Actually, you guys are nuts. Even SY owns the same piece of test equipment that we used for this test. It is not inaccurate, the ONLY problem is that the distortion in question was NOT noted numerically on the graph. It COULD have been, but it did not make sense at first, and apparently ignored for that reason. I REALLY doubt that the aberrant distortion is 2f2-11f1. It would be in the right place, BUT what about the magnitudes of the distortion byproducts in question? Can they be THAT HIGH, compared to lower orders of IM distortion? 2f2-8f1 is still off by 40 Hz or so.
Last edited:
John,
It is a photocopy of a plotter output, it may even be a photocopy of a photocopy, you cannot assume you can measure anything accurately from that plot. You can see the photocopy blots all over it, they are not accurate, just try putting some graph paper in an old one, you will see how bad they can be.
As others have said earlier in this thread, if you can generate something that isnt a mixer product or a stray from somewhere else in the chain then you would get so much adulation even you wouldnt know what to do with it.
Wrinkle
It is a photocopy of a plotter output, it may even be a photocopy of a photocopy, you cannot assume you can measure anything accurately from that plot. You can see the photocopy blots all over it, they are not accurate, just try putting some graph paper in an old one, you will see how bad they can be.
As others have said earlier in this thread, if you can generate something that isnt a mixer product or a stray from somewhere else in the chain then you would get so much adulation even you wouldnt know what to do with it.
Wrinkle
The graph IS monotonic, the analyzer accurate to 1Hz, and the graphics plotter is a professional one, not some 'modern' printer. It is true that small deviations preclude super accurate estimates, but 40 Hz is just too much. What amazes me is what we get from our printers on our home computers. Now that, is suspect. Please remember that I ran many, many original plots, and they were accurate enough. The VTT government research lab in Finland has to keep its equipment running properly and calibrated. Audio is just a part time sideline for them. This test equipment with plotter cost 1/2 years salary of an American engineer, at the time. HP don't come cheap.
Please my friends, let's come down to earth again. We can measure - yes the maeurement methods are getting better / or not - and we can calculate, but in the end the question is when do I stop to design and bring the amplifier into a product... and much important is its sound.
Last edited:
The whole point of this exercise is to find new and better ways of designing amplifiers and preamplifiers. Anyone can just look up in a textbook how to make an amplifier. Personally, I don't understand why most here would even want to. I have been making amplifiers for more than 40 years. I am ALWAYS trying to improve them. However, what is 'improvement' is debatable. For example, I can use a lot of global negative feedback to get very low frequency and midrange distortion. Is that all there is to it? OR I can use two stage compensation and get very low distortion up to 100KHz or so. Is that better? Besides avoiding gross distortion, what is important in an amplifier design? Perhaps, we should make amplifiers that are very very fast. But, I do not know why this would be an improvement. I am not getting anywhere with this, so I will simply state what works for ME. I make the best circuitry that I can, use as little global negative feedback as possible, and make sure the slew rate is 100V/us or so, even in my preamp designs. This works for me. Your opinion can differ.
I think that it is important to continue researching for better ways to design audio equipment. In my opinion, no audio design is successful, until it has been listened to and accepted by serious audio listeners, over a period of time. This is harder than it looks.
Let me give an example:
About 10 years ago, the HCA-3500 was introduced to the marketplace by Parasound. I, of course, had specified the general topology, and tested the final prototype, before release.
Much of the support circuitry, was done by engineers in Taiwan, and was similar to previous designs, such as the HCA-2200. It was made to be a more powerful successor to the HCA-2200, with a slight change in drive circuitry. However, at this time, I was seemingly being distanced by Taiwan, in many design decisions. The passive parts and the layout were sourced almost entirely by the Taiwanese engineers, using common stocks, for the most part, apparently made extremely cost effective by bulk purchasing, where many parts, such as connectors, wiring, resistors, caps, etc are used in a great many designs, even those made for competitive companies. At this time, I was rather tired of 'riding herd' on the Taiwanese engineers, and let them do, what they wanted to do. After all, maybe I was TOO careful with device selection, etc. The product line was established, we had gotten a good review on the HCA-2200, what could go wrong? I also BENCH TESTED the amp extensively to make sure it was stable, low distortion, especially higher order odd harmonic distortion, and fast.
