I think John is right in this. I remember that when ML (the man) left ML (the company), there was a legal action where ML (the man) tried to stop HK from using the ML name, but lost.
jan didden
I would again like to remind people of my audio design philosophy, where it came from, and what it is today. Just to keep the record straight.
I didn't start listening to hi fi quality sound. Throughout my first 20 years of life, I was exposed to radio and phono consoles that most everyone had. When I started making electronics kits, they were test equipment, mostly. There were only 2 people, I personally knew, in those 20 years that had good sound equipment, one a relative, who made his own diy, and a high school acquaintance, who built some Heathkit equipment. I could hear the difference, even then, more extended highs and lower overall distortion, but I was not that interested in hi fi, itself. I was more interested in playing and collecting guitars, both classical and electric. Here, I found that I personally had an aptitude to discriminate between musical instruments, and even musical instrument amplifiers, apparently more than the average person, and this kept me busy for years, trying to get the best sounding classical guitar(s) for example.
Later, when I found that my playing ability was not up to professional par, and didn't seem to be going anywhere, I shifted my attention to hi fi reproduction. As I was an upper division student studying engineering-physics at the time, my funds were limited to used audio equipment, cast offs of others, like Dyna tube equipment, and an Acoustic Research AR-1. I put together a pretty good MONO hi fi system that I used for years.
This is when I decided to concentrate on being an audio design engineer. It didn't involve making military weapons or rockets, which was the primary area of electronics at the time that held easy and well paying jobs. People actually tried to discourage from doing so, stating that I was being impractical.
Well, enough of that early history of 50 years ago. When I finished college, I had a BA in Physics (not very useful in industry) with a minor in electronic engineering (much more useful). I was first hired as a designer or jr. Engineer, because I knew enough electronics and had the math background to learn engineering on the job at a professional level. I often think that it is the math that is the most important thing that I learned in college. It helps give a deeper understanding of the basic processes that make things work, such as loudspeakers.
In any case, hi fi design was put aside, while I learned to do electronic design and testing. Just to make it clear, I had assembly techs who worked under me, and I was made responsible for troubleshooting their efforts, and making all the measurements.
In the background, at home, I kept improving my hi fi with better speakers, turntable, phono cartridges, and even a professional Ampex portable 2 ch tape recorder.
Finally, in 1968, I arrived at my goal, audio design, being put on staff in the Audio Division of Ampex. I really thought that I was on top of the world, and the opportunities to learn, were enormous. Huge company library, one of the best audio libraries I have ever seen in my life experience, night classes given by noted professors from nearby universities, and a huge assembly of parts, some considered too exotic for audio use.
In one example, it was easy for me to try to make a complementary differential input amplifier, because I had matched pairs of bipolar transistors available to me, in both NPN and PNP, right off the shelf. Made it easy.
More later.
I didn't start listening to hi fi quality sound. Throughout my first 20 years of life, I was exposed to radio and phono consoles that most everyone had. When I started making electronics kits, they were test equipment, mostly. There were only 2 people, I personally knew, in those 20 years that had good sound equipment, one a relative, who made his own diy, and a high school acquaintance, who built some Heathkit equipment. I could hear the difference, even then, more extended highs and lower overall distortion, but I was not that interested in hi fi, itself. I was more interested in playing and collecting guitars, both classical and electric. Here, I found that I personally had an aptitude to discriminate between musical instruments, and even musical instrument amplifiers, apparently more than the average person, and this kept me busy for years, trying to get the best sounding classical guitar(s) for example.
Later, when I found that my playing ability was not up to professional par, and didn't seem to be going anywhere, I shifted my attention to hi fi reproduction. As I was an upper division student studying engineering-physics at the time, my funds were limited to used audio equipment, cast offs of others, like Dyna tube equipment, and an Acoustic Research AR-1. I put together a pretty good MONO hi fi system that I used for years.
