Great post Jan. I agree with SY that Barrie Oliver should be on the list. I have seriously learned from about 1/2 of the people on this list of remarkable linear engineers, through their papers and personal interchange.
Please remember that linear audio engineering is a subset of linear engineering, so we may have a different list.
For example: In 1974, at an ISSCC Conference, I asked Bob Widlar, in front of 100's of people, what to do with the TIM problem due to limited slew rate in IC's. He told me DIRECTLY that IC's were NOT made for audio and that we should make our own circuits. Walt Jung was there, you can ask him if you want. This is important, because we often try to force fit analog IC's into audio circuitry and hope for the best. Intelligence and even commercial success does not remove prejudice or ignorance of what it takes to make a very high quality audio product.
Please remember that linear audio engineering is a subset of linear engineering, so we may have a different list.
For example: In 1974, at an ISSCC Conference, I asked Bob Widlar, in front of 100's of people, what to do with the TIM problem due to limited slew rate in IC's. He told me DIRECTLY that IC's were NOT made for audio and that we should make our own circuits. Walt Jung was there, you can ask him if you want. This is important, because we often try to force fit analog IC's into audio circuitry and hope for the best. Intelligence and even commercial success does not remove prejudice or ignorance of what it takes to make a very high quality audio product.
Well, it must be nice to have personally met all those people. I enjoy that as far as I can make it happen. So far, I've met Bob Pease and Jim Williams in the flesh.
I can imagine that 1974 ICs would wreak havoc on audio. Your discrete designs at the time would run rings around those 741's and 709's.
OTOH, if I see what those IC smiths turn out in 2009, ICs get very competitive. As long as you can live with +/-15V supply, that is. For higher levels, there's still no substitute for discretes.
Jan Didden
I can imagine that 1974 ICs would wreak havoc on audio. Your discrete designs at the time would run rings around those 741's and 709's.
OTOH, if I see what those IC smiths turn out in 2009, ICs get very competitive. As long as you can live with +/-15V supply, that is. For higher levels, there's still no substitute for discretes.
Jan Didden
Actually Jan, IC's have not changed as much as you might first guess. They have gotten somewhat cheaper, and a few advanced techniques have evolved, but look back at the Harris IC's in the 1970's, made for the military and note the similarity to what is made today.
NOW we worry about the output stage crossover distortion as much as earlier on.
NOW we worry about the output stage crossover distortion as much as earlier on.
I think I'm with janneman here. Output stage distortion tends to be better, stability is improved and I think no one made Op-Amps that had noise performance in the 2-3nV/Square Root Hz bandwidth, not to mention significant reductions in the so called pop-corn noise.
Still,l everything I read from knowledgeable people, says Op-Amps still don't achieve the sound performance of discretes.
Still,l everything I read from knowledgeable people, says Op-Amps still don't achieve the sound performance of discretes.
There are many opamps for audio these days that have output stage class A currents of over 5mA. For 600 ohm loads, that's over 4V in class A, enough to blow your speakers with even a low sensitivity poweramp. For a regular poweramp input of 10k or more, even 1mA of standing current would be OK.
Jan Didden
Jan Didden
There were pretty good hybrid designs made for the military-industrial complex that did just about everything, that were made by NSC. Check it out.
5mA of ClassA output into 600r is 3Vpk or about 2Vac.janneman said:
But, if the impedance is reactive as in headphones, then expect the ClassA current to be expended with <1Vac output.
If you mean 5mA of output stage bias current, then give some/one example that suits audio.
5mA biased push-pull is class A up to 6Vpk/600 ohm. 5mA bias currents results in 10mA class A.
LSK389
My intentions is to buy 20 LSK389 B and C, is it worth to deal with, are they as good as the original 2SK389 BL and V.
The firm who offer me the devices will bill me about 3£ each, maybe it is better to buy something else, a lot of single JFET and do some matching.
I feel that I need to make something like the Vendetta as a home project.
Advice please!
Kamskoma
My intentions is to buy 20 LSK389 B and C, is it worth to deal with, are they as good as the original 2SK389 BL and V.
The firm who offer me the devices will bill me about 3£ each, maybe it is better to buy something else, a lot of single JFET and do some matching.
I feel that I need to make something like the Vendetta as a home project.
Advice please!
Kamskoma
john curl said:
The Harris linear ICs were indeed remarkable for their time. I remember cutting some of them open in the early '70's at Bell Labs to see what they were doing. If I recall correctly, they were diectrically isolated with complementary PNPs.
BTL was one of the earliest, if not the first, to commercialize a full complementary junction-isolated linear IC process, called CBIC, in the mid-'70's. This enabled us to rid ourselves of the terrible lateral PNPs while having a cost-competitive process.
The availability of complementary PNPs allowed us to make much better output stages as well, in some cases using folded emitter follower drivers that permitted pretty good control over class AB bias current. Good op amps were required for precision active filters, even for telephone (voiceband) work.
Apart from the Harris ICs being a bit expensive, I believe that they did not have JFETs and did not have any (or a good) on-chip trimming process. Those are two improvements that were important as time went on, in my opinion.
BTW, Paul Brokaw of Analog Devices also comes to mind for the list.
Cheers,
Bob
Bob Cordell said:
BTW as one would expect this list is very US centric. I would certainly consider Willy Sansen, or Rudy van de Plassche.
If you read to the end (a bad place to say it) the list is not in any order other than historical. The author asks for your input and contributions for the list.
Go for it.
_-_-bear
Go for it.
_-_-bear
john curl said:
I'm surprised you would repeat a 35yr. quote as if nothing has changed. There are plenty of amplifiers with 1000X the slew rate and TIM is a concept that is best put in the grave.
bear said:
Indeed, the list is pretty subjective, and US centered. Willy Sansen is certainly a candidate, but isn't really an industry guy, more academic.
I liked the background info about those guys you read about but don't really know.
OTOH, if this is the peer selection, it reflect the perception of the analog engineering community. As democratic as a presidential election 😉
Jan Didden
How long has Bob Widlar been dead? Do you know how many years? What have you, Scott Wurcer, added to IC design that was not essentially known by Harris in the 1970's? Please provide evidence, and listening feedback. Please don't say that you invented low Rbb'. We known about that since 1966.
- Status
- Not open for further replies.