Well the amp went for review, and I got my own personal HCA-3500, that I also tested to my satisfaction that it was biased right and operating properly. I gave it to a friend to use at CES. So what happened in the review and with my friend? More later.
Let me give an example:
About 10 years ago, the HCA-3500 was introduced to the marketplace by Parasound. I, of course, had specified the general topology, and tested the final prototype, before release.
Much of the support circuitry, was done by engineers in Taiwan, and was similar to previous designs, such as the HCA-2200. It was made to be a more powerful successor to the HCA-2200, with a slight change in drive circuitry. However, at this time, I was seemingly being distanced by Taiwan, in many design decisions. The passive parts and the layout were sourced almost entirely by the Taiwanese engineers, using common stocks, for the most part, apparently made extremely cost effective by bulk purchasing, where many parts, such as connectors, wiring, resistors, caps, etc are used in a great many designs, even those made for competitive companies. At this time, I was rather tired of 'riding herd' on the Taiwanese engineers, and let them do, what they wanted to do. After all, maybe I was TOO careful with device selection, etc. The product line was established, we had gotten a good review on the HCA-2200, what could go wrong? I also BENCH TESTED the amp extensively to make sure it was stable, low distortion, especially higher order odd harmonic distortion, and fast.
Well the amp went for review, and I got my own personal HCA-3500, that I also tested to my satisfaction that it was biased right and operating properly. I gave it to a friend to use at CES. So what happened in the review and with my friend? More later.
Well, first, my friend, someone who I just talked to yesterday, took the HCA-3500, freshly off my test bench and LISTENED to it on the speakers that he was taking to CES. He didn't tell me, at first, but he REJECTED the sound of the amp, and refused to take it to CES, even as a back-up. Wow, measurements good, sound 'bad'. Well, the amp then went to by business partner, Bob Crump. He listened to the amp, compared it to a very fine amp that he owned, made by our 'BEAR' who contributes here, and rejected the HCA-3500 as well. What to do? It measured well, was very cost effective, yet I could NOT get my friends and associates to 'like' the sound of the amp. Should I have found new friends and associates, or subjected them to an ABX double-blind test to PROVE that they could not hear any difference? Some here would. More later.
At the same time, more or less, 'Stereophile' got the HCA-3500 in for review. They reviewed it, measured it, and it came out well, BUT they did NOT rate it. The expected rating of an amp of this price and design sophistication should get a 'B'. The HCA-2200 had a B rating, why not the HCA-3500? In any case, the amp was not very successful in the audio marketplace, even before the recession reduced sales, and it was quietly remaindered off with a discount mail order dealer. Well, now what? More of the same? I think not.
So, my business partner, Bob Crump, and I decided to modify the HCA-3500, and using Bob's exceptional ears and hi fi playback system, as our reference comparison.
First, I changed the feedback resistor. This resistor was an 'unknown' brand with great specs. However, I knew from experience, that this feedback resistor could be important, because the amp's overall response is completely dependent on it, and it is subject to voltage extremes of perhaps +/- 100V, on occasion. In any case, I had initially established that the Holco 1/2W resistor, worked for the HCA-2200. This was our first 'mod' and it made an audible difference.
Second, Bob Crump found that transformer leakage could be important sonically. Why, I cannot easily explain, but Bob found that the 'housekeeping' transformer to be leaking AC voltage on the chassis and that reversing its windings, helped.
Bob Crump then said at this time that just THESE 2 mods made the sound of the amp acceptable to most listeners, but what if we did more?
So we did, replacing the power supply bypass caps, rectifier diodes, input and output connectors, internal wiring, etc. Guess what? We had an amp that sounded darn good, and almost everybody who listened to it wanted one. We decided to go into business modifying these amps. We used them with the CTC Blowtorch at upcoming CES shows and got great feedback. We both started to use the amp, ourselves in our own hi fi system. More later.