This is when I decided to concentrate on being an audio design engineer. It didn't involve making military weapons or rockets, which was the primary area of electronics at the time that held easy and well paying jobs. People actually tried to discourage from doing so, stating that I was being impractical.
Well, enough of that early history of 50 years ago. When I finished college, I had a BA in Physics (not very useful in industry) with a minor in electronic engineering (much more useful). I was first hired as a designer or jr. Engineer, because I knew enough electronics and had the math background to learn engineering on the job at a professional level. I often think that it is the math that is the most important thing that I learned in college. It helps give a deeper understanding of the basic processes that make things work, such as loudspeakers.
In any case, hi fi design was put aside, while I learned to do electronic design and testing. Just to make it clear, I had assembly techs who worked under me, and I was made responsible for troubleshooting their efforts, and making all the measurements.
In the background, at home, I kept improving my hi fi with better speakers, turntable, phono cartridges, and even a professional Ampex portable 2 ch tape recorder.
Finally, in 1968, I arrived at my goal, audio design, being put on staff in the Audio Division of Ampex. I really thought that I was on top of the world, and the opportunities to learn, were enormous. Huge company library, one of the best audio libraries I have ever seen in my life experience, night classes given by noted professors from nearby universities, and a huge assembly of parts, some considered too exotic for audio use.
In one example, it was easy for me to try to make a complementary differential input amplifier, because I had matched pairs of bipolar transistors available to me, in both NPN and PNP, right off the shelf. Made it easy.
More later.
Last edited:
By 1970, I had left Ampex, first taking a detour to the Research Division. It was hard to leave Ampex, but my engineering duties were more video and motor drive, than audio, and I wanted to get back to audio. This is when I got hired by Alembic, a subsidiary of the Grateful Dead, to make solid state designs to replace their aging and fragile tube equipment, that sounded pretty good. I learned a lot about live sound working with this company. We did live concerts virtually every week, and I had to help. We also owned a 16 channel Ampex audio recorder, that we used for making live and studio recordings.
In any case, we built a multi-channel 'live' mixer to replace the tube (open loop) mixer electronics that we were using, and which worked amazingly well. We built the new mixer with CERAMIC coupling caps, and used the best IC that we could find, the HA990 from Harris Associates, or Radiation Inc, their original name. These devices were made to be radiation resistant, and were very expensive to produce, at the time.
Well, it went on the road, but FAILED the GD listening experience test, and I was let go to find my way in audio, elsewhere. This taught me a BIG lesson. Measurement, especially SMPTE IM measurement, did NOT give me enough information to be sure that the audio that I designed, sounded good enough. The tube electronics that they LIKED (and I did as well) measured FAR WORSE with my distortion measuring equipment. What was going on here? Did I like extra distortion? No, not really. In fact, the extra distortion kind of 'muffled' the sound, BUT the tube equipment did NOT effect the sound quality in the BAD way that solid state seemed to do. WHY? More later.
In any case, we built a multi-channel 'live' mixer to replace the tube (open loop) mixer electronics that we were using, and which worked amazingly well. We built the new mixer with CERAMIC coupling caps, and used the best IC that we could find, the HA990 from Harris Associates, or Radiation Inc, their original name. These devices were made to be radiation resistant, and were very expensive to produce, at the time.
Well, it went on the road, but FAILED the GD listening experience test, and I was let go to find my way in audio, elsewhere. This taught me a BIG lesson. Measurement, especially SMPTE IM measurement, did NOT give me enough information to be sure that the audio that I designed, sounded good enough. The tube electronics that they LIKED (and I did as well) measured FAR WORSE with my distortion measuring equipment. What was going on here? Did I like extra distortion? No, not really. In fact, the extra distortion kind of 'muffled' the sound, BUT the tube equipment did NOT effect the sound quality in the BAD way that solid state seemed to do. WHY? More later.
No - I read your opinions on this thread with interest. Of course, opinions are just that, so I'll compare them with my own experience and that acts as a kind of filter.
