Haha, you crack me up!
So JC and Richard Marsh and Charles Hansen etc have no clue about electronics and have no strong fundament?
Great sense of humor you've got
jan
Well said, Vlad.
Now to attempt to answer your question about Ft in output devices:
In the beginning, at least in my design lifetime, the practical Ft of complementary output devices started at about 2MHz. It is, of course, possible to build audio amplifiers with a device with an Ft of 2MHz, and most of the early power amps, like the Ampzilla, were made with devices like these. These transistors did not have many other 'quality' factors going for them, either, except for a higher second breakdown capability than the lower safe area 4MHz devices that were popular also. In fact, the 4MHz devices, from the same manufacturer, showed twice the practical linearity as a follower, and twice the potential slew rate, as well.
So if you wanted to make a complementary output high power amplifier, you would have to use 2MHz parts, (Like the Gale amp, in my case), and if you wanted more inherent linearity, you would use the 4MHz parts, but be stuck with low working voltages, (like the JC-3, in my case), in fact, anything beyond +/- 25V was getting into dangerous territory.
Now here come the Japanese devices. Apparently, these devices were thought out independently from the American designs, and they came to show a number of attributes that are not easily denied. They have good safe area, allowing power supply voltages up to +/- 90V, they are much higher in Ft, and they have very good beta linearity. They are just BETTER overall.
Now what is the advantage with much higher Ft? I can only point out two things: The effective slew rate can be increased dramatically, and the output coil can be removed if you throw away some of the effective slew rate potential and compensate the amp to be unconditionally stable with a difficult load.
While 2MHz devices are still made, they have few real advantages. One potential advantage is more peak current, up to 50A. We compensate for this by just using more output devices in parallel. We get the peak current, a better heat spread, to improve cooling, and we still get the linear beta and high speed. I hope this helps.
Now to attempt to answer your question about Ft in output devices:
In the beginning, at least in my design lifetime, the practical Ft of complementary output devices started at about 2MHz. It is, of course, possible to build audio amplifiers with a device with an Ft of 2MHz, and most of the early power amps, like the Ampzilla, were made with devices like these. These transistors did not have many other 'quality' factors going for them, either, except for a higher second breakdown capability than the lower safe area 4MHz devices that were popular also. In fact, the 4MHz devices, from the same manufacturer, showed twice the practical linearity as a follower, and twice the potential slew rate, as well.
So if you wanted to make a complementary output high power amplifier, you would have to use 2MHz parts, (Like the Gale amp, in my case), and if you wanted more inherent linearity, you would use the 4MHz parts, but be stuck with low working voltages, (like the JC-3, in my case), in fact, anything beyond +/- 25V was getting into dangerous territory.
Now here come the Japanese devices. Apparently, these devices were thought out independently from the American designs, and they came to show a number of attributes that are not easily denied. They have good safe area, allowing power supply voltages up to +/- 90V, they are much higher in Ft, and they have very good beta linearity. They are just BETTER overall.
Now what is the advantage with much higher Ft? I can only point out two things: The effective slew rate can be increased dramatically, and the output coil can be removed if you throw away some of the effective slew rate potential and compensate the amp to be unconditionally stable with a difficult load.
While 2MHz devices are still made, they have few real advantages. One potential advantage is more peak current, up to 50A. We compensate for this by just using more output devices in parallel. We get the peak current, a better heat spread, to improve cooling, and we still get the linear beta and high speed. I hope this helps.
The question is not that ambiguous as one might think. The guys who achieved something in high-end design, they do not devote all their efforts to discussion of FFT THD and IMD, and they do not claim that they have mastered TOP AUDIO electronics. Those who 100% believe in FFT THD IMD fundament, they are not high-end guys.
Last edited:
You appear to be saying that the best buildings are constructed on wobbly foundations. Or you could be erecting a straw man: where are these alleged 100% believers in THD etc? THD is most often mentioned on here by two groups: newbies, and those who falsely claim that some others are 100% believers in THD. Real engineers understand the limitations of THD and have been debating for decades how best to come up with a better measure of distortion.
I would expect a genuine high-end designer to understand elementary engineering such as Fourier analysis and sampling theory, and of course genuine high-end designers do; they need to understand it in order to exploit it, because physical objects can't avoid it however much wishful thinking goes on.
