Touching a sacred cow can appear as rude, in particular to those brown nosing around.
The few attempts I tried to provide constructive feedback were welcomed with a dismissive and patronizing attitude, which I also happen to find extremely rude.
That was an obscure old ThermalTrak DIYAUDIO thread with many hundreds of posts, and no post was cited anyway. My book was plainly available at the time and was the appropriate reference. C'mon.
BTW, I did credit you in my book for the technique of mounting the bias sensor on top of the power transistor. I also credited you for the terms "gm doubling" and "wingspread plot"; these are all significant contributions. I did not credit you with the current mirror IPS load, the Darlington VAS and Miller compensation, all of which are old as dirt. You failed to credit Barney Oliver for the 26mV Vt concept and Bart Locanthi for the 3EF Triple.
If you will send me a list of places in my book where I have overlooked deserving references to your work, I will happily include them in my second edition.
Stingy referencing does not make you more of an authority and generous referencing will endear you to your peers.
Bob
Hi Jan,
It does occur to me that sensing being too fast might be a concern if the bypass capacitor in the Vbe multiplier creates a time constant that is near that of the sensor, thus possibly giving one two (or more) too-closely-spaced poles in what is ultimately a feedback loop. I have not given this a lot of thought or plugged in numbers.
The reason that I bring up this concern is that I worry about bias spreader "collapse" when an amplifier clips. This can happen when the VAS current goes to zero.
With normal-size bias spreader bypass capacitors (say, 10uF or 47uF) that capacitor may discharge by the Vbe multiplier resistive divider current, such that the amplifier may be briefly under-biased when exiting the clipping interval. One could consider mitigating the under-bias condition by using a rather large (say, 1000uF) bias spreader bypass capacitor. This would, of course, tend to introduce a longer time constant into the bias spreader. Might a bias instability be created? I don't know.
BTW, bias spreader collapse can also be mitigated by using a normal-sized bypass capacitor with a Darlington or CFP bias spreader that can employ a much larger divider resistance.
On a true DC basis, one could also put a fixed 1mA current source (or the equivalent - think cascode VAS as an example) in paralle with the VAS transistor and use a bias spreader that needed less than 1mA to stay on.
Cheers,
Bob
Which is exactly what I believe about your book, Mr. Cordell, with the addition of lots of pompous hand waving, and at least 25% treated just because your book was supposed to cover everything (class D, for example). And trying to make happy everybody, including as much as possible the Golden Ear Brigade. I believe it's also a new world record of a 600 pages technical book with less than 100 references. And please don't ask for more precise pointers, I am not paid to be your book's professional reviewer (and no, I don't want the job).
But then we reached here the crux of the matter: two famous audio book writers disliking each other, each with his own, ego based, fan club. Cool.
Doug and I are both waiting for you to write YOUR book
.If there ever was someone on this forum who constantly tried to bring himself up by bringing others down, it is you.
Cheers,
Bob
Well-put, Richard. None of us is perfect or all-knowing, but we all usually learn a thing or two when there are technical disagreements that are aired out.
Cheers,
Bob
I know I'm not sacred and would be insulted to be called a cow
.On those very rare instances when you have provided constructive feedback, it has certainly not been met by a dismissive and patronizing attitude; don't confuse that with honest disagreement or honest agreement.
When I was young and graduated from college I also thought I knew it all; then as I grew older I learned how little I knew. I've learned more in audio from those with whom I have disagreed than from those with whom I've agreed, and that includes folks like Doug and Matti Otala and John Curl.
Cheers,
Bob
Dependent on frequency and the Z of the VAS when it clips.
(below) shows this effect.
At 5 hz , we do go into a "collapse" ... but we are clipping ,
who cares. Nothing to worry about - amp and speakers typically survive.
I realized this years ago , combined it with real world solutions.
The shunt R at the VAS and the spreader bypass cap determines what freq.
the "collapse" happens at.
Last post in this thread .... between the bickering and being ignored
despite very good contributions - I give up.
OS
Attachments
Hello Bob
Nicely put, but the 7th edition has not even reached the planning stage yet, though I have accumulated a good deal of new material for it.
I do however have a book coming out this autumn, the 3rd edition of Self On Audio. This adds all my recent articles to the compilation, but more importantly each article now has a much extended preface which in most cases includes new and unpublished stuff, as well as added historical perspectives.
You are right, and I didn't either, so my post was a bit less than firm ;-).
Then again, it's 'just' another control loop problem and lots of people here are very good at solving such.
Nevertheless, TT-devices are quite usefull - I use the Sanken TT Darlingtons in my last few designs and they are really rock-stable, but they cannot keep up with program-dynamic-related thermal transients as I noted in my article. But that is really only of academic interest I would think.
Jan
Hello Bob
Permit me to say that I have never claimed to have invented the term "gm doubling" and I can't say I like it. I know others like Cherry disagree, but I have never found working in terms of gm helpful in the study of output stages. Dagnabbit, it is a voltage amplifier, after all. Possibly I was the first to draw attention to the way it which makes Class-AB such a doubtful proposition, but I certainly don't claim so.
IIRC, I proposed "gain-deficit halving" as a more accurate term, but I said it wouldn't catch on and it hasn't.
"wingspread plot" for sure. I don't think anyone has detected the subtle reference in that one yet.
I wonder whether this mechanism accounts for some audible differences between amps? I mean, beyond the obvious generation of harmonics when clipping?
It would be interesting to see these waveform with a low level higher freq signal say 1kHz superimposed?
Jan
Exactly as I did when deciding on the best diode arrangement in my Leach back in 2006. But P297 of your excellent book, halfway down the fourth paragraph, you mention waiting 20 secs before taking that measurement. I could never understand this.
And I think there are a couple of maths errors on page 322, lines 2 and 3 (if it hasn't already been mentioned).
Brian
I thought that the Sankens were a problem because there was not enough output transistor base pull-down current to suck out minority carriers fast enough at high frequencies, and that there was no opportunity to increase it by external means. Am I wrong? What is the nominal amount of base turn-off current built into these devices?
Cheers,
Bob
Me too!!
Hello folks,
I have always advocated that in the true DIY spirit that these designs, as seen in said referenced books, be available as pcbs for everyone to make. In other words offer CAM data and BOMs, so others can order pcb/parts. Yes put it up on a web site.
Okay I understand that it is a lot of work to capture and layout but at least offer a few examples of the circuits. And yes I know that some of Doug Self's designs are available from Elektor and others.
Also layout is just as important as the design itself. Not including the layout is only doing half the job in designing a complete amplifier.
I can offer one design, the variable freq tone control ckt that I did a few years ago, done in SMT, of course
I never fab'd it, but if anyone is interested I can provide the data and proto it myself first, to give it my blessings. I guess I could go over it again and upgrade it to the low noise version. I would obviously have to be given the rights, as buying the article/book is one thing, making/posting a pcb for public use is another thing.
Actually to be fair I should offer to do one of Doug Self's and one of Bob's designs
I would act as the capture/layout person and the EE's ( Bob/Doug) review the design, similar to my old job as a designer.Cheers
Rick
Last edited:
The primary source for my designs, in PCB, kit or fully built and tested form, is the Signal Transfer Company, of which I am one of the partners. See:
The Signal Transfer Company
Very true indeed.
You will understand that I am in this business as a professional. I won't be reviewing anybody elses's designs unless they pay my consultancy rates.Actually to be fair I should offer to do one of Doug Self's and one of Bob's designs