Your earlier comment stated that to you, it seemed I was not aware of this. That's the comment that I was referring to when I mentioned a distorted perception. Don't read things into what I write, take the words as they stand, at face value. What you think is 'implied' is where the perceptual distortion is occurring - the 'slander' is entirely in your mind. I am indeed aware that Mark Levinson the man was not involved with the company of the same name at that time.
Clearer now?
Are you speaking of the man here, or the product? The man is apparently the Russian president, Dmitry Medvedev.
Dmitry Medvedev's stereo system has Russia's bloggers buzzing | World news | The Guardian
Last edited:
This is where I made MY personal breakthrough in what is important in audio design, back in the early '70's. I was fortunate that someone else, experienced in their own musical production, gave me a serious appraisal of what I thought was important in audio design, and I failed. Most engineers think that designing audio equipment is easy, all you have to do is to make it measure well with the distortion measurements of your choice, and you are OK. This is not so, and I have proven this with dozens of designs from microphone preamps to large power amps, including studio boards and master analog tape recorders. In fact, I found that I could be a consultant for other companies, and improve their subjective designs, without lifting a finger, just giving them good design advice. However, I ultimately found that IF I really want to make a world class audio design, perhaps enjoyed by someone of the stature of Medvedev, I have to control virtually everything, myself, as virtually every design decision can make a significant difference, even if some here do not believe it. It would be the same if I designed the fastest and best driving automobile, camera, or even a TV. Refinement can be discerned by those who count, usually serious audio enthusiasts, who have a great deal of experience in listening to and comparing audio equipment. More later.
What I essentially found, was that with the best audio designs that the quality of the contacts and connections, along with a dozen other factors, competed with measured distortion and circuit topology to make a successful, world class, design. It appeared to be as much physics as design engineering. This is a major departure from much of electronic engineering, in my experience. Again, this seems to show that the human ear is an amazing distortion analyzer.
This is where it gets difficult to convey info and advice that is not seriously attacked by many others. I have 'grounded' this recent section of this thread with two parallel paths, the 'objective' test history route and the 'subjective' "What did I do wrong?" route, covering the time up to about 1972, a long time ago. Now, continuing with the 'subjective' route and combining it with the 'objective' route at the time, created a breakthrough for making designs and even products that 'satisfied' the demands of both professional musicians and hi fi enthusiasts.
This involved combining the most elegant circuit topologies available at the time, the 'exotic' ideas of Matti Otala, and the finest construction and parts quality, to make gain modules that worked for virtually every situation in audio processing. Of course, we needed a 'control' to see if we were not kidding ourselves. This was provided by the Mark Levinson LMP-2 preamp, that contained IC based modules, that could be directly compared by designers and independent reviewers alike. This preamp preceded the JC-2 preamp, and was as beautifully built (if not better) as the JC-2, using the same circuit board material, construction techniques and even the same RIAA equalization.
And guess what? These were the VERY SAME IC's that I had selected for the 'live board' that we built in 1970, for the GD. These IC's, however, were SELECTED for lowest distortion and then packaged in a hybrid module by Dick Burwin, who made virtually all the decisions about the passive parts, wiring, switches, etc. for the LMP-2, and used as well in the JC-2.
The Grateful Dead also tried one of Mark Levinson's portable mixers, made with these IC based modules, and generally approved its performance. At least, Mark used Polycarbonate and not CERAMIC coupling caps, as well as the top pick of the best IC's then available. This obviously made a difference with the band, in their evaluation.
So now we had evaluated the difference between a conventionally made studio board, with carbon resistors, ceramic caps, and unpicked IC's to a studio board with metal film resistors, polycarbonate caps, and lower distortion picks of the same IC's.