I would expect a genuine high-end designer to understand elementary engineering such as Fourier analysis and sampling theory, and of course genuine high-end designers do; they need to understand it in order to exploit it, because physical objects can't avoid it however much wishful thinking goes on.
You cannot believe in FFT for 80%. If the FFT shows 0.5% 2nd, that's 0.5% 2nd.
Actually, in some sense your statement sometimes is true. People who don't know a lot of electronics and have no strong fundament can be successfull in high end audio. They are very good in marketing and have a nameless slave who IS very good in electronics design their stuff.
I could give examples except that I might end up getting sued...
Edit: Vlad, see JC's post above. ALL about electronics and strong fundamental knowledge. Are you denying that JC is a successfull high-end designer?
jan
Actually, in some sense your statement sometimes is true. People who don't know a lot of electronics and have no strong fundament can be successfull in high end audio. They are very good in marketing and have a nameless slave who IS very good in electronics design their stuff.
I could give examples except that I might end up getting sued...
Edit: Vlad, see JC's post above. ALL about electronics and strong fundamental knowledge. Are you denying that JC is a successfull high-end designer?
jan
Last edited:
A good example how important listening is, is the sometimes (or somewhere) better sound late in the night. If there would be only "measurers" and no "listeners", I doubt that the whole line disturbance topic would have been researched.
Now we know better, and this is easily explainable with measurements.
Generally speaking, without listening, we would not be where we are now.
This is the big advantage (and the reason of their success) People like Charles Hansen, John Curl, Nelson Pass, Wayne Colburn, Jocko Homo and others have:
a) They listen, and they can hear things others cannot, either because those do not believe that there are differences or because those do not have the (trained) ears to do so. Same as I'm not able to do things with my body that others can, this has been shown to me again these days watching the olympic games....
.
b) At the same time, they have the theoretical background not to start at zero but develop on a sound knowledge base.
Many have only a) OR b) and this gets difficult if you want to make well engineered AND good sounding gear.
Now we know better, and this is easily explainable with measurements.
Generally speaking, without listening, we would not be where we are now.
This is the big advantage (and the reason of their success) People like Charles Hansen, John Curl, Nelson Pass, Wayne Colburn, Jocko Homo and others have:
a) They listen, and they can hear things others cannot, either because those do not believe that there are differences or because those do not have the (trained) ears to do so. Same as I'm not able to do things with my body that others can, this has been shown to me again these days watching the olympic games....
.b) At the same time, they have the theoretical background not to start at zero but develop on a sound knowledge base.
Many have only a) OR b) and this gets difficult if you want to make well engineered AND good sounding gear.
Last edited:
That was the intention. The target is usually those of us who trust our ears, trust evidence, and do our best to exclude non-auditory factors.
It is reasonable to assume a wide range of talents both viewing and posting. That is the essence of this site IMHO.
A somewhat odd statement, given what I've read from others (and posted myself IMHO). It makes me wonder why you decided to say such an interesting thing...
Oh, that explains it.
Richard. It is not necessary to denigrate all others in order to bolster one individual.
We agree. Analog is the lost art. E/M field theory is basically abandoned.
I have found very few post docs who really get it.
I lament the current state of education in these fields.
Get used to what? Insults? I think your choice of words was most unfortunate.
Writing to a larger audience is great and very important. It is not necessary to insult the entire audience in order to write to them. Sounds like boot camp.
jn
Other argument, with equal validity is, that those who can hear, hear simply because they believe in differences (or simply they wish to hear "something", so they "hear" it .), but no one of them is able to show this ability in DBT, but they simply reject DBT. .And we can continue forever..
For hi rel, I'm sticking to gold over...and hermetic.
My app requires about 6 thousand normally closed switches, only one contact problem can shut the machine off. Hermetic gold is the only way to survive...
jn
Jan,
RC filters are old but most treatments today cover corner frequency, passband, stop band, etc. Tilt seems forgotten. It was a big deal in the late 40's and started going away by the 60's.
Dick,
With the vast number of posts the S/N ratio is extremely low. But for the really good stuff you have to sign an NDA.
SY,
Are you really confused about the difference between Fourier theory and applications?
