"Some people" will always think all sorts of things.
Also, using just one speaker, or even one not too popular TYPE of speakers, to prove a point is hardly convincing. Although I agree that slew rates as such, isolated on their own, don't mean much really.
By the same token, I could now call my own speakers (local manufacture) the standard, even if they are an exceptionally easy load to drive, nominal impedance 8 Ohms, minimum 6.5 Ohms, worst case phase shift 20...20,000 Hz -25 degrees at around 200 Hz, efficency 92 dB/2.83V/1m. They would in fact level the playing field to unrealistic comfort levels, enabling tha humblest of amplifiers to achieve its best. But this is unrealistic because in the other room, I have a pair of AR 94 speakers, which are a bitch to drive and make quite a few amps huff and puff for breath.
Square waves are an invaluable tool in designing, as they represent a multitude of even simpler sine waves, and no matter what anyone says, they do let us see at least approximately what happens with our amp as it is subjected to various drive regimes. Not ideal, but certainly very useful.
Agree on the square wave deal ... 
I was reading an article last week , about a test done by one of the B&K designers , according to him , this setup have the ability to test amplifiers from .0001 watt to 1 watt and by his evaluation, amplifiers that show .00001% thd @ 100 watts and sound subjectively average, exhibit 1 % thd or higher @ .01 watt . The belief is this new testing procedure , has the best correlation between measurements and good sound ..
I was reading an article last week , about a test done by one of the B&K designers , according to him , this setup have the ability to test amplifiers from .0001 watt to 1 watt and by his evaluation, amplifiers that show .00001% thd @ 100 watts and sound subjectively average, exhibit 1 % thd or higher @ .01 watt . The belief is this new testing procedure , has the best correlation between measurements and good sound ..
Agree on the square wave deal ...
I was reading an article last week , about a test done by one of the B&K designers , according to him , this setup have the ability to test amplifiers from .0001 watt to 1 watt and by his evaluation, amplifiers that show .00001% thd @ 100 watts and sound subjectively average, exhibit 1 % thd or higher @ .01 watt . The belief is this new testing procedure , has the best correlation between measurements and good sound ..
Well, not a moment too soon. It was about time somebody did something really worthwhile in that arena, even if it is no longer so "in" as it used to be.
On the subject of voltage slew rate, overall I feel it's too exaggerated. In my view, as long as your amp can slew 1V/usec per every peak volt of output, I believe you're home and dry regarding SID, TIM and such like. Using FET inputs, this is very likely to be more to much more anyway, without any special effort.
Looking back, it seems to me that the problems encountered many years ago were in fact the result of irresponsible compensation - the easiest thing to do is to throw in a compensation cap until it stops oscillating or exhibiting signs of instability, certainly far easier and cheaper than further researching and refining your circuit. But, as ever, after the battle, everybody's a general.
In my view, and I am totally with Thorsten on this, the trick is to judge a proper balance between local and global NFB, and this will vary from case to case, from one topology to another, from one set of requirements to another. Like T., I also happen to think it's better to clean up house locally, leaving the global NFB to be more or less like icing on a cake.
Where we differ (I think?) is in how much should be left over for global. T. mentioned that 12 dB was ideal; I don't know how he got that particular number, but I am not convinced that's always ideal. Yes, my own HK 680 integrated uses exactly that much, but overall, I tend more towards 20 dB as being ideal, 12 dB still has reminants of that "not quite completed" design I dislike. But at 20 dB, of the several amps I have heard, not one had even a hint of it.
Since amps using that much (or little) global NFB don't exactly grow on trees, my sample was small, and all of it was hand built by enthusiasts. Not a terribly reliable sample, I'm afraid.
BTW, my simulator says my voltage slew rate is around 390 V/usec (1.3 mA and 3.3 pF). Thus, even in the very unlikely case of having to double the value of the compensation caps, I still have more than I could ever possibly use. Open loop full power bandwidth is just over 80 kHz at -3 dB point and stable (according to the sim) - into a 4 Ohm load.