We also had the new discrete modules, designed by me, based on Otala's recommendations, and built by Mark Levinson, put into the same sized hybrid module. The GD not only liked, but became enthusiastic about the discrete modules. I had FINALLY satisfied the GD with an acceptable gain amplifier to replace their tube equipment. Mark Levinson, at this point, realized that these discrete modules really did sound better than the IC based modules, and we went on to build the Levinson JC-2 preamp that put both Mark Levinson and me, 'on the map', so to speak. More later.
This involved combining the most elegant circuit topologies available at the time, the 'exotic' ideas of Matti Otala, and the finest construction and parts quality, to make gain modules that worked for virtually every situation in audio processing. Of course, we needed a 'control' to see if we were not kidding ourselves. This was provided by the Mark Levinson LMP-2 preamp, that contained IC based modules, that could be directly compared by designers and independent reviewers alike. This preamp preceded the JC-2 preamp, and was as beautifully built (if not better) as the JC-2, using the same circuit board material, construction techniques and even the same RIAA equalization.
And guess what? These were the VERY SAME IC's that I had selected for the 'live board' that we built in 1970, for the GD. These IC's, however, were SELECTED for lowest distortion and then packaged in a hybrid module by Dick Burwin, who made virtually all the decisions about the passive parts, wiring, switches, etc. for the LMP-2, and used as well in the JC-2.
The Grateful Dead also tried one of Mark Levinson's portable mixers, made with these IC based modules, and generally approved its performance. At least, Mark used Polycarbonate and not CERAMIC coupling caps, as well as the top pick of the best IC's then available. This obviously made a difference with the band, in their evaluation.
So now we had evaluated the difference between a conventionally made studio board, with carbon resistors, ceramic caps, and unpicked IC's to a studio board with metal film resistors, polycarbonate caps, and lower distortion picks of the same IC's.
We also had the new discrete modules, designed by me, based on Otala's recommendations, and built by Mark Levinson, put into the same sized hybrid module. The GD not only liked, but became enthusiastic about the discrete modules. I had FINALLY satisfied the GD with an acceptable gain amplifier to replace their tube equipment. Mark Levinson, at this point, realized that these discrete modules really did sound better than the IC based modules, and we went on to build the Levinson JC-2 preamp that put both Mark Levinson and me, 'on the map', so to speak. More later.
We finally had a solid state module that could compete with tube quality and sound. It did NOT do everything right, but at least, it was acceptable to discerning listeners. It should be noted that this new design did not deviate much, in principle, from tube designs, in that it had low feedback, and high open loop bandwidth, Class A operation, and a very smooth transfer function, even without global negative feedback.
The IC competition, on the other hand, had thermal feedback, crossover distortion, and often an almost unrecognizable open loop transfer function, that could ONLY be put right by lots of global negative feedback. The IC's had low slew rates as well, compared to the 100V/us of the GD module.
Now, we must admit, that over the decades, IC designers have fixed many of these drawbacks, but did they go far enough? More later.
The IC competition, on the other hand, had thermal feedback, crossover distortion, and often an almost unrecognizable open loop transfer function, that could ONLY be put right by lots of global negative feedback. The IC's had low slew rates as well, compared to the 100V/us of the GD module.
Now, we must admit, that over the decades, IC designers have fixed many of these drawbacks, but did they go far enough? More later.
Thanks for the Hirata reference
Hey John,
Long time no speak. I found this thread looking for something else, and it's taken hours to read through this fascinating discussion.
I have been looking for the Hirata papers on and off for many years. The main reason I couldn't find it was a memory failure -- for some reason I thought it was by HIRAGA! Of course, Jean Hiraga has done some interesting stuff, but not this....
When I first encountered the idea, it was one of those concepts that was clearly right. Didn't one of the articles show that piano notes exhibited a similar kind of asymmetry? I never could understand why it didn't become a core tool for design and test.
Peter Tait
Hey John,
Long time no speak. I found this thread looking for something else, and it's taken hours to read through this fascinating discussion.