As to patents, Don Lancaster opines that with enough research everything can be shown as prior art. I did an article on how to make a non-adaptive filter feedback eliminator for live PA use. After that Bell Labs filed for and received a patent on the algorithm. It is clear they never actually tried it as their version had serious problems. (It sounded dreadful.) Now that their patent has expired I am building actual working version for my clients. (It is possible I may patent the parts they did not cover and are important to the proper function.)
RC filters are old but most treatments today cover corner frequency, passband, stop band, etc. Tilt seems forgotten. It was a big deal in the late 40's and started going away by the 60's.
Dick,
With the vast number of posts the S/N ratio is extremely low. But for the really good stuff you have to sign an NDA.
SY,
Are you really confused about the difference between Fourier theory and applications?
As to patents, Don Lancaster opines that with enough research everything can be shown as prior art. I did an article on how to make a non-adaptive filter feedback eliminator for live PA use. After that Bell Labs filed for and received a patent on the algorithm. It is clear they never actually tried it as their version had serious problems. (It sounded dreadful.) Now that their patent has expired I am building actual working version for my clients. (It is possible I may patent the parts they did not cover and are important to the proper function.)
Having written my PhD dissertation about it and run R&D at Nicolet, I think not. I won't answer for Scott, who seems to know it at least as well.
Lancaster is fun to read, but he truly does not understand patents and how to monetize them.
No thanks. I stated what I believe is "real". No need for endless discussions.
One closing remark from my side:
I prefer to listen to those who have a track record of being successful in the audio market. Although it is nice to have, I am not in need for scientific proof for their statements.
I have already mentioned some of them, there are others too. Their experience and what they communicate about what works and what not saved/saves a lot of time for me. Interestingly, they agree on many things (what does that tell us btw?).
I am very grateful about their contributions and help in open accessible fora.
I didn't really think so, but noise by definition is random. So of course we use an approximation to it and an FFT based analyzer to tune sound systems.
And exactly what was you dissertation, since I thought your PHD was in chemistry?
As to monetizing patents, locally I am quoted $30K by folks who don't actually understand the issues. A DC patent attorney is much more reasonable, but I just don't have confidence in his abilities either. So while there are two items I think may be worth it, I am still looking.
Patents:
You should be able to file a patent for $3-4K. However the prosecution could make the tab $10K+ and if it needs to go to appeal (either you are weak or the examiner is a dunderhead) it could get much more expensive. Complex patents like the insides of a microprocessor could get much more expensive. But this is trivial compared to defending a patent. If there isn't a lot of money at stake the patent is an expensive piece of paper on the wall. The big guys get patents like kids get trading cards and use them that way. Startups do it to create a "tangible" value for the investors, often leading to the founders getting screwed and the investors owning worthless paper. If you are not a big player its difficult to get a return. I would not recommend getting a patent for your personal benefit unless you like the look of the document.
You should be able to file a patent for $3-4K. However the prosecution could make the tab $10K+ and if it needs to go to appeal (either you are weak or the examiner is a dunderhead) it could get much more expensive. Complex patents like the insides of a microprocessor could get much more expensive. But this is trivial compared to defending a patent. If there isn't a lot of money at stake the patent is an expensive piece of paper on the wall. The big guys get patents like kids get trading cards and use them that way. Startups do it to create a "tangible" value for the investors, often leading to the founders getting screwed and the investors owning worthless paper. If you are not a big player its difficult to get a return. I would not recommend getting a patent for your personal benefit unless you like the look of the document.
Fourier transform infrared photoacoustic spectroscopy, specifically of electrically conducting polymers. My work involved acquiring and manipulating extremely small acoustic signals (nominally not periodic!) using condenser microphones. That's where I learned low noise design, since the commercially available preamps were not satisfactory.
Irrespective of whether noise is random (it certainly is by definition), any finite length time domain noise signal can be treated as periodic and Fourier transformed to give the frequency representation. And back again to generate the original noise signal. That's really basic stuff which has been explained to you again and again.
That turns out not to be the case. In fact, your best chance of a return is if you get ripped off by a big company and hire a shark on contingency. Monetizing patents does not end with license fees, but it does depend on our fabulously corrupt legal system.
- Status
- Not open for further replies.