So, my business partner, Bob Crump, and I decided to modify the HCA-3500, and using Bob's exceptional ears and hi fi playback system, as our reference comparison.
First, I changed the feedback resistor. This resistor was an 'unknown' brand with great specs. However, I knew from experience, that this feedback resistor could be important, because the amp's overall response is completely dependent on it, and it is subject to voltage extremes of perhaps +/- 100V, on occasion. In any case, I had initially established that the Holco 1/2W resistor, worked for the HCA-2200. This was our first 'mod' and it made an audible difference.
Second, Bob Crump found that transformer leakage could be important sonically. Why, I cannot easily explain, but Bob found that the 'housekeeping' transformer to be leaking AC voltage on the chassis and that reversing its windings, helped.
Bob Crump then said at this time that just THESE 2 mods made the sound of the amp acceptable to most listeners, but what if we did more?
So we did, replacing the power supply bypass caps, rectifier diodes, input and output connectors, internal wiring, etc. Guess what? We had an amp that sounded darn good, and almost everybody who listened to it wanted one. We decided to go into business modifying these amps. We used them with the CTC Blowtorch at upcoming CES shows and got great feedback. We both started to use the amp, ourselves in our own hi fi system. More later.
All fluff that is irrelevant once the original has been put through a cheap photocopier, it may well still be monotonic, but it is no longer linear between paper and the real frequency that was measured. Now if it was a modern scan of the original graph paper then we might have more of a chance of believing the results.
Wrinkle
Wrinkle, 'believing the results' depends on the position of the observer/evaluator. For example, is every UFO sighting just the planet Venus? Some might think so.
In the case of 2f2-8f1, it should be 4.56 KHz. That is what I compute, please check my math, if you want to. I would appreciate it. Now the UNKNOWN FREQUENCY appears between 4.08 KHz and 5.46 KHz. Do you see it? Where do you think it is positioned on the graph. This is simple linear interpolation. Now, do you think the unknown frequency is in the middle, between the two KNOWN frequencies? If so, it would 4.77 KHz. Does this make sense? IF, you, like me, see it over closer to 2/3 of the way between the two points, then it might be as much as 5 KHz. Does this make sense to you? MAYBE, 2f2-11f1 would fit. Is this what we are seeing? NOW, resolution is EVERYTHING and we now don't have it to this extent. IF, 2f2-8f1 was on the graph, it would be less than 1/3 across the graph. Do you think we work with 'fun house' optics for this to be possible?
In the case of 2f2-8f1, it should be 4.56 KHz. That is what I compute, please check my math, if you want to. I would appreciate it. Now the UNKNOWN FREQUENCY appears between 4.08 KHz and 5.46 KHz. Do you see it? Where do you think it is positioned on the graph. This is simple linear interpolation. Now, do you think the unknown frequency is in the middle, between the two KNOWN frequencies? If so, it would 4.77 KHz. Does this make sense? IF, you, like me, see it over closer to 2/3 of the way between the two points, then it might be as much as 5 KHz. Does this make sense to you? MAYBE, 2f2-11f1 would fit. Is this what we are seeing? NOW, resolution is EVERYTHING and we now don't have it to this extent. IF, 2f2-8f1 was on the graph, it would be less than 1/3 across the graph. Do you think we work with 'fun house' optics for this to be possible?
Please do not mix apples and pears. The issue was not audible difference or 'A' or 'B' rating, but evidence of PIM. I am saying that the old plots bring no reliable evidence. Why do not we have any current PIM evidence with better methods available. Please do not confuse the scientific issue by Stereophile stories, too cheap.
Thanks for the kind words John and I agree with you. May I suggest one point more, I trust not only serious audio listeners to rate my designs. I ask serious, very sensitive, music lovers what opinion they have regarding the sound.
Back to the 3500 saga.
Bob Crump and I had a modified HCA3500 that we could use and also sell to customers to go with the CTC Blowtorch. Bob Crump did all the work of modifying the amps.