I have been looking for the Hirata papers on and off for many years. The main reason I couldn't find it was a memory failure -- for some reason I thought it was by HIRAGA! Of course, Jean Hiraga has done some interesting stuff, but not this....
When I first encountered the idea, it was one of those concepts that was clearly right. Didn't one of the articles show that piano notes exhibited a similar kind of asymmetry? I never could understand why it didn't become a core tool for design and test.
Peter Tait
Recognizing that may be a little side tour, but in the phono preamp discussion a question puzzled me: how to protect a multi-k$ cartridge in case the preamp input is broken? given the very small size of those coils I would think they can only handle a few mW of heat before they give....
To sense a (DC) overcurrent at the input would probably ruin all low noise efforts, I guess, so it might have to be done elsewhere in the preamp input stage - again, hopefully not ruining the noise performance. And where would the protection circuit have to sit - could it be a relay that just short-circuits the input to make sure the current doesnt flow through the cartridge?
To sense a (DC) overcurrent at the input would probably ruin all low noise efforts, I guess, so it might have to be done elsewhere in the preamp input stage - again, hopefully not ruining the noise performance. And where would the protection circuit have to sit - could it be a relay that just short-circuits the input to make sure the current doesnt flow through the cartridge?
Now, that we had a representative model for a discrete amplifier design that 'worked' and the consumer and audio reviewer response was enthusiastic, who could ask for anything more? That is what I thought for a couple of years, mainly 1974-1976, at least for consumer audio design. In that time, I acquired the very first Electrocompaniet prototype, and compared it directly with the latest Marantz and SAE power amps at the time, and found that the Lohstroh-Otala design easily beat the other competition. In fact, I used this little amp, off and on, until 1991, when it was destroyed in a firestorm. It was good enough to drive a pair of WATT-Puppies, the best speaker that I had at the time. Again, low feedback, linear circuitry, high open loop bandwidth, 100V/us slew rate.
Then, I started finding little problems in the JC-2 preamp. First, the volume controls that we used had 'bad spots' that when the wiper of the pot contacted them, created LISTENABLE and measurable distortion. To make a long story, short, Mark ultimately got Penny and Giles to make an audio rotary pot, that we often see in hi end equipment, even today.
Then, in 1976, at Frank Van Alstine's audio establishment, I heard the same vinyl recording, essentially the same speakers, as I was using at the time, IMAGE between the speakers, better than I heard it previously through the JC-2 preamp. This was traced to an asymmetrical crosstalk, due to the layout of the JC-2 motherboard, for the most part. I would never have believed that this small amount of crosstalk asymmetry could make any difference, but it did. This actually crippled the JC-2 from being virtually perfect, and I went on, over the next years to build a successor. More later
Then, I started finding little problems in the JC-2 preamp. First, the volume controls that we used had 'bad spots' that when the wiper of the pot contacted them, created LISTENABLE and measurable distortion. To make a long story, short, Mark ultimately got Penny and Giles to make an audio rotary pot, that we often see in hi end equipment, even today.
Then, in 1976, at Frank Van Alstine's audio establishment, I heard the same vinyl recording, essentially the same speakers, as I was using at the time, IMAGE between the speakers, better than I heard it previously through the JC-2 preamp. This was traced to an asymmetrical crosstalk, due to the layout of the JC-2 motherboard, for the most part. I would never have believed that this small amount of crosstalk asymmetry could make any difference, but it did. This actually crippled the JC-2 from being virtually perfect, and I went on, over the next years to build a successor. More later
As you may note, I was slowly assembling list of do's and don'ts in order to make a 'successful' audio design. We first found: high speed, class A, j-fets where practical, minimum stage count, low distortion passive parts, high channel separation, high open loop bandwidth, and a number of other things, including connector design and plating.
Some of these 'do's', were debated strongly then, and some important ones, may be challenged even today. The 'proof' is in the listening, much like the 'proof' is in the driving, with an automobile, etc.
