Hello all !!
I'm kind a new member on this forum, but I must say I find
this forum very interesting. One can learn a lot from people
in this forum.....and it's very good to read as many opinions as possible on a subject.
I'v read a lot about discrete opamps in this forum, and many
has mentioned the JC-2, Elektor preamp, Great River... but I
can't see a thread about the 990 opamp (The one that is sold
by JH company, and if I'm not wrong was developed by Jensen).
Is there a reason for that? Did anyone buy/build this opamp
or had a chance to listen to it? How does it compare to
other discrete designs?
Udi.
I'm kind a new member on this forum, but I must say I find
this forum very interesting. One can learn a lot from people
in this forum.....and it's very good to read as many opinions as possible on a subject.
I'v read a lot about discrete opamps in this forum, and many
has mentioned the JC-2, Elektor preamp, Great River... but I
can't see a thread about the 990 opamp (The one that is sold
by JH company, and if I'm not wrong was developed by Jensen).
Is there a reason for that? Did anyone buy/build this opamp
or had a chance to listen to it? How does it compare to
other discrete designs?
Udi.
Don't forget Forssell
Check out Mr. Fred Frossell's design here:
I have a line level pre built with these....wonderful!
Greg
Check out Mr. Fred Frossell's design here:
I have a line level pre built with these....wonderful!
Greg
It IS simple, but the difference is in the power supply that you can operate it from. Being able to go all the way up to +-24v gives LOTS of headroom. The original design for the pre that I built was used in a pro +4dBu environment for a mastering engineer, mine however is only for home use, so maybe it is overkill, but that's what I like!
Greg
Greg
Consider one of the discretes from Borbely Audio...
Quite excellent - you can package the design tighter like the 990 or API modules if you need compactness...
I built up some 990s years back - they were very good in their day, but don't have the performance available today.
The main deal was the supermatched bipolars for super low noise at the input and the little inductors on the legs of the diff input pair... the whole thing is written up in JAES.
Is there a schematic for the 992?
Deane Jensen was ahead of the pack in his day.
_-_-bear
Quite excellent - you can package the design tighter like the 990 or API modules if you need compactness...
I built up some 990s years back - they were very good in their day, but don't have the performance available today.
The main deal was the supermatched bipolars for super low noise at the input and the little inductors on the legs of the diff input pair... the whole thing is written up in JAES.
Is there a schematic for the 992?
Deane Jensen was ahead of the pack in his day.
_-_-bear
udip said:
I'm kind of confused by this... what are you trying to get done?
Is this just for messing around or is there a particular application?
If you want superior sound & studio output level capability then a design like the Borbely stuff is hard to beat.
If you don't need studio output levels, then it's hard to beat some of the newer monolithic opamps. Darn near impossible actually.
_-_-bear
> I think that a good discrete opamp design will outperform any monolithic opamp
Why?
The modern chip processes make essentially the same transistors as we buy individually. The sad old days of lateral PNPs are still here, but we can also get perfectly good PNPs on some processes.
Chip designers are sometimes clueless about real audio quality, but some are good and some are very-good.
Thermal feedback used to be an issue but today it is probably less than PCB trace resistance coupling.
If you are driving 10V 10K levels, the only real issue is that chip Class AB outputs are sometimes underbiased. This looks "better" to chip makers, but is usually bad news for sound.
If you are driving 10V into honest 600 ohms (very rare today), then you are at the upper limit of common small chips. Even there, either the old 5534 or some of the new kids will beat +18dBm with ease.
What Jensen's original 990 did, aside from quite low noise voltage, is give GOBS of output so you could use massive fan-out and low-Z feedback without strain. 14V RMS across 75 ohms is not a real problem, and that is over 2.5 WATTs.
I respect Mr Curl's view, and sure agree that there are many places where the sonic advantage of "soft" FETs are better than "naked junction" BJTs. On the other hand, sometimes you want to go with the serious gain and drive that comes easiest with BJTs.