About this time, Parasound, wanting to make a replacement for the HCA-3500, came up with a new, prettier chassis, with even bigger heat sinks. They wanted to call it the JC-1, even before I knew about this project.
This is where I had to 'draw the line in the sand'. I couldn't just let the Taiwanese engineers do what they wanted and put MY initials on the product. Also, I knew from experience that it would fail in the marketplace. This is when I wrote a serious letter to Parasound to convince them to let me again have control of the essence of the project, including parts selection, layout, as well as circuit topology control. Parasound decided to give in, even though this would put the amp in a different price point, perhaps double or even triple the price of the HCA-3500.
The amp would be mono, giving me at least double the heatsink area. This gave me the ability to have 25W or so Class A standing current. This was a given from the start. Then, in order to add more power, we doubled the power supply from HCA3500. Of course, the Taiwanese thought we were 'nuts' but NOW we could drive 800W into 4 ohms, and even print it on the spec sheet. This made driving very large electrostatics easy enough, and one electrostatic company started to show with us at CES, because of this.
But that was not ALL, or even the most important part.
What really made the JC-1 'successful' was the printed circuit layout by Carl Thompson, who also did the Vendetta Research, Lineage, and CTC Blowtorch circuit boards with me, and parts selection by Bob Crump, based on our experience with modifying the 3500. There were lots of small, but important details, including what kind of bypass caps, resistors, connectors, interior wiring. One IMPORTANT detail was MAGNETIC material, like steel. It had to be minimized. Cheap connectors attracted a magnet, big-time. Successful connectors did not. Another detail was substituting high speed diodes and rectifiers, EVERYWHERE in the power supplies. Snubbers might have worked, but we went to the cause of the problem. We also relay bypassed the low inrush nonlinear power resistor, after initial turn on. We found this to be audible, also.
In fact, Bob and I found just about everything to be important.
Now this took time and money, but it was worth it.
When the JC-1 was reviewed, it got an A rating from 'Stereophile' and while being very much more expensive than the HCA-3500, it was denoted as being a 'bargain' in the hi end audiophile marketplace. We had succeeded in our task: Parasound had a 'flagship' that gave it notice from the high end people, electrostatic speaker companies got a very powerful amp, that sounded pretty good, and Bob and I got the satisfaction that we could make a successful amp, without having to modify it extensively later.
Please note: It is still an HCA-3500, underneath. The essence of the original design is still there, YET it is now sonically acceptable to the serious audiophile community. That was my experiment, from the first.
Now, did this make it a 'perfect' product? No, it still had less heatsink area than Nelson Pass, and more negative feedback than Ayre.
The reality came clear, when someone we knew, sold his JC-1's to buy an Ayre power amp. The JC-1 was still a compromise, a good compromise, but still not the essence of perfection. That is still out there, and why I continue to 'preach' that on this thread. I personally am happy with the CTC Blowtorch preamp. That is when 100% control of EVERYTHING was given to Bob Crump, Carl Thompson, and me. The CTC of the company. I doubt that I will ever replace the CTC with anything 'better'. It is 'done' so to speak.
Not so with the JC-1 and its family of products, I will continue to 'improve' it where I can.
Bob Crump and I had a modified HCA3500 that we could use and also sell to customers to go with the CTC Blowtorch. Bob Crump did all the work of modifying the amps.
About this time, Parasound, wanting to make a replacement for the HCA-3500, came up with a new, prettier chassis, with even bigger heat sinks. They wanted to call it the JC-1, even before I knew about this project.
This is where I had to 'draw the line in the sand'. I couldn't just let the Taiwanese engineers do what they wanted and put MY initials on the product. Also, I knew from experience that it would fail in the marketplace. This is when I wrote a serious letter to Parasound to convince them to let me again have control of the essence of the project, including parts selection, layout, as well as circuit topology control. Parasound decided to give in, even though this would put the amp in a different price point, perhaps double or even triple the price of the HCA-3500.