The designers and sponsors of the audio product are secondary to the overall response of the serious listening public, and this may take several years before real success or failure can be ascertained. In hindsight, look at the audio classics of the past: The Marantz 9 power amp, the Marantz 10 tuner, both vacuum tubes, the Audio Research SP3 and the Levinson JC-2, preamps, etc.
We now get into 'subjective' territory and ANY opinion can be attacked by others as being biased or self serving. However, I use this acquired history of products to show that really well made products will be regarded as 'special', long after they have been discontinued and replaced with other designs.
For example, how many of you know much about the Levinson LNP-2 or the Levinson JC-1? These were successful products in their time, but they lacked 'something' and are generally forgotten in audio history. What they 'lacked' is what we have learned to design around and out, if possible.
Now, given significant improvements in IC design, we are under continuous pressure to forgo discrete circuit design and chose an IC for the task. However, so far, I have been only partially successful in doing this. Why? I think that the REDUCTION in open loop bandwidth compared to historically SUCCESSFUL audio designs is a KEY reason.
Except for VIDEO IC op amps, usually the open loop bandwidth is 100Hz or less.
While this can be explained away by both math and measurement, I think this is the KEY to continued successful audio circuit design. There has been research in this area over the past 30 years, but it has been ignored for the most part, except by hi end designers, like Charles Hansen of Ayre, and myself, just to name two. Tube designers essentially get high open loop by default, not necessarily their own choosing, but solid state designers have to do it on purpose, either by running open loop, where the effective open loop bandwidth is the working bandwidth of the device, OR minimum feedback and high speed design, to push the open loop bandwidth as high as possible, at least above 1KHz, and with 10K-20K preferred. This is because of PIM distortion, an FM based aberration that is 'invisible' to our normal test methods at the moment. This is still where we will find new research in audio design to be extremely helpful along with Hirata and TIM-SID based measurements.
Some of these 'do's', were debated strongly then, and some important ones, may be challenged even today. The 'proof' is in the listening, much like the 'proof' is in the driving, with an automobile, etc.
The designers and sponsors of the audio product are secondary to the overall response of the serious listening public, and this may take several years before real success or failure can be ascertained. In hindsight, look at the audio classics of the past: The Marantz 9 power amp, the Marantz 10 tuner, both vacuum tubes, the Audio Research SP3 and the Levinson JC-2, preamps, etc.
We now get into 'subjective' territory and ANY opinion can be attacked by others as being biased or self serving. However, I use this acquired history of products to show that really well made products will be regarded as 'special', long after they have been discontinued and replaced with other designs.
For example, how many of you know much about the Levinson LNP-2 or the Levinson JC-1? These were successful products in their time, but they lacked 'something' and are generally forgotten in audio history. What they 'lacked' is what we have learned to design around and out, if possible.
Now, given significant improvements in IC design, we are under continuous pressure to forgo discrete circuit design and chose an IC for the task. However, so far, I have been only partially successful in doing this. Why? I think that the REDUCTION in open loop bandwidth compared to historically SUCCESSFUL audio designs is a KEY reason.
Except for VIDEO IC op amps, usually the open loop bandwidth is 100Hz or less.
While this can be explained away by both math and measurement, I think this is the KEY to continued successful audio circuit design. There has been research in this area over the past 30 years, but it has been ignored for the most part, except by hi end designers, like Charles Hansen of Ayre, and myself, just to name two. Tube designers essentially get high open loop by default, not necessarily their own choosing, but solid state designers have to do it on purpose, either by running open loop, where the effective open loop bandwidth is the working bandwidth of the device, OR minimum feedback and high speed design, to push the open loop bandwidth as high as possible, at least above 1KHz, and with 10K-20K preferred. This is because of PIM distortion, an FM based aberration that is 'invisible' to our normal test methods at the moment. This is still where we will find new research in audio design to be extremely helpful along with Hirata and TIM-SID based measurements.
Last edited:
- Status
- Not open for further replies.