I don't know if it has been linked before in the pages of this thread, but here is the ''Otala - Lohstroh'' amplifier commercial story from Electrocompaniet's mouth. http://home.online.no/~tsandstr/OtalaStory.htm
This first Electrocompaniet amp is still a masterpiece. I have one in use, today. One of the best amps ever designed. How about that, Jan?
Salas said:
This text the link, is saying that high negative feedback cause TIM distortion 😕
I know that is the low Slew Rate would cause TIM distortion !
Rafael.luc said:
Actually, both are equally needed. For HF stability one stage has to be dominant in terms of slowness, so the previous one in the loop would have to deliver more power than is nominally needed. The loop is round, so any stage that precedes the slower one being pushed by a feedback would try harder generating more of non-linearities than could be expected during normal operations.
Rafael.luc said:
''So, we realised than the balance of distortions were the essential factor to consider. Not only do you have to balance THD against TIM, but also low frequency distortion against high frequency distortion, frequency and phase response against non-linear distortion in general, and so on.
This insight triggered the Great Change:
One night (it always happened at night times!) we increased the feedback 10dB, to a total amount of 30dB feedback. The sound improvement was staggering!!! And, contrary to common belief in our own community !
After this we only adjusted the amplifier slightly. It had found the form it should have for its remaining life at Electrocompaniet. And, the amplifier was a success!''
We have found 'open loop bandwidth' to still be just as important for sonic quality as slew rate, etc.
One of their basic specifications that they were trying to match in their production against the prototype indeed.
Rafael.luc,
there`s a tight relationship between these things. Dynamic distortions (DIM, TIM, SID) occur only in connection with global feedback. Slew-rate limitation, that is poor bandwidth and the associated phase errors directly causes TIM.
there`s a tight relationship between these things. Dynamic distortions (DIM, TIM, SID) occur only in connection with global feedback. Slew-rate limitation, that is poor bandwidth and the associated phase errors directly causes TIM.
John,
Wide bandwidth is generally needed for linearity, not just to keep dynamic distortions low.
Lumba Ogir said:
Dear Lumba;
dynamic distortions actually are manifestations of non-linearities that increase with frequency.
Lumba Ogir said:
Absolutely not necessary. Any stage may be overdriven by high frequency signals without any feedback. Feedback just highlight the problem not being able to react so fast to compensate errors. Local feedback, like an emitter degenerator, may show similar problems due to capacitances. Miller effect in tubes can cause similar errors without any feedback applied.
Wavebourn said:
Indeed! And slew rate has nothing to do with open-loop bandwidth, only the input stage bias current and compensation capacitor. It's possible to make radical changes in open-loop bandwidth without changing the slew rate at all, by crippling the VAS with a resistive load.
In my opinion, if the amp is prepared for a high negative feedback, that would be the differential pair with a considerable degeneration to receive a big difference.
Only the negative feedback would not cause any distortion (including Dynamic distortions ) the cause of distortions TIM, would low Slew Rate and TDH in HF..
Only the negative feedback would not cause any distortion (including Dynamic distortions ) the cause of distortions TIM, would low Slew Rate and TDH in HF..
Even if you do not use negative feedback, is recommend you use differential pair, it has less distortion (Attached)Wavebourn said:
Wavebourn,
The properties of all devices (passive and active) deteriorate with increasing frequency. The ideal rate of bandwidth is infinite (lower than that means distortion).
distortions are insufficiency symptoms, of course.
The properties of all devices (passive and active) deteriorate with increasing frequency. The ideal rate of bandwidth is infinite (lower than that means distortion).
Rafael.luc said:
Thanks, I see no reason in partial compensation in small signal mode of even order distortions. Such solution of non-significant problems causes bigger problems, more significant.
Wavebourn,
It is that way as stated. Otala`s and others comprehensive work shows that. Otherwise, it can be realized by logical thinking.
Lumba Ogir said:
No, logical thinking doesn't help here. It may be intuitive, but nevertheless not correct.
I believe that even Matti later realised that it was more subtle than that, and that you can have all those pesky TIM, DIM, SR etc without feedback. Several discussions on this forum also made that clear.
BTW Why do you think that more bandwidth leads to lower distortion? I can't see than connection.
Edit: are you thinking about freq response deviations as linear distortions (which IMHO they are)?
jd
Makes sense. We are still unable to explain sound differences only by degree of FB, OL corner and SR. Just hints, not sufficient conditions.
andy_c said:
Absolutely correct, Andy.
IF open-loop bandwidth has anything to do with sound quality, it is NOT because it has anything to do with slew rate, TIM, PIM etc. It has been shown that with proper design those distortions are just as low or lower with high-feedback, low open-loop bandwidth designs.
IF open-loop bandwidth has anything to do with sound quality, it is not something we have yet been able to measure, to the best of my knowledge.
Cheers,
Bob