I agree Scott. The only parallel with Benveniste I can see here is that both sets of claims are a red rag to debunkers
This is my 2'nd try to talk about something that might show what Chas and I are all about. This is the technical comparison between two Parasound amps, both designed with me involved in the heart of the circuit. One was a relative failure, and one was a success. What were the differences and could they have been avoided?
The HCA-3500 was introduced about 10 years ago, maybe longer, and met its specs which were pretty good. It was fast, over 100V/us, lowish distortion, reasonable class A region, and great 'bang for the buck'. We hoped for a 'B' designation, much like the HCA-2200 got previously.
'Stereophile' reviewed it, but they did NOT give it a recommendation. Why, were they mad at us? Well, I finally got my own personal unit and I loaned it to Brian Cheney, my associate, to use at a CES. He took it home, listened to it, and rejected it outright!
Wow, and I had just put it through a whole array of measurements and it measured just fine. Perfect? Of course not, but very competitive.
I then sent the unit to my business partner at the time, Bob Crump. He listened to it, and said that it was awful, too! NOW WHAT? In desperation, I looked for oversights that may have been made. First, I found that the Taiwanese had changed the main feedback resistor to a tiny little thing, that spec'd well, but was a completely different brand. We replaced it with a Resista. Then Bob noticed that it had lots of caps in the power supply with no apparent purpose, so we removed those and put our Rel Caps in, instead, then Bob noticed that the wiring and the connectors were really marginal, so we replaced them. Finally Bob found that one of the housekeeping transformers was 'out of phase' or leaking ground current, excessively. By now, this amp came to life. It started to sound pretty good! Later, we removed the inrush limiter and used a circuit breaker as an input switch, and high speed diodes instead of typical diodes. Mostly we REMOVED parts or replaced them with NON-MAGNETIC equivalents, rather than change the circuit, and it became so successful, that we took it to CES the very next year, with some interest shown by others because of our demo. We then decided to make a limited production of this design by purchasing stock HCA3500's and rebuilding them. By this time the HCA-3500 was already being remaindered out at $1500. We could buy them up, add value, and sell them for $5000, with happy customers as well. (more later)
The HCA-3500 was introduced about 10 years ago, maybe longer, and met its specs which were pretty good. It was fast, over 100V/us, lowish distortion, reasonable class A region, and great 'bang for the buck'. We hoped for a 'B' designation, much like the HCA-2200 got previously.
'Stereophile' reviewed it, but they did NOT give it a recommendation. Why, were they mad at us? Well, I finally got my own personal unit and I loaned it to Brian Cheney, my associate, to use at a CES. He took it home, listened to it, and rejected it outright!
Wow, and I had just put it through a whole array of measurements and it measured just fine. Perfect? Of course not, but very competitive.
I then sent the unit to my business partner at the time, Bob Crump. He listened to it, and said that it was awful, too! NOW WHAT? In desperation, I looked for oversights that may have been made. First, I found that the Taiwanese had changed the main feedback resistor to a tiny little thing, that spec'd well, but was a completely different brand. We replaced it with a Resista. Then Bob noticed that it had lots of caps in the power supply with no apparent purpose, so we removed those and put our Rel Caps in, instead, then Bob noticed that the wiring and the connectors were really marginal, so we replaced them. Finally Bob found that one of the housekeeping transformers was 'out of phase' or leaking ground current, excessively. By now, this amp came to life. It started to sound pretty good! Later, we removed the inrush limiter and used a circuit breaker as an input switch, and high speed diodes instead of typical diodes. Mostly we REMOVED parts or replaced them with NON-MAGNETIC equivalents, rather than change the circuit, and it became so successful, that we took it to CES the very next year, with some interest shown by others because of our demo. We then decided to make a limited production of this design by purchasing stock HCA3500's and rebuilding them. By this time the HCA-3500 was already being remaindered out at $1500. We could buy them up, add value, and sell them for $5000, with happy customers as well. (more later)
Excellent stuff John, those kinds of stories are what makes this thread a must-read. Keep it coming...
What is important is that it is not obvious what will work. We have to be: Openminded, enterprising, and take chances in order to be successful with audio design.
The alternative is like: "The operation was successful, but the patient died." That is not the way we find works as well.
What Charles was hoping to communicate with his comparison of programs was basically that we DO listen, and even listen pretty darn well, but we are open to NOT completely understanding WHY something works, and we will go with it anyway. It just has to work 'for us' and this usually means that it will work for the vast majority of our customers as well. If it did not work that way, we would have been out of this business, long ago.
In fact, I was told, later, that I almost lost my position with Parasound, BECAUSE of the subjective failures of the HCA3500 power amp and the PLD2000 preamp. I had left too much to the Taiwanese to decide.
I knew at that time that I HAD to push for subtle changes that Bob and I came up with, that the Taiwanese did NOT necessarily want to go through with. We also had to tighten up with the design and not allow unannounced changes without going through us first. It is still a battle, with both the dealers, and the foundry. It is NOT easy to find complementary jfets, for example, in the quantities that we use, but they are still necessary. And so it goes.