It's a habit I picked up from the Germans decades ago. Make it work well into 4 Ohms and it'll work even better into 8 Ohms.
Another habit I picked up also from the Germans, specifically Studer/Revox (although a nominally Swiss company, their actual home audio manufacturing was just across the border, in Loeffingen, Germany) is to get the amp to work to say 300 kHz, and then put in a first order filter at the input, limiting this to say 200 kHz. Seems prudent to me.
Jan will probably know better, but if memory serves, Philips used that gig as well.
Looking back, it seems to me that the problems encountered many years ago were in fact the result of irresponsible compensation - the easiest thing to do is to throw in a compensation cap until it stops oscillating or exhibiting signs of instability, certainly far easier and cheaper than further researching and refining your circuit. But, as ever, after the battle, everybody's a general.
In my view, and I am totally with Thorsten on this, the trick is to judge a proper balance between local and global NFB, and this will vary from case to case, from one topology to another, from one set of requirements to another. Like T., I also happen to think it's better to clean up house locally, leaving the global NFB to be more or less like icing on a cake.
Where we differ (I think?) is in how much should be left over for global. T. mentioned that 12 dB was ideal; I don't know how he got that particular number, but I am not convinced that's always ideal. Yes, my own HK 680 integrated uses exactly that much, but overall, I tend more towards 20 dB as being ideal, 12 dB still has reminants of that "not quite completed" design I dislike. But at 20 dB, of the several amps I have heard, not one had even a hint of it.
Since amps using that much (or little) global NFB don't exactly grow on trees, my sample was small, and all of it was hand built by enthusiasts. Not a terribly reliable sample, I'm afraid.
BTW, my simulator says my voltage slew rate is around 390 V/usec (1.3 mA and 3.3 pF). Thus, even in the very unlikely case of having to double the value of the compensation caps, I still have more than I could ever possibly use. Open loop full power bandwidth is just over 80 kHz at -3 dB point and stable (according to the sim) - into a 4 Ohm load.
It's a habit I picked up from the Germans decades ago. Make it work well into 4 Ohms and it'll work even better into 8 Ohms.
Another habit I picked up also from the Germans, specifically Studer/Revox (although a nominally Swiss company, their actual home audio manufacturing was just across the border, in Loeffingen, Germany) is to get the amp to work to say 300 kHz, and then put in a first order filter at the input, limiting this to say 200 kHz. Seems prudent to me.
Jan will probably know better, but if memory serves, Philips used that gig as well.
Don't know about Philips in detail, but they were very conservative, more like "it measures great therefor it MUST sound great". They were at the time an engineers' outfit, not very good at marketing. Their video casette system was arguably better but lost from VHS marketing power.
A few years ago I spoke with the guy who invented the moving magnet cartridge, but Philips thought that piezo cartridges were all that would ever be needed. This guy went to work for Bauer at Shure and the rest, as they say, is history.
Sorry for rambling on like that...
jan
I couldn't agree more with you.
I used to say that somebody should do Philips a favor and send shock troop into the marketing department, with orders to use extreme prejudice - shoot anything that moves.
I had their fantastic N4520 open reel deck, all 28 kilos of it - it was a truly outstanding product, which never really made it, like their Black Tulip range, because I was sure they had like the worst marketing in the world.
Having worked with Philips on their PC products for five years, I know that their technical and commercial staff was one of the best in the business - but again, their marketing slowed things down like I couldn't believe. Their SE European HQ was in Vienna, relatively near, so I saw a lot of them. To this day, I think very fondly of them.
Agree on the square wave deal ...
I was reading an article last week , about a test done by one of the B&K designers , according to him , this setup have the ability to test amplifiers from .0001 watt to 1 watt and by his evaluation, amplifiers that show .00001% thd @ 100 watts and sound subjectively average, exhibit 1 % thd or higher @ .01 watt . The belief is this new testing procedure , has the best correlation between measurements and good sound ..
