Hi Joachim,
a bit off-topic question but I have to grab you here when I saw your reply and in particular the part I'm quoting here above whereby a question raised in my head which I would like to forward over to you.
As you have a profound knowledge of loudspeaker designing (Audiophysics) I would like to ask what is your view of amplifiers with high amount of feedback in conjunction with loudspeakers in terms of overall system distortion?
The reason I ask is as we know, in general high feedback amounts into also very low output impedance of the amplifier and as such I am in particular keen on your view when such amplifier is connected with a loudspeaker.
btw, I remember very well your interview performed by Lars Mytting 1996 and I want to tell I appreciate a lot of your views concerning loudspeaker designing, thanks for your time.
Cheers Michael
TIM : for me it boils down to slew limmiting. when a very fast signal reaches the amplication stage it can happen that a phase delay ocures in the feedback loop. the feedback is not "fast" enough to track the signal. one simple solution is bandwidth limmiting at the input in the form of a lowpass filter usually of 1. order in the form of a shunt capacitor or a series inductor. the signal is "slowed down" and can not provoke TIM.
that seems to be very popular in england also for the reason to filter all garbage that is not in the audible range. another aproach is taken by companies like Spectral that make equipment with very wide bandwidth so it has no problem to cope with fast signals. filtering is them sometimes done in the speaker cables.
i do not want to polute this thread with subjective opinions any more so compare yourself. i see no concern for TIM in Syn08s phonostages. when i see all this cascoding and high speed feedback loops i asume they have no problem to track the signal. overload margins seem to be generous too so ticks and pops on the record should not wreck havoc . i measue TIM with the DAAS system and you can study the manual if you care on that link how it is done: http://www.vandenhul.com/userfiles/docs/Daas4 - pro - manuale.pdf
i do not know why this thread provokes so much off topic issues but phonostage seem THE HOT item momentarily.
that seems to be very popular in england also for the reason to filter all garbage that is not in the audible range. another aproach is taken by companies like Spectral that make equipment with very wide bandwidth so it has no problem to cope with fast signals. filtering is them sometimes done in the speaker cables.
i do not want to polute this thread with subjective opinions any more so compare yourself. i see no concern for TIM in Syn08s phonostages. when i see all this cascoding and high speed feedback loops i asume they have no problem to track the signal. overload margins seem to be generous too so ticks and pops on the record should not wreck havoc . i measue TIM with the DAAS system and you can study the manual if you care on that link how it is done: http://www.vandenhul.com/userfiles/docs/Daas4 - pro - manuale.pdf
i do not know why this thread provokes so much off topic issues but phonostage seem THE HOT item momentarily.
Another sample of messed up technobabble, promoting messed up myths and messed up legends, denoting a lack of understanding of basic physics and EE principles. Here, the feedback loop chasing it's tail, violating the fundamental causality principle.
Joachim, you are entirely correct in your understanding of TIM. Negative feedback does not create TIM, but it reduces TIM, provided that the base amp is fast enough. OTOH, an amp without feedback can be very well be a significant source of TIM. This has been discussed, analyzed and experimentally verified in the literature and in the BT thread ad nauseum. And yes, Matti Ottala was flat wrong in this respect. Which does not diminish his contribution to analyzing TIM, in particular through his measuring method.
Indeed, what would the cure according to you be to address such problem, could perhaps a LP after the variable R||C load network (if such is used) to prevent the input seeing too high levels of such HF content? Not straight forward addressable with a wide variety of source impedances from different cartridges I would guess, would require another variable R-C network but would make the product even less user friendly for the end consumer.
Cheers Michael
TIM is created by a combination of a nonlinear input-second stage, and the slew rate limit that is generally set by gain-bandwidth and inversely by the Gm of the input stage. Most op amp designers use high Gm input stages in order to get the open loop gain as high as possible, and the input noise as low as possible. Tubes and jfets have inherently lower Gm than bipolar transistors, so they tend to have higher slew rate with the same gain bandwidth. Bipolar input stages are sometimes compromised by Gm degeneration by adding significant emitter resistors, but this will make the input stage noisy, very noisy, in most cases. Global negative feedback of a significant amount must be used in a given amplifier design just to create slew rate. If the global negative feedback is too low, usually 20 dB or less, like many tube amp designs of the past, or with low feedback designs made by some solid state designers, today, the amplifier will only rise-time limit, not slew rate limit, and this does NOT add any TIM like artifacts. This is an ideal state, but seldom done really effectively, to be both low distortion and high open loop bandwidth, that keeps whatever feedback used, equal at all frequencies, and from allowing slew rate limiting, allowed by the fact that the open loop bandwidth has been exceeded by the musical signal and its distortion artifacts, already present in the program.
Those of you who like tube designs, should look carefully as to how they are not easily slew rate limited, and are usually rise-time limited, for this very set of reasons.
Those of you who like tube designs, should look carefully as to how they are not easily slew rate limited, and are usually rise-time limited, for this very set of reasons.
Last edited:
Please stop bashing the phds.
This goes well beyond a honest mistake or misunderstanding. Its a disgrace for anybody calling himself an engineer.
John, the text you wrote above should go straight to the IEEE board. They ought to know who they have as lifetime member. Its a shame and if you would dare to publish such nonsense in any IEEE journal (assuming it would pass the peer review process
) that would clearly show why the "high end audio" industry is today nothing but a mockery.Your personal comments as above would also show to everybody your knee jerk moral stature.
Sorry for pooping your party boys, now go ahead and pour your crap.
Joachim,
I agree with almost all of this, I just would like to comment on your first statement. Generally the slew rate limiting occurs in the forward path of the amp as the feedback path is generally much faster than the forward path. That is also the reason that TIM (caused by slew rate limiting) occurs even without feedback being present. What feedback can do is prolong the slew rate limiting because the error signal overloads internal amp stages. The solution is simple: design your amp so that it doesn't slew limit and/or has good overload margin.
And your tips on bandlimiting are one way to do it. Matti O gave us another way, and there may be yet others.
BTW I have done an interview with Jan Lohstroh who actually designed and build the 'Otala amp" and he has a view of this as well. It will be published early next year I think.
jd
This is patently untrue, but since you avoid backing it up with anything, there no real discussion possible is there?
jd
Stop making a fool of yourself John, several people here are getting 'embarassement by proxy'.
jd
So I gather some are disputing Otala's findings in favor of Bob Cordell's findings with regard to TIM and other dynamic distortions and the use of negative feedback. So negative feedback, in large doses, is OK these days? Is that a fair statement?
What has changed for this change in design philosophy since Otala and others have made their findings in the 1970's? I am not aware of a lot of these things. I was still under the impression that negative feedback was sinister and to be avoided like the plague, at least in heavy doses. Is it now a panacea for all circuits and designs so that low open loop bandwidth distortions are no longer important? I am trying to understand when this happened and why.
What has changed for this change in design philosophy since Otala and others have made their findings in the 1970's? I am not aware of a lot of these things. I was still under the impression that negative feedback was sinister and to be avoided like the plague, at least in heavy doses. Is it now a panacea for all circuits and designs so that low open loop bandwidth distortions are no longer important? I am trying to understand when this happened and why.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.