The Jensen 990 is still a gold-standard in the recording industry, especially in heavy work. Jensen did lay out all the problems of a good audio op-amp, and solved them in best 1970s style. Fads come and go, and we have better FETs now than then, and Deane himself might have worked on an FET amp (if he had lived). But it would not be better in all ways, just in some ways.
You can get Jensen's AES paper from his company, VERY good reading. (They can't post it for copyright, but are happy to paper-mail upon request.)
John Hardy is using (essentially) the original Jensen 990 design. See John's paper http://www.johnhardyco.com/pdf/990.pdf
JLM has a 990-pinout module, called JLM99V, http://www.jlmaudio.com/JLM99V.htm that is FET input, and is not zero DC offset. This seems to be very sweet in many applications that don't need the extreme low noise voltage or output drive (or DC coupling) of a full 990. It can eat +/-34V, and will drive 600 ohms. Some sticklers would object that a 0.6V offset defies the "op-amp" concept; but in audio, we can often overlook DC error.
> I think that a good discrete opamp design will outperform any monolithic opamp
I don't think you can let yourself off the hook that easy. You can find perfectly good chips and perfectly awful modules. (You should see some of the Burr-Brown antiques the mikepre nuts find on eBay.) A lot depends on what you are doing with them: a gold-plated Deane-blessed 990 would be useless as a condenser mike head-amp, just as the 627 is not the bee's-knees for transformerless operation from 50-ohm ribbon mikes on harpsichord.
Why?
The modern chip processes make essentially the same transistors as we buy individually. The sad old days of lateral PNPs are still here, but we can also get perfectly good PNPs on some processes.
Chip designers are sometimes clueless about real audio quality, but some are good and some are very-good.
Thermal feedback used to be an issue but today it is probably less than PCB trace resistance coupling.
If you are driving 10V 10K levels, the only real issue is that chip Class AB outputs are sometimes underbiased. This looks "better" to chip makers, but is usually bad news for sound.
If you are driving 10V into honest 600 ohms (very rare today), then you are at the upper limit of common small chips. Even there, either the old 5534 or some of the new kids will beat +18dBm with ease.
What Jensen's original 990 did, aside from quite low noise voltage, is give GOBS of output so you could use massive fan-out and low-Z feedback without strain. 14V RMS across 75 ohms is not a real problem, and that is over 2.5 WATTs.
I respect Mr Curl's view, and sure agree that there are many places where the sonic advantage of "soft" FETs are better than "naked junction" BJTs. On the other hand, sometimes you want to go with the serious gain and drive that comes easiest with BJTs.
The Jensen 990 is still a gold-standard in the recording industry, especially in heavy work. Jensen did lay out all the problems of a good audio op-amp, and solved them in best 1970s style. Fads come and go, and we have better FETs now than then, and Deane himself might have worked on an FET amp (if he had lived). But it would not be better in all ways, just in some ways.
You can get Jensen's AES paper from his company, VERY good reading. (They can't post it for copyright, but are happy to paper-mail upon request.)
John Hardy is using (essentially) the original Jensen 990 design. See John's paper http://www.johnhardyco.com/pdf/990.pdf
JLM has a 990-pinout module, called JLM99V, http://www.jlmaudio.com/JLM99V.htm that is FET input, and is not zero DC offset. This seems to be very sweet in many applications that don't need the extreme low noise voltage or output drive (or DC coupling) of a full 990. It can eat +/-34V, and will drive 600 ohms. Some sticklers would object that a 0.6V offset defies the "op-amp" concept; but in audio, we can often overlook DC error.
> I think that a good discrete opamp design will outperform any monolithic opamp
I don't think you can let yourself off the hook that easy. You can find perfectly good chips and perfectly awful modules. (You should see some of the Burr-Brown antiques the mikepre nuts find on eBay.) A lot depends on what you are doing with them: a gold-plated Deane-blessed 990 would be useless as a condenser mike head-amp, just as the 627 is not the bee's-knees for transformerless operation from 50-ohm ribbon mikes on harpsichord.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.