that next to pointing out possible "flaws", we also need to praise the parts of the work that became the foundation for many other "useful" discoveries. Is not all of this part of the learning curve? Not everyone can be Bach or Pythagorus and still be 100% after hundreds of years.
If the sport becomes knowing so much that you can shoot holes in anyones arguments, why not start making better products yourself instead of peeing on others. That is the curse of investigative journalism. We end up applying our most valuable resources to destruction instead of the collective good.
I think that we are pretty lucky that a handful of these "pioneers" even bother to put up with it.
I think that next to pointing out possible "flaws", we also need to praise the parts of the work that became the foundation for many other "useful" discoveries. Is not all of this part of the learning curve? Not everyone can be Bach or Pythagorus and still be 100% after hundreds of years.
If the sport becomes knowing so much that you can shoot holes in anyones arguments, why not start making better products yourself instead of peeing on others. That is the curse of investigative journalism. We end up applying our most valuable resources to destruction instead of the collective good.
I think that we are pretty lucky that a handful of these "pioneers" even bother to put up with it.
- Status
- Not open for further replies.