.0001W. Yea right. In the noise. Is .01 watt valid? ONE watt would be. Going for a standard 10X resolution, start at .1W. I think it was mentioned a few thousand posts ago of an Italian mag that tested at many powers across the spectrum and produced a nice 3-d waterfall plot. That would be insightful information. It still leaves out the difference in amps in proportion of high to low order harmonics and if that changes across the spectrum or with power. We need at least a four dimensional plot. Frequency, power, distortion, ratio high to low. We need to be able to slice the plot on any axis. If you have the data, Matlab can do the plot.
VHS won for one very simple reason: They were the first consumer unit with 3 hours and could record a full American football game. End of competition.
THE PROBLEMS WITH MEASUREMENTS
All audiophiles should at least once bring an accurate sound-level pressure gauge into their listening rooms. This is especially important for those music lovers who listen mainly to acoustical music; classical, jazz, folk etc., like me.
After an evening of listening to different music and observing the meter, most audiophiles are very surprised at how low the dB readings are on the gauge; usually between 60 to 85dB. It will on occasion go lower, and only for extremely brief instances will it ever rise above 90db. Much more important than the pure numbers are the ultimate implications of all this. What do these surprisingly low dB numbers really mean for the on-going scientific and practical attempts of measuring a component's ability to reproduce music?
They are devastating, and here is why.
Let's start with a speaker of fairly high sensitivity, which is the trend these days and also what I recommend above. Let's say the sensitivity is 90dB/1 watt. This means that at an 80dB loudness level, this speaker is receiving a total of 1/10th of 1 watt of power from the amplifier. This is the point where most audio magazines stop measuring, but this is the exact point where they really should begin "fine measuring", because the 80dB is only the peak/accumulated loudness at that moment. All the real, fine musical details and information; the harmonics, decays, sense of space, dynamic inflections etc. are still 20 to 30dB (or more) below the 80dB peak.
What does this all mean?
The truly unique and distinguishing musical information is being reproduced with only 1/10,000 of a watt or even less power!
At a softer 60dB loudness level, which is not that unusual, the power level of even 1 Millionth of one watt becomes important! Which audio "tech/guru" or scientist measures what is happening in an amplifier from 100th to 1,000,000th of one watt?
The Answer: Not even one.
This same basic principle holds true for measuring preamplifiers, speakers and everything else. (It is also a very plausible explanation why some components appear to sound better after some "break-in".) Until it is possible to scientifically measure low-level musical information, we will have to trust our imperfect and unscientific ears and let them choose what component has the most "magic".
This inevitably brings us to the next logical and unavoidable subject.
-Salvatore
All audiophiles should at least once bring an accurate sound-level pressure gauge into their listening rooms. This is especially important for those music lovers who listen mainly to acoustical music; classical, jazz, folk etc., like me.
After an evening of listening to different music and observing the meter, most audiophiles are very surprised at how low the dB readings are on the gauge; usually between 60 to 85dB. It will on occasion go lower, and only for extremely brief instances will it ever rise above 90db. Much more important than the pure numbers are the ultimate implications of all this. What do these surprisingly low dB numbers really mean for the on-going scientific and practical attempts of measuring a component's ability to reproduce music?
They are devastating, and here is why.
Let's start with a speaker of fairly high sensitivity, which is the trend these days and also what I recommend above. Let's say the sensitivity is 90dB/1 watt. This means that at an 80dB loudness level, this speaker is receiving a total of 1/10th of 1 watt of power from the amplifier. This is the point where most audio magazines stop measuring, but this is the exact point where they really should begin "fine measuring", because the 80dB is only the peak/accumulated loudness at that moment. All the real, fine musical details and information; the harmonics, decays, sense of space, dynamic inflections etc. are still 20 to 30dB (or more) below the 80dB peak.
What does this all mean?
The truly unique and distinguishing musical information is being reproduced with only 1/10,000 of a watt or even less power!
At a softer 60dB loudness level, which is not that unusual, the power level of even 1 Millionth of one watt becomes important! Which audio "tech/guru" or scientist measures what is happening in an amplifier from 100th to 1,000,000th of one watt?