The amp would be mono, giving me at least double the heatsink area. This gave me the ability to have 25W or so Class A standing current. This was a given from the start. Then, in order to add more power, we doubled the power supply from HCA3500. Of course, the Taiwanese thought we were 'nuts' but NOW we could drive 800W into 4 ohms, and even print it on the spec sheet. This made driving very large electrostatics easy enough, and one electrostatic company started to show with us at CES, because of this.
But that was not ALL, or even the most important part.
What really made the JC-1 'successful' was the printed circuit layout by Carl Thompson, who also did the Vendetta Research, Lineage, and CTC Blowtorch circuit boards with me, and parts selection by Bob Crump, based on our experience with modifying the 3500. There were lots of small, but important details, including what kind of bypass caps, resistors, connectors, interior wiring. One IMPORTANT detail was MAGNETIC material, like steel. It had to be minimized. Cheap connectors attracted a magnet, big-time. Successful connectors did not. Another detail was substituting high speed diodes and rectifiers, EVERYWHERE in the power supplies. Snubbers might have worked, but we went to the cause of the problem. We also relay bypassed the low inrush nonlinear power resistor, after initial turn on. We found this to be audible, also.
In fact, Bob and I found just about everything to be important.
Now this took time and money, but it was worth it.
When the JC-1 was reviewed, it got an A rating from 'Stereophile' and while being very much more expensive than the HCA-3500, it was denoted as being a 'bargain' in the hi end audiophile marketplace. We had succeeded in our task: Parasound had a 'flagship' that gave it notice from the high end people, electrostatic speaker companies got a very powerful amp, that sounded pretty good, and Bob and I got the satisfaction that we could make a successful amp, without having to modify it extensively later.
Please note: It is still an HCA-3500, underneath. The essence of the original design is still there, YET it is now sonically acceptable to the serious audiophile community. That was my experiment, from the first.
Now, did this make it a 'perfect' product? No, it still had less heatsink area than Nelson Pass, and more negative feedback than Ayre.
The reality came clear, when someone we knew, sold his JC-1's to buy an Ayre power amp. The JC-1 was still a compromise, a good compromise, but still not the essence of perfection. That is still out there, and why I continue to 'preach' that on this thread. I personally am happy with the CTC Blowtorch preamp. That is when 100% control of EVERYTHING was given to Bob Crump, Carl Thompson, and me. The CTC of the company. I doubt that I will ever replace the CTC with anything 'better'. It is 'done' so to speak.
Not so with the JC-1 and its family of products, I will continue to 'improve' it where I can.
Last edited:
WOW, John, this is interesting stuff!
In general, I'm puzzled by the lack of good relations between subjective and objective measurements in audio, in particular in amps.
So I wonder, could any of the improvements you made to the HCA3500 be verified by any "objective" measurements?
No direct measurements show the difference. So, according to many here, it must be my imagination, or false hopes. It might be some sort of subjective 'magic' where my 'good intentions' get put into each piece of equipment, or my 'pathological' negative opinions changed people's subjective impression of the 3500. [Wow, am I powerful! ;-) ]
Of course, I am just kidding, but I found decades ago, that it is more a matter of PHYSICS rather than engineering, that makes these subtle differences. You know, what is really happening at a very low level, when passing audio information. I wish that I could prove my assertions to everyone's satisfaction, but I can't. I only recommend that people listen, and if they hear differences, they probably are real.
Of course, I am just kidding, but I found decades ago, that it is more a matter of PHYSICS rather than engineering, that makes these subtle differences. You know, what is really happening at a very low level, when passing audio information. I wish that I could prove my assertions to everyone's satisfaction, but I can't. I only recommend that people listen, and if they hear differences, they probably are real.
Ones in a closed feedback loop, differences in components if good but not exotic get swapped easily. For example a good but not expensive polypropylen capacitor does not have enough distortion ( even measured on it´s own outside a feedback loop ) to explain why it should be subjectively different from say an expensive teflon cap. The University in Salford in England undertook research why this can be and they found that there are differences in the mechanical behaviour of cappacitors under music exitation.
- Status
- Not open for further replies.