Model the input impedance. 20Hz to 20kHz should be enough as a first estimate. What you have then is called Transfer Linearity. Even in a shunt feedback, input inclusive situation that matters because the effective feedback resistors are not linear with frequency any more. This is mostly a high frequency phenomenon and distortion rises with frequency if the Transfer Linearity is compromised.
Hi
That thing is not a CFA, but a VFA.
In my opinion it is overly complex in the "wrong" places.
All the best
bicycle repair man
C'mon now, let's keep it polite.
Well, I think that topic was beaten to death here some years ago.
Not perhaps the best choice as an example. My thermal simulation work was first published in May and June 1996, in Electronics World. It then appeared in every edition of APAD from the start. Your book was published in 2011, and the the thermal stuff is totally derived from mine. For example, your Fig 14.21 is clearly derived from my Fig 8 in the May article. I only reference original work.
As I said, I only reference original work.
I fear you miss the point. The Blamelessness of the amplifier is not due to the configuration (though that certainly helps, if you compare others) but attending to several detailed points of design, and making sure you attend to all of them. It's not just a configuration, it's also a philosophy for avoiding mistakes. Thus Blameless..
That may or may not be so, but I don't think anyone would doubt that I was the first person to apply them to audio power amplifiers. The British Patent Office certainly thought so. Will your next edition have anything on crossover displacement?
I cover a great many more amplifier architectures, which I jokingly term "Selfless".
Go ahead. But I suggest you keep an eye on Chapter 8 of Audio Power Amplifier Design (6th edition) where it is shown (for, I believe the first time) that all the common approaches are unable to maintain the critically important balance in an input LTP, and unable to maintain a stable current through a bias generator. These are very real problems, as I assume you agree, and need solving in any amplifier that aims to be better than Blameless.
6 feet from my elbow is a Blameless that gives 0.0007% at full power at 20 kHz. It's had a tweak or two.
Don't you think that topic is done to death too? You think you've found a mistake I've made, and you seem to be obsessed with it. If that's the only mistake in 700-page book, I wouldn't feel too bad. But since I draw attention to it and say 'I screwed up!' Here's how to fix it!' it is perhaps not appropriate to keep chastizing on about it.
And I still think that Politeness is the Privilege of Amplifier Designers. Dumb?
By the way, I've got a new book coming out in April....
Last edited:
Its still a CFA even with the added input buffer for balanced operation.
Not really that complex... just some cascodes around but still basic topology of CFA and some good techniques.
-RNM
Dear friends, may I offer, as contribution to this discussion, the relevant part of a chapter from Audio Power Amplifiers?
Dr. Kolinummi was specifically interested in investigating distortion mechanisms in open loop stages, so this may be of interest.
Jan
Dr. Kolinummi was specifically interested in investigating distortion mechanisms in open loop stages, so this may be of interest.
Jan
No it is not a CFA, but a VFA
I'm not talking about the "not needed" buffer, but how the whole circuit works.
I agree that it isn't an overall complex thing, but I think that it is unnessesary complex in some stages, I guess that is because you have to compensate for a "not a very good design"
All the best
Reodor, you mean the buffer on the inverting input which when removed would put R1 (820 Ohms) to ground and the 22K Rf as the gain network? Yes, then this is not a CFA. The feedback is still buffered and changing gain by increasing or decreasing R1 does not change the open-loop transfer function of the main amplifier. H-bridge style amplifiers are not CFA's they are slew enhanced VFA's.
Last edited:
Yes you are correct of course. I meant my CFA with added buffer at inverting input and it became VFA with all advantage of previous amp. This is possible to do with my amp just changing low impedance CFA feedback board with the other small board with the higher impedance feedback with buffered inverting input.
I think we had that discussion before with the same amp and it showed VFA behavior in a simulation.
Last edited:
We agree to disagree. It is time to return to our regularly scheduled programming.
Bob
yes and I agree how it behaves externally as VFA but I choose to call it CFA with buffered input so the topology is other-wise under stood. We can agree to disagree on how you want to determine operation. It definitely isn't the Blameless topology.
The tests I refer to and showed THD, however, does Not use the buffered input version CFA->VFA but the UnBal version only is tested and used.
Though CMR can be improved, if desired, and diff input has potential to excel for CMR with Bal ips --- I have not found the added circuitry for Bal input to be needed in a rather benign home environment with relatively short interconnect lengths with preamp and PA gains significantly > 1.
In fact, I have Never used the Bal I/O on any of my home gear - ever.
THx-RNMarsh
Last edited:
I know this is probably a cop-out answer, but I think it depends on what is the dominant source of the input current distortion in the first place. For example, if the original input current distortion was basically due to the input-referred signal of the open-loop, and its distortion, then little improvement might be expected. By input-referred, I mean the signal current at the input of the amplifier to obtain the required output. This might be thought of as the open-loop transimpedance of the amplifier - just thinking out loud here.
On the other hand, if the input current distortion were the result of some shortcoming in the input stage, like a poor tail current source or junction capacitance effects, the even-order input current distortions might be canceled to some extent, depending on how well-matched the complementary parts of the input stage are, and how well-matched their shortcomings are.
Cheers,
Bob