The Answer: Not even one.
This same basic principle holds true for measuring preamplifiers, speakers and everything else. (It is also a very plausible explanation why some components appear to sound better after some "break-in".) Until it is possible to scientifically measure low-level musical information, we will have to trust our imperfect and unscientific ears and let them choose what component has the most "magic".
This inevitably brings us to the next logical and unavoidable subject.
-Salvatore
MEASURING ULTRA LOW-LEVEL MUSICAL AMPLIFIER SIGNALS
"My name is Steve Keiser, the 'K' of B&K components, and presently design engineer with Luminance Audio. I have developed a measurement system which is able to quantify distortion levels at micropower quantities down to 1/1000th of a watt. These measurement techniques are unprecedented, and have revealed a number of revelations of amplifier distortion characteristics, at micropower levels, which are in direct opposition with traditional and scientific assumptions up unto this point.
The emphasis of my work is to definitively quantify low level signal linearity measurements of power amplifiers, and attempt to correlate these measurements with subjective listening results, as well as establishing the significance of low level distortion. Conventional test equipment generally does not resolve meaningful distortion measurements below 100mw, since the measurements become predominated by noise.
I have modified a spectrum analysis software program, which uses time-averaging to effectively cancel out noise products, leaving an identifiable signal and its related harmonics. This time-averaging approach is to identify extremely weak signals from spacecraft, amid a very high noise ambient background. Using this method, I am able to resolve a standardized total harmonic distortion measurement down to 1/1000 watts, and an approximate measurement down to 1/500,000th of a watt.
My measurement results oppose common engineering supposition, in that it is commonly believed that very low signal linearity is essentially 'virtually perfect', and that only high level signal linearity is a relevant parameter. My measurements indicate exactly the opposite is true of this common held assumption, particularly for amplifiers employing solid state devices.
To give you an example: the Halcro DM58 amplifier measures .007%THD at 2 watts, whereas at 1/1000th watts, THD measures 8.9%! By contrast a Wavac SH833 measures .57%THD at 2 watts and 1.6%THD at 1/1000th watts. The tube Wavac exhibits significantly lower THD at low signal levels by orders of magnitude than the Halcro. I have measured numerous amplifiers, both solid state and tube, which I will provide to you as well as any other information you may want pertaining to this work.
Correlative Listening Tests
...I will now provide a supplementary addition regarding correlative listening tests with a panel of 5 evaluators. Some of the tests were conducted using a blind A/B comparison method, in order to satisfy militant objectivists. The two amplifiers compared were a Wavac SH833 and Halcro DM58. In 10 trials, with listeners blindfolded, every listener on the panel preferred the Wavac by several orders of magnitude, with commentary such as describing the Halcro as sounding: transistory, thin, harsh, dark, closed in spatially, as well as having poor sound floor resolution.
Every listener described the Halcro as being 'unlistenable', while the Wavac enjoyed universal positive accolades. These listening tests correlate exactly with the comparative measurements I outlined.
My research into this characteristic is currently ongoing, and I would enjoy sharing my results with audio enthusiasts, editors, and designers. If this correlation between measurement technique and listening impressions holds up consistently, it could mean a whole new approach to audio engineering could be opened up resulting in significant breakthroughs in design performance. The main idea is to let our ears continue to be the final arbitrator of component performance and allow objective science to enhance our subjective appreciation."
- Steve Keiser
"My name is Steve Keiser, the 'K' of B&K components, and presently design engineer with Luminance Audio. I have developed a measurement system which is able to quantify distortion levels at micropower quantities down to 1/1000th of a watt. These measurement techniques are unprecedented, and have revealed a number of revelations of amplifier distortion characteristics, at micropower levels, which are in direct opposition with traditional and scientific assumptions up unto this point.
The emphasis of my work is to definitively quantify low level signal linearity measurements of power amplifiers, and attempt to correlate these measurements with subjective listening results, as well as establishing the significance of low level distortion. Conventional test equipment generally does not resolve meaningful distortion measurements below 100mw, since the measurements become predominated by noise.