The alternative is like: "The operation was successful, but the patient died." That is not the way we find works as well.
What Charles was hoping to communicate with his comparison of programs was basically that we DO listen, and even listen pretty darn well, but we are open to NOT completely understanding WHY something works, and we will go with it anyway. It just has to work 'for us' and this usually means that it will work for the vast majority of our customers as well. If it did not work that way, we would have been out of this business, long ago.
In fact, I was told, later, that I almost lost my position with Parasound, BECAUSE of the subjective failures of the HCA3500 power amp and the PLD2000 preamp. I had left too much to the Taiwanese to decide.
I knew at that time that I HAD to push for subtle changes that Bob and I came up with, that the Taiwanese did NOT necessarily want to go through with. We also had to tighten up with the design and not allow unannounced changes without going through us first. It is still a battle, with both the dealers, and the foundry. It is NOT easy to find complementary jfets, for example, in the quantities that we use, but they are still necessary. And so it goes.
So what I'm getting from this John is that we'd all do very well to listen to you when you tell us what works and what doesn't, because you're experienced at listening and the proof is you're still in business. But we'd also do well take any explanations with a pinch of salt because the 'whys' don't bother you half as much as the 'whats'.
As an example you say that wide bandwidth sounds better. I get that. But when you go on to say why it sounds better we'd better start thinking for ourselves because that's not where you're at.
I'm not quite sure what conclusion we were meant to draw from JC's post #12203. My conclusion is that a poor implementation of a good design can sound bad, for perfectly rational reasons. I thought we all knew that. A second conclusion is that some people can hear things that others can't; not new either. A third conclusion is that something can measure well and sound bad: either the wrong things were being measured, or people prefer certain distortions - again nothing new.
Or were we supposed to conclude that we cannot measure the right things, and that components with identical electrical/mechanical characteristics in all respects can still sound different? I suppose then we just have to trust the gurus.
Or were we supposed to conclude that we cannot measure the right things, and that components with identical electrical/mechanical characteristics in all respects can still sound different? I suppose then we just have to trust the gurus.
Basically CH posted his experience because of Scott telling that "they" would not do listening tests without peeking.
Obviously CH did a controlled listening test without peeking (only the coder did know which one was the correct answer- if i got it right) and normally Scott should admit that his assertion was wrong.
Obviously CH did a controlled listening test without peeking (only the coder did know which one was the correct answer- if i got it right) and normally Scott should admit that his assertion was wrong.
Afair the "grand debate" started because some people claimed that two amplifiers were sounding different, although both passed a standard set of measurements and should - related to the known thresholds of hearing abilities - be indistinguishable (under normal working conditions) .
So, i think John´s story is related to that situation.
Sometimes it is hard to believe that this grand debate even started, because it was (and is) based on psychoacoustic results, and as controlled listening tests can not _prove_ something, these results are open to changes anytime.
A good example are the ´curves of equal loudness´- there exists a nice paper which lists the various studies on this matter.
Small number of participants, very different conditions and quite different results; first it was Fletcher/Munson, then it was Davidson/Richardson and the new standard is something in between.
Last edited:
Agreed; but that doesn't make him a wizzard does it? Chances are you, or - gasp! - even me would also have heard the difference, if it was audible.
jan didden
Now to continue:
Our success with the modified HCA3500 and the new CTC Blowtorch together gave us a real boost in confidence.
However, Parasound had independently decided to make a MONO version of the HCA3500 and call it the JC-1. They had a 'mock up' of the unit at the CES in another building. I didn't even know about it, until I happened by and saw it. It was attractive, even beautiful, and it had a BIG heatsink, but while these cosmetic changes were just great, what about the insides?
This is when I wrote a strong E-mail to my boss at Parasound urging him to let Bob Crump, Carl Thompson, and me, (CTC) at the insides, and have veto power over the way it was put together and with what parts and materials.
I DELIBERATELY kept the same basic design as the HCA3500, just like Parasound had initially decided to do, anyway, BUT changes would be made. This would make the design more expensive, but we were designing for an 'A' rated product this time. Sort of a 'go for broke' approach.
I decided that I could move the rated power to 400W into 8 ohms, and 800W into 4 ohms, as we had the heatsink for it. We DOUBLED the power supply values, and used high speed power diodes throughout the unit. That's twice the transformer, and twice the power caps, as we had the room, inside.
We insisted that Carl Thompson, who initially laid out the Vendetta Research products, CTC and many more efforts, do the internal layout. This was a big effort, but certainly worth it. Bob Crump was brought in to choose the input, output connectors, wiring, tranformer phasing, etc, all his 'tricks of the trade' and have veto power over the product's component selection. Trust me, it was NOT easy to convince the people overseas, who thought they could do it as well and cheaper.
