Joachim,
I think it is inherent. For stability, you need to roll of the gain with increasing frequency to 0dB gain before you have 180 degrees phase shift. That said though, you can easily built a stable amp with bandwidth much larger than audio.
But I'm sure you know that
jd
The JC-1 power amp as well as the rest of the amps that I design, have a minimum of 100V/us slew-rate, and unconditional stability with load (as far as I know) without any output coils necessary. It is a matter of gain-bandwidth of the output devices and proper compensation.
Of course, all else being equal, the Gm of the input stage is also a factor in minimizing the slew rate. If you have never read Solomon, or equivalent texts, and have no idea what I am talking about, you should avoid designing power amps.
Of course, all else being equal, the Gm of the input stage is also a factor in minimizing the slew rate. If you have never read Solomon, or equivalent texts, and have no idea what I am talking about, you should avoid designing power amps.
Yes, i do. I finished my survay into non global feedback phonostages, especially headamps. Now i am looking at global feedback because you reach a certain bondary (although very low and debatable if audible) where you can not improve distortion any more. I ran into some stability problems because i use enormous gain. Allen Wright called it "scary" and plan to look into that glas more deeply.
I stated to play with Microcap.
Anyway, an "open loop" circuit can be made very fast without stability problem quite easily. With feedback it is more difficult but of cause solvable.
I aways had a week spot for fast circuits if only for the technical aspect. I do not want to start a discussion if they sound any better. That could get quite lengthy and tiring here.
Well, i did a preamp that has a bandwidth of 50Mhz -3dB) with feedback but that uses a transimpedance amp.
I stated to play with Microcap.
Anyway, an "open loop" circuit can be made very fast without stability problem quite easily. With feedback it is more difficult but of cause solvable.
I aways had a week spot for fast circuits if only for the technical aspect. I do not want to start a discussion if they sound any better. That could get quite lengthy and tiring here.
Well, i did a preamp that has a bandwidth of 50Mhz -3dB) with feedback but that uses a transimpedance amp.
Joachim, it is NOT the question, but the lack of seriousness between us, as to finding the answer. I was designing for about 5 years, before I first saw this Solomon article. It answered virtually everything in this area of design. I even went back to university, before I read this article, and was taught by the very same professors who taught Solomon, BUT I still overlooked some important aspects. Read first, then we can discuss the subject easily.
Joachim, please learn to fill the holes in your education, as I do with my other associates who teach me. It is not a personal thing, just something that must be discussed. I don't have the time or energy to re-write what Solomon has already put down. By the way, by the time that I first read this article, I had designed dozens of power amps, and was working on the design of the JC-3 power amp. One amp that I made for Ampex Research, in 1969, was a balanced bridge floating current output with a +/-50A swing making it about 2500W worst case. Still, I did NOT know what Solomon put down so clearly.
I take your advice seriously. My speciality is low level circuits and loudspeakers. That´s where i am really goud at i think. As a speaker designer i am of cause very interested in the amp-speaker interface but all amp designers i worked with over the years really do not take the loudspeaker into account much. When the amp drives two Ohms and some capacitance they are happy.
My measurements with the Klippel Distortion Analyser tell me a totally different storry. I can measure distortion in the voltage and in the CURRENT with high precission. Maybe i could contribute a thing or two that is not mentioned in that tutorial. Just a guess and a gut feeling that the dynamically changing loudspeaker impedance is not that easy a problem. A lot of hystheresis is going on too, making it difficult to predict with linear analysis. Just my 50 cents of words.
My measurements with the Klippel Distortion Analyser tell me a totally different storry. I can measure distortion in the voltage and in the CURRENT with high precission. Maybe i could contribute a thing or two that is not mentioned in that tutorial. Just a guess and a gut feeling that the dynamically changing loudspeaker impedance is not that easy a problem. A lot of hystheresis is going on too, making it difficult to predict with linear analysis. Just my 50 cents of words.
I might point out to everyone that the key to high slew rate, all else being essentially equal, is reduction of the Gm of the input stage, when the circuit is frequency compensated AFTER the input stage. A single transistor or even a transistor differential pair, is ALWAYS weak in slew rate, IF this is not taken into consideration. I solved the problem in the early 1970's by switching to comp diff fet input from comp diff bipolar input, used for approximately 5 years previously. Another method is emitter resistor degeneration, that is used by many designers, the third method is a different form of bandwidth equalization, as used by Otala and Lohstroh. The only other real way to get good slew rate with stability is to increase the effective gain bandwidth. What is amazing, is that nothing else really matters much.
understood.