I have modified a spectrum analysis software program, which uses time-averaging to effectively cancel out noise products, leaving an identifiable signal and its related harmonics. This time-averaging approach is to identify extremely weak signals from spacecraft, amid a very high noise ambient background. Using this method, I am able to resolve a standardized total harmonic distortion measurement down to 1/1000 watts, and an approximate measurement down to 1/500,000th of a watt.
My measurement results oppose common engineering supposition, in that it is commonly believed that very low signal linearity is essentially 'virtually perfect', and that only high level signal linearity is a relevant parameter. My measurements indicate exactly the opposite is true of this common held assumption, particularly for amplifiers employing solid state devices.
To give you an example: the Halcro DM58 amplifier measures .007%THD at 2 watts, whereas at 1/1000th watts, THD measures 8.9%! By contrast a Wavac SH833 measures .57%THD at 2 watts and 1.6%THD at 1/1000th watts. The tube Wavac exhibits significantly lower THD at low signal levels by orders of magnitude than the Halcro. I have measured numerous amplifiers, both solid state and tube, which I will provide to you as well as any other information you may want pertaining to this work.
Correlative Listening Tests
...I will now provide a supplementary addition regarding correlative listening tests with a panel of 5 evaluators. Some of the tests were conducted using a blind A/B comparison method, in order to satisfy militant objectivists. The two amplifiers compared were a Wavac SH833 and Halcro DM58. In 10 trials, with listeners blindfolded, every listener on the panel preferred the Wavac by several orders of magnitude, with commentary such as describing the Halcro as sounding: transistory, thin, harsh, dark, closed in spatially, as well as having poor sound floor resolution.
Every listener described the Halcro as being 'unlistenable', while the Wavac enjoyed universal positive accolades. These listening tests correlate exactly with the comparative measurements I outlined.
My research into this characteristic is currently ongoing, and I would enjoy sharing my results with audio enthusiasts, editors, and designers. If this correlation between measurement technique and listening impressions holds up consistently, it could mean a whole new approach to audio engineering could be opened up resulting in significant breakthroughs in design performance. The main idea is to let our ears continue to be the final arbitrator of component performance and allow objective science to enhance our subjective appreciation."
- Steve Keiser
Last edited:
Well of course he would have to use signal averaging, and nothing wrong with that per se. But the details of the apparatus should be very interesting. And the spectra of these supposed low-level distortions more interesting still! And what about IM distortion, of far more significance to sound quality than mere harmonic distortion?
We also get into territory here where the presence of some noise may actually enhance low-level detection (so-called stochastic resonance), and as well where some amplifying devices have noise that is signal-modulated. Tubes have been cited as having noise that seems to "float" independently of the signal, and in ways that are somehow easier to separate from the signal.
Brad
We also get into territory here where the presence of some noise may actually enhance low-level detection (so-called stochastic resonance), and as well where some amplifying devices have noise that is signal-modulated. Tubes have been cited as having noise that seems to "float" independently of the signal, and in ways that are somehow easier to separate from the signal.
Brad
Thanks for S.K. comments. I know of this approach, but I am not sure it is as accurate as it is being presented. This short piece does not say anything about the noise floor. Pulling signals from below the floor is quite a trick. Time averaging beats the heck out of autocorrelation, but it is not magic. ( another technical method brought to you by a large US government agency whom wishes to remain low profile).
Exactly as he states, unprecedented, so I think it requires a bit more explanation. I would love to read more and see where this led.
How old is this posting? I find the links on the WEB to their products dead and their site a search redirect.
Exactly as he states, unprecedented, so I think it requires a bit more explanation. I would love to read more and see where this led.
How old is this posting? I find the links on the WEB to their products dead and their site a search redirect.
B& K is out of Business , he is with Luminance Audio, send him an email there I guess ...
http://www.youtube.com/watch?v=C7TQJT4HlL0
i sent Steve an email, he may respond ....
- Status
- Not open for further replies.