Bias the opamp output stage into class A. Very easy to do. Then you can have the pick of any VFA or CFA topology device you care to try. And, plenty of choice on te JFET front as well.
You seem to assume that complementary JFETs are the one and only option for what you call "high end". Is this correct, or is this just another excuse to avoid (good) opamps?
There are several processes that would support them but there is no compelling reason to spend the money for the extra layers. In the real world fashion architectures don't have much value. Also for reasons mentioned before they are complementary in only the grossest sense.
You know what is funny, Scott. I thought that the same thing would happen with complementary bipolar topologies, and that nobody would ever build a complementary symmetry bipolar topology, BUT Harris did! Mil requirements seem to allow for advanced topologies. By the way, your AD797 topology appears to be derived from an earlier Harris topology. However, it would seem, and I will try to find the evidence, that complementary different cmos or mosfet topologies are coming of age. What do you know? '-)
Last edited:
The basic idea behind complementary differential input stages, whether bipolar or jfet, comes from the desire to have a BALANCED SECOND STAGE. This can only be easily done with a complementary differential input stage. A 2 piece complementary jfet input stage is also possible, and POSSIBLY an improvement over the jfet complementary differential input stage, but is is difficult to get the offset out. I abandoned it in 1973, for the full complementary differential jfet input stage, because of this.
CMOS has complementary in its name so they were on to it in 1967 (for different but related reasons), Harris just applied the same logic to bipolar (pun intended).
Don't waste your time I only moved the take off for the output when I noticed it enabled the crossover nulling. You won't find that, on the surface it's illogical since it removes one stage of buffering the output now needs to be a triple.
With IC's, one rarely knows the application it will be put towards. So All specs are important to get to be their best. BUT, in audio line level stages or power amps, there can be a laxing of the input for lowest noise in the whole wide world. And, if you dont rely primarily on gnfb to make the thing linear as possible, the good complimentary jFEts issue subsides and can be made to work very well with more local degeneration and modest gnfb used for excellent results. This was my thought and implimentation of desired complimentary jFET topologies for line level amps. - Thx RNMarsh
Last edited:
These are all fashion design principles common to a narrow body of audio designers.
Man jewelry as I call it. But it lets me buy my preferred forms of man jewelry.
I believe J.C. have never tried a "GOOD" modern OPA, and when we read his comment, we just have to replace "OPA" by "µ471".
The main thing is OPAs are not fashionable, and not good for guru designers to sell mystery.
I wonder where can be evil class B, when measured distortion can be < -100 dB @ 100 kHz
Last edited:
I have 100's of IC's in my office, and samples of just about everything that I can find useful for audio. I have been using and testing IC's since 1966, a long time ago. In 1969, I used uA741's for servo design, wonderful devices for that.
In 1970 I used the Harris HA911 in a studio board, with the slew rate of +5/-2.5V/us and +/- 24V operation. This was my FIRST failure and I tend to believe that the IC's were part of the reason for the failure. The ceramic caps were the second reason.
Mark Levinson used IC's FIRST in the LNP-2 preamp, hand selected by Dick Burwen for audio quality, in advance. This is when first I saw the XOVER DISTORTION in many of the samples of the HA-911. Mark's use of the selected HA-911 worked pretty well, BUT ultimately he had me design a discrete replacement for it. After 1975 or so, all LNP-2 and LNC-2 products use my discrete 'upgrade'.
In 1970 I used the Harris HA911 in a studio board, with the slew rate of +5/-2.5V/us and +/- 24V operation. This was my FIRST failure and I tend to believe that the IC's were part of the reason for the failure. The ceramic caps were the second reason.
Mark Levinson used IC's FIRST in the LNP-2 preamp, hand selected by Dick Burwen for audio quality, in advance. This is when first I saw the XOVER DISTORTION in many of the samples of the HA-911. Mark's use of the selected HA-911 worked pretty well, BUT ultimately he had me design a discrete replacement for it. After 1975 or so, all LNP-2 and LNC-2 products use my discrete 'upgrade'.
Last edited:
Despite what i read everywhere, on my experience, slew rate is one of the most important thing to care of when we chose an OPA for audio.
20V/µs is far to be enough, and more it is, more transparent is the device.
Slew rate can be > 1000V/µs, Noise can be as low as 1.5nV/√Hz, output current >150ma, distortion so low that it is not a concern.
Of course, implementation is important, and, often, in a high end system, you cannot replace one specific OPA by an other. Need to tune the value of the feedback resistance, parallel cap for no overshoot etc...
And, of course, great care of power supply and HF decoupling.
Looking for a perfect sample can be time consuming, there are so many available devices, and there is often one of your requisite you have to sacrifice.
20V/µs is far to be enough, and more it is, more transparent is the device.
Slew rate can be > 1000V/µs, Noise can be as low as 1.5nV/√Hz, output current >150ma, distortion so low that it is not a concern.
Of course, implementation is important, and, often, in a high end system, you cannot replace one specific OPA by an other. Need to tune the value of the feedback resistance, parallel cap for no overshoot etc...
And, of course, great care of power supply and HF decoupling.
Looking for a perfect sample can be time consuming, there are so many available devices, and there is often one of your requisite you have to sacrifice.
So what, Esperado? What makes YOUR OPINION any better than MY opinion? I might agree with you that high slew rate is important, but maybe not 1000V/us for a preamp. Seems a little excessive, when 5V/us is considered OK in every test, EXCEPT FOR FM DISTORTION, that we have made with real audio signals.
The fact that i'm not in concern with the sell of anything ? You share your opinion, when i don't agree, i share mine. Am-i not allowed ?
Anyway, i believe generalizations like "OPA not good for audio" are obviously suspect, like all generalizations.
I use to compare my preamp circuits with strait wires. When i'm not able to hear a real difference, i believe i do not have to worry anymore. And , as i said previously, i achieved that many time with various OPAs. Funny, never with OPA rated for audio.
Last edited:
- Status
- Not open for further replies.