Hi ostripper, I understand now. Your experiment proves that TMC works, which I don't think anybody here is debating. How could you though estimate the improvement over other compensation methods? Was the Miller compensation still leaving the crossover glitches visible? If so, then as far as your amp quality is concerned you have a much larger matzo ball than TMC in your hands.I have to get Douglas Self article, otherwise I am really surprised, so much ego driven discussions around who's who and who did what about TMC and nobody took the time to breadboard a TMC compensated Blameless, optimize it, and compare the optimized results to standard Miller and two pole compensations having with the same unity loop gain frequency. No Audio Precision is required, all you need is a decent sound card.Never mind, I'll do it myself asap.
Last edited:
Hi Richiem,
Thank you very, very much. The book being received like this makes it all worth it. I am truly glad that you are happy with it. That tells me I have met a very big part of my goal.
Thanks!
Bob
Hi Bob
I think Pete Basel has nailed it; 680R is what was meant. Peter was one of the great audio designers of the twentieth century, but I am sure he would have been the first to admit that he was not completely infallible. I vote typo.
I think Pete Basel has nailed it; 680R is what was meant. Peter was one of the great audio designers of the twentieth century, but I am sure he would have been the first to admit that he was not completely infallible. I vote typo.
Last edited:
I don't have any special "inside" knowledge
the LT1028 datasheet shows the simplified internal circuit at the bottom, the over-compensation C is added from the output to the "final VAS" Q23 base - there is an "extra" diff pair gain stage in the middle so there are multiple compensation loops
the TDA7293/4 both have the text in the datasheet:
"A significant aid in keeping the distortion contributed
by the final stage as low as possible is provided
by the compensation scheme, which exploits
the direct connection of the Miller capacitor
at the amplifier’s output to introduce a local AC
feedback path enclosing the output stage itself."
the LT1028 datasheet shows the simplified internal circuit at the bottom, the over-compensation C is added from the output to the "final VAS" Q23 base - there is an "extra" diff pair gain stage in the middle so there are multiple compensation loops
the TDA7293/4 both have the text in the datasheet:
"A significant aid in keeping the distortion contributed
by the final stage as low as possible is provided
by the compensation scheme, which exploits
the direct connection of the Miller capacitor
at the amplifier’s output to introduce a local AC
feedback path enclosing the output stage itself."
Last edited:
It looks like they "nailed it" !!
I deal in reality (below) , TMC has NOT resulted in ANY performance decrease in amps that I know "inside and out" (Blameless and beta enhanced balanced VAS -symasym). Scope shots and all are soon to come !! The "fabulous world of simulation" can come scary close to the real thing provided the proper models.
OS
Attachments
Baxandall's statement includes the words "or" and "perhaps" leaving it open to interpretation. If you really wish to take the statement literally he says turn the pot to maximum or open the circuit which suggests my interpretation. He would have written that maximum approximates an open circuit if that was what he intended.
What is clear however, is that his intent was to provide TMC and it is illogical to believe that he was unable to determine component values that worked as his description intended.
Last edited:
Very good , It is infantile to debate whether it works. Move on to how to apply it (better) and to other topologies. It was still good to see it's history /origin(s).
Why not have a TMC/ETMC combined thread (compensation in general as well).
This topic will benefit more members (and non-burning/better sounding amps as well
).OS
If we accept both of your statements as truths it requires the belief that Baxandall lacks the ability to select proper component values - something I find absurd given his obvious talent. This leads to my opinion that it was a typo.
Very good , It is infantile to debate whether it works. Move on to how to apply it (better) and to other topologies. It was still good to see it's history /origin(s).
Why not have a TMC/ETMC combined thread (compensation in general as well).
This topic will benefit more members (and non-burning/better sounding amps as well
OS
I also liked seeing the history. I prefer independent threads because these threads get so long that it will just help to keep the discussion on point. They are independent now, so it is also probably easier.
I too am interested in the history and it is clear that Baxandall got the topology and the verbal description of the functionality right. If one suggested resistor value throws it all off, my point is that it is probably a typo or not well thought out, but it does not make all of his work different or wrong. One would be expected to try different values in a prototype (or compute them based on further analysis) in order to provide behavior that matches what is described in the text.
It is clearly documented on this site that you found a good optimization of the values and certainly you should get credit for it.
Last edited:
Opinions on Baxandall taking the feedback after the load isolation network? This did not sit well with me at all on first seeing it, however when we consider that it is only the Miller loop and that TMC transitions to simple Miller at HF it seems more reasonable. I've not simulated it or done any further analysis so I am curious to hear other opinions.