It should be noted that there is some compensation in that the second stage devices are complementary to their respective input stage devices. Also, VOLTAGE DRIVEN second stages have essentially the same gain with bipolar transistors on both sides. This can be very important.
Last edited:
PMA, I would like to acknowledge your input here. You are (in my opinion) on the right track, the same path as Charles Hansen and I have followed over the decades. It is significant that you have been skeptical in the past, and 'had to prove it to yourself' before accepting what we were 'promoting', and do your own measurements, simulations, and listening tests, without any direct input from us.
All we care about, in the final summation, is making the best quality audio that we can, and I am sure you are with us on this path.
All we care about, in the final summation, is making the best quality audio that we can, and I am sure you are with us on this path.
This kind of casual copying of IP bothers me, even though I know the intent was/is positive. Out of context w/o rest of article, also.
I am in the process of posting the full article on my own web site, and will provide a link here when that is complete.
Walt Jung
Hi John,
I'm sure I was not the first to achieve 100 V/us in a power amp (although my real number was 300 V/us in a 50-watt amp).
However, I'm curious, what power amps did you have in mind that achieved over 100 V/us over 35 years ago?
Cheers,
Bob
A number of pro amps tried high slew rates, only to find out with long cable runs from multiple amplifiers in the same conduit there were problems. So they reduced the slew rates to allow a full power bandwidth of 20k. But their front end designs and the whole amplifier was based on massive feedback.
For that massive feedback it was difficult to make it ultimately stable, I would guess. Anyway, power bandwidth of 20k is too low.
The link to the full "Taming Ampzilla" article is here:
the "By Request" page of WaltsBlog
This is the blog portion of my website, which allows users to comment, etc.
Walt Jung
the "By Request" page of WaltsBlog
This is the blog portion of my website, which allows users to comment, etc.
Walt Jung
Pavel,
On a recent stadium project we aimed for 8K upper frequency response. One fellow complained that was not enough, he wanted at least 14K. I ran a few numbers and mentioned to get from 8 to 12K would require a power increase of 630,000 times the existing design. Anyone care to guess what throw distance would require that?
ES
Pavel,
The pro amplifier guys often don't understand the difference. In the early days of solid state amplifiers they became very popular in recording studios.
They also contributed to the development of solid state design. They still make bold new designs.
Their gear is usually better than most consumer mass-fi. But you are correct by the limited view here they are not in consideration for state of the art.
On the same project I am giving them an option to use a less well known brand of amplifier that will cost them $100,000.00 less. Not because the amplifiers cost less, but because they are more reliable! I expect to save even more than that on warranty costs!
Thanks for the link to KV2 audio, PMA It is always interesting to note the progress in pro audio. I started in pro audio, back in 1968, and kept up with it until 1995, or so when the Grateful Dead dissolved as a group. I hope that you are connected with these people, somehow, in order to keep them from being 'compromised'.
Hitachi s 1978 amps did achieve at least 125V/uS using the famed
2SK135/2SJ50 pairs , with very good stability.
Models range from HMA7500 , 7500MK2 and 8500 , the latter
did use a non switching circuitry.
Second that..
For numerous reasons , complementary differentials is the best topology..
Last edited:
Before things get out of hand in the 'slew rate race', understanding the fundamental limitations of the output devices is important here.
The first popular complementary devices had an Ft of 4MHz, and 80V breakdown. However, higher voltage devices came out, apparently with a similar die, but with 140V breakdown, and an Ft about 2MHz. Some sort of necessary tradeoff.
In 1975, I compared both with a new amp project for Gale Electronics. I found with the SAME circuit, that I could only get 50V/us from the higher voltage parts, yet 100V/us with the lower voltage parts. This would mean that the Ampzilla and its relatives could only achieve 50V/us due to some inherent limitation of the output transistor, itself. Yes, 1/2 the slew rate and twice the distortion with the higher voltage devices.
It should be noted that the JC-3 and the Lohstroh amps used higher Ft devices than Ampzilla, and this, along with the input topology modifications, gave us 100V/us.
However, about 1978, or even a few years sooner, complementary Mos power fets came on the scene, from NEC, Sony, and Hitachi etc. They were much faster than the typical, or even high grade but low voltage, American or European power transistors then available and allowed the circuit techniques we pioneered in the early 1970's to be made even faster. So, of course, greater than 100V/us was possible, and we all gave it a go. Then, very high Ft complementary devices appeared from Japan, about 1980. I took one of my classical designs from 1975 and added the high speed bipolar transistors for an amplifier design for Brian Cheney of Itone audio. It was a balanced bridge, but the output slew rate went from 50V/us /side (100V/us total) to well over 500V/us/side or over 1000V/us total. We used it for several years at CES shows.
However, on reconsideration, I realized this was a race to 'nowhere' and I began to look, like Nelson Pass before me, to find more effective ways of improving amp quality.
One way, as shown by Parasound, before I actually did it, was to eliminate the OUTPUT COIL. This would require more internal compensation for stability, but it gave a better tradeoff in sound qualtiy vs slew rate. That is why we have scaled back to around 100V/us these days. Enough is enough, all else being equal. Around 100V/us is a good minimum slew rate, even for a fairly powerful amplifier, as stated by Walt Jung, Otala, and others.
The first popular complementary devices had an Ft of 4MHz, and 80V breakdown. However, higher voltage devices came out, apparently with a similar die, but with 140V breakdown, and an Ft about 2MHz. Some sort of necessary tradeoff.
In 1975, I compared both with a new amp project for Gale Electronics. I found with the SAME circuit, that I could only get 50V/us from the higher voltage parts, yet 100V/us with the lower voltage parts. This would mean that the Ampzilla and its relatives could only achieve 50V/us due to some inherent limitation of the output transistor, itself. Yes, 1/2 the slew rate and twice the distortion with the higher voltage devices.
It should be noted that the JC-3 and the Lohstroh amps used higher Ft devices than Ampzilla, and this, along with the input topology modifications, gave us 100V/us.
However, about 1978, or even a few years sooner, complementary Mos power fets came on the scene, from NEC, Sony, and Hitachi etc. They were much faster than the typical, or even high grade but low voltage, American or European power transistors then available and allowed the circuit techniques we pioneered in the early 1970's to be made even faster. So, of course, greater than 100V/us was possible, and we all gave it a go. Then, very high Ft complementary devices appeared from Japan, about 1980. I took one of my classical designs from 1975 and added the high speed bipolar transistors for an amplifier design for Brian Cheney of Itone audio. It was a balanced bridge, but the output slew rate went from 50V/us /side (100V/us total) to well over 500V/us/side or over 1000V/us total. We used it for several years at CES shows.
However, on reconsideration, I realized this was a race to 'nowhere' and I began to look, like Nelson Pass before me, to find more effective ways of improving amp quality.
One way, as shown by Parasound, before I actually did it, was to eliminate the OUTPUT COIL. This would require more internal compensation for stability, but it gave a better tradeoff in sound qualtiy vs slew rate. That is why we have scaled back to around 100V/us these days. Enough is enough, all else being equal. Around 100V/us is a good minimum slew rate, even for a fairly powerful amplifier, as stated by Walt Jung, Otala, and others.
- Status
- Not open for further replies.