Finally, I improved the basic design to be more efficient with the existing power supplies and more effortless in its operation by using 9 pairs of output devices, instead of 8, and other subtle improvements, but I kept to the basic topology of my previous designs, including the HCA-3500, so we could make a comparison in future, to a known design. I also wanted to vindicate my original design, as well, as I knew it was better than it had been shown to be, previously. There are other design approaches, such as balanced bridged, that I have made amps with, but this one was to be more conventional.
Well, the amp came out, and we got our 'A' rating. Bob had to battle 'parts substitution' by using a magnet to ferret out the 'phony connectors' (try it yourself, sometime) and we retired from making the modified HCA-3500's, because we had a new amp to show at CES. This is more or less, the end of the story.
Where is the engineering? Where is the SPICE analysis? Where is the 'improvement'?
Yet, the difference between the stock HCA-3500 and the JC-1 is obvious to reviewers and customers, alike. (at least to those who believe in and care about these sorts of things)
Our success with the modified HCA3500 and the new CTC Blowtorch together gave us a real boost in confidence.
However, Parasound had independently decided to make a MONO version of the HCA3500 and call it the JC-1. They had a 'mock up' of the unit at the CES in another building. I didn't even know about it, until I happened by and saw it. It was attractive, even beautiful, and it had a BIG heatsink, but while these cosmetic changes were just great, what about the insides?
This is when I wrote a strong E-mail to my boss at Parasound urging him to let Bob Crump, Carl Thompson, and me, (CTC) at the insides, and have veto power over the way it was put together and with what parts and materials.
I DELIBERATELY kept the same basic design as the HCA3500, just like Parasound had initially decided to do, anyway, BUT changes would be made. This would make the design more expensive, but we were designing for an 'A' rated product this time. Sort of a 'go for broke' approach.
I decided that I could move the rated power to 400W into 8 ohms, and 800W into 4 ohms, as we had the heatsink for it. We DOUBLED the power supply values, and used high speed power diodes throughout the unit. That's twice the transformer, and twice the power caps, as we had the room, inside.
We insisted that Carl Thompson, who initially laid out the Vendetta Research products, CTC and many more efforts, do the internal layout. This was a big effort, but certainly worth it. Bob Crump was brought in to choose the input, output connectors, wiring, tranformer phasing, etc, all his 'tricks of the trade' and have veto power over the product's component selection. Trust me, it was NOT easy to convince the people overseas, who thought they could do it as well and cheaper.
Finally, I improved the basic design to be more efficient with the existing power supplies and more effortless in its operation by using 9 pairs of output devices, instead of 8, and other subtle improvements, but I kept to the basic topology of my previous designs, including the HCA-3500, so we could make a comparison in future, to a known design. I also wanted to vindicate my original design, as well, as I knew it was better than it had been shown to be, previously. There are other design approaches, such as balanced bridged, that I have made amps with, but this one was to be more conventional.
Well, the amp came out, and we got our 'A' rating. Bob had to battle 'parts substitution' by using a magnet to ferret out the 'phony connectors' (try it yourself, sometime) and we retired from making the modified HCA-3500's, because we had a new amp to show at CES. This is more or less, the end of the story.
Where is the engineering? Where is the SPICE analysis? Where is the 'improvement'?
Yet, the difference between the stock HCA-3500 and the JC-1 is obvious to reviewers and customers, alike. (at least to those who believe in and care about these sorts of things)
John, you have just described some engineering! Some people might want to argue with some of your choices, and say that some of them make no difference, but that is a separate issue. You seem to be tilting at windmills. By all means attack a caricature if that makes you happy, but the main effect is that people will take less notice of you not more. This would be a pity.
Sorry but no evidence was offered that there was no chance of bias. Self refereed DBT's are a problem.
I've had an F-bomb sent my way because I, my boss at the time, and my wife even could not hear that a change in lead frame material made some IC sound like ****. The customer socketed a preamp off the assembly line and sent 4 each of ones they had sorted. We spent a lot of our time and energy gratis and in good faith tried to find/hear a difference. Nothing, nada and all we got is that we can't hear a kuffing thing.
Why is it there is never anyone in the room who doesn't hear it?
Once when I was on top of the mountain of garbage, there was an issue with a sibilance or resonant hangover in a particular loudspeaker. They measured it and their software at that time showed more than a 360 degree phase shift at an upper mid range frequency, which they read as a software error. Two of the engineers could hear the problem and three could not. So did I imagine the problem or was it there?
Jan,
I do not ascribe magical powers, the point was the issue was there and some can hear it and some can't.
I had a salesman onetime stop in with a sample of a new stadium sized loudspeaker. He was anxious for me to hear it as he "knew" it was priced much lower than the similar product we were using and that it also sounded much better. He had wonderful data to show that. So we hooked it up and listened to it. I was not impressed. So much to the salesman's surprise I brought the original product out of the back room. (72" x 48" x 48" in size, the back room is quite large.) I plugged it in to the same amplifier and played the same music through it. I have never seen that salesman again.
- Status
- Not open for further replies.