If you like i could send you latest measurement of a german TÜV and ISO certified lab, the TEST FACTORY that made measurements on the amp-speaker interface. They test amps with speakers connected. They put in an MLS signal and do an FFT of the first 100msec. A lot of amps with high feedback and perfectly good standart measurements ( low distortion, high damping factor etc.) had a lot of trouble to control that load in this short time segment. I will not alaborate more now what they found out because this is stuff for a lot of controversy. When you have looked through the material they collected over many years you can decide if this is worthy to discuss here.
If you like i could send you latest measurement of a german TÜV and ISO certified lab, the TEST FACTORY that made measurements on the amp-speaker interface. They test amps with speakers connected. They put in an MLS signal and do an FFT of the first 100msec. A lot of amps with high feedback and perfectly good standart measurements ( low distortion, high damping factor etc.) had a lot of trouble to control that load in this short time segment. I will not alaborate more now what they found out because this is stuff for a lot of controversy. When you have looked through the material they collected over many years you can decide if this is worthy to discuss here.
Joachim, that sounds like an interesting subject, but I think that it is somewhat different from what is considered normal stability analysis. Usually we can use any square wave with a difficult range of capacitive loads, and we look for ringing, and even oscillation. Your test sounds like something different, is it?
They claim that an MLS signal is more music like. It exites the amp at all frequencies at ones. When they measured with a swept sine the differences where much smaler.
The ouput of that test looks like a damped version of the impedance curve. The high feedback amps had trouble especially in the treble but also at lower frequencies. The amps with the best results ( more or less a straigt line) were the ones with little or no global feedback like the Ayre, Pass, Brinkmann et.all.
Tubes with output transformer did well too. They concluded that they had found a mechanism why that amps sounded so good in the subjective apraisal but had less good
statik measurements like distortion etc. That result makes that test so controversial. What is interesting is the time table: first they got news from subjective reviewers that for example the Brinkmann sounds excellent but measured mediocre in the lab and THEN they stated to search for an answer to bridge that gap. That is an example that a "subjectist" was insisting on a sound difference ( this time in favour for the not so well measuring amps) and the lab searched for a measurement that could explain that and not the usual way around like : "It measures perfect so your listening is at fault".
Peter Schüller and Hannes Mayer presented there findings in public during the Munich High End Show so theese technicians are not afraid to stand behing their findings. Janemann may have heard about that experimens too.
The ouput of that test looks like a damped version of the impedance curve. The high feedback amps had trouble especially in the treble but also at lower frequencies. The amps with the best results ( more or less a straigt line) were the ones with little or no global feedback like the Ayre, Pass, Brinkmann et.all.
Tubes with output transformer did well too. They concluded that they had found a mechanism why that amps sounded so good in the subjective apraisal but had less good
statik measurements like distortion etc. That result makes that test so controversial. What is interesting is the time table: first they got news from subjective reviewers that for example the Brinkmann sounds excellent but measured mediocre in the lab and THEN they stated to search for an answer to bridge that gap. That is an example that a "subjectist" was insisting on a sound difference ( this time in favour for the not so well measuring amps) and the lab searched for a measurement that could explain that and not the usual way around like : "It measures perfect so your listening is at fault".
Peter Schüller and Hannes Mayer presented there findings in public during the Munich High End Show so theese technicians are not afraid to stand behing their findings. Janemann may have heard about that experimens too.
By chance noticed a very interesting subject I have been waiting for years to be discussed in depth; amplifier design philosophy AND loudspeakers... for about 4 years I was amused by a discussion of the never ending story is a ultra low distorting/high damp factor amplifier the answer, is a Halcro (they have become personified with the statement worlds "best" or "lowest" distorting amplifier) and alike the right path and how does the amplifier and loudspeaker in conjunction look like regarding the system distortion and why.
Joachim did the German test team ever look at the loudspeaker electro-dynamic behaviour or did they solely concentrate around the amplifier, if you don't mind I would kindly ask for a copy as well I would be most delighted, you have already my e-mail address.
Cheers Michael
Joachim did the German test team ever look at the loudspeaker electro-dynamic behaviour or did they solely concentrate around the amplifier, if you don't mind I would kindly ask for a copy as well I would be most delighted, you have already my e-mail address.
Cheers Michael
- Status
- Not open for further replies.