The advantage of monolithic is that if you want to throw 25 extra transistors and diodes in to improve performance, that's easy. Another is that all those extra parts can be accommodated without long inductive and hum-prone leads. And yet another is that matching in critical circuit areas can be incredibly tight, much better than you can do with discrete. The performance of modern monolithic opamps (putting aside fashion and totally unproven claims) reflects this, with amazingly low distortion, excellent bandwidth, high CMR and PSR, and much lower offsets. Noise performance is comparable.
Amazingly low distortions for opamps, wide bandwidth for opamps, and PSRR that is needed for opamps only. Couple of vacuum pentodes and MOSFETS and I have better audio amp than can be made with any opamp, with amazingly low distortions and excellent bandwidth, without any feedback and no CMRR is needed. ADD feedback, and compare to opamp. If CMRR is needed I would better use input transformer that provides high linear CMRR on wider bandwidth that is needed because of cellphones and other sources of RF energy.
The bear with tubes is the excess low frequency noise. The high gm triodes are actually pretty astonishingly quiet above the noise corner. If the plate is very lightly loaded the distortion can be quite low. Not that I have anything against transistors
MC preamp on planar triode 6S17K-V with microphone transformer 150 Ohm : 40 kOhm is dead quiet. The same combination with dynamic microphone is of course even quieter, but I use mostly large diaphragm condensers. What else to dream about? To fit such preamp inside of handheld digital recorder? Impossible, unfortunately. But if to use external microphones it is not a big deal to add external preamps.
Old news though with some new twists. Maybe 15 years ago, a buddy of mine (who was a researcher at IMB on STM methods) was fabricating arrays of cold cathode vacuum tubes in silicon, using a nanoscale pyramid for a cathode, with electron flow via tunneling to the anode. The control grid was a ring.
Another can of worms. 6S17K with 8 micron grid wires already have kind of ultra-linear curves. Tubes are good because of their properties. For such properties they have to be relatively big. When I want low input capacitance and high input resistance I prefer JFETs.
Oh boy, this is evolving towards another X vs Y contest.
Go on and argue all you want, but I think it's pointless.
Wave, if someone doesn't like the "typical" (whatever that might be) sound of tubes, there is no sense expounding their virtues. Anyway, nobody is arguing against them.
I have to say I have heard a few preamps made around a pack of op amps which sounded really good to my ears. About as frequent as with any other technology as such.
It's not what you make it with, it's how you make it. Wave, without ever even seeing, much less hearing, any of your products, I will happily concede that you can make it come on song. Basically, I base this on the fact that you are a tube affictionado, and have worked with them a lot, and have over the years mastered them. I think this stands to reason.
Same goes for John and trannies - he's been at it for 40+ years, so it's reasonable to assume he knows his preferred pack of devices (which we all have) inside out, in terms of sound, almost certainly better than the man who designed it.
Who comes out better is a matter of auditioning and, let's be honest, personal taste to a fair extent. The most difficult part of auditioning any piece of audio equipment is putting one's own personal tastes aside as much as you can and judging it on its own merits.
Surprisingly few people can do that.
Go on and argue all you want, but I think it's pointless.
Wave, if someone doesn't like the "typical" (whatever that might be) sound of tubes, there is no sense expounding their virtues. Anyway, nobody is arguing against them.
I have to say I have heard a few preamps made around a pack of op amps which sounded really good to my ears. About as frequent as with any other technology as such.
It's not what you make it with, it's how you make it. Wave, without ever even seeing, much less hearing, any of your products, I will happily concede that you can make it come on song. Basically, I base this on the fact that you are a tube affictionado, and have worked with them a lot, and have over the years mastered them. I think this stands to reason.
Same goes for John and trannies - he's been at it for 40+ years, so it's reasonable to assume he knows his preferred pack of devices (which we all have) inside out, in terms of sound, almost certainly better than the man who designed it.
Who comes out better is a matter of auditioning and, let's be honest, personal taste to a fair extent. The most difficult part of auditioning any piece of audio equipment is putting one's own personal tastes aside as much as you can and judging it on its own merits.
Surprisingly few people can do that.
It is 2012 and I observe that people still seem to divide into 'camps'. I use FETs, BJTs, MOSFETs and valves. I also use soft HF domes, metal HF domes, plastic cones, paper cones, dome mid ranges, horns etc. The conclusion I have reached is that no one technology is better than the rest. Even the outstanding performance of full range electrostatic speakers has a weakness because they can not be used in all environments. Each technology needs a tuned implementation. When this is achieved, hearsay perceptions of signature characteristics disappear, eg. horns sound like any other speaker when dealt with properly. Too often the 'tuning' is ignored because it is too arduous or expensive.
Samuel Groner made a very valuable study on the distortion of various opamps. This is in itself a study on how the ICs have been tuned. On pure distortion performance many perform well. However there is only one out of the 45 chips tested that I would prefer to use for audio. On this forum Nelson Pass has an ultra simple amplifier circuit which has trimmers for distortion balancing. I haven't built the circuit, but I have considerable confidence that those who build it and experiment with it, would learn a great deal about the nature of distortion.
My personal endeavours now exploit the differences between technologies rather than try and crush them into one uniform result. I have learned a lot from making distortion my friend, not my enemy.
Samuel Groner made a very valuable study on the distortion of various opamps. This is in itself a study on how the ICs have been tuned. On pure distortion performance many perform well. However there is only one out of the 45 chips tested that I would prefer to use for audio. On this forum Nelson Pass has an ultra simple amplifier circuit which has trimmers for distortion balancing. I haven't built the circuit, but I have considerable confidence that those who build it and experiment with it, would learn a great deal about the nature of distortion.
My personal endeavours now exploit the differences between technologies rather than try and crush them into one uniform result. I have learned a lot from making distortion my friend, not my enemy.
I used to make distortions at late 70'th, when designed analog synthesizers and guitar effects. As the result I somehow calibrated perception, so when I hear some distortions I imagine changes of waveforms in dynamic and know what to measure, what to look for, how to minimize.
Of course it is not wise to be limited by selected set of technologies, because each active component has advantages for certain functions. It was exactly my point: if opamps are suboptimal for high end audio amplification, why to struggle and use them when other active components work better?
Of course it is not wise to be limited by selected set of technologies, because each active component has advantages for certain functions. It was exactly my point: if opamps are suboptimal for high end audio amplification, why to struggle and use them when other active components work better?
I would like to make the point that I basically agree with Wavebourn. IF I were to use EVERY possible device available, there would be tubes in my designs, at least, once in a while. Sometimes, I now use IC's for the audio path.
Now, why do I stick with jfets, avoid tubes, and seldom use IC's.
For two reasons:
Tubes are hard to use with the topologies I am known for developing. All I can do with them is to copy someone else's design, developed throughout the history of vacuum tube design. They require a separate high voltage supply, and they can shock you, if you accidently touch their circuitry in the wrong place.
All else being equal, it would be MORE EXPENSIVE to add tubes to MY designs.
What about IC's, then?
IC's are NOT designed with enough current to be truly Class A, in almost every case. I like Class A operation, because it gives lower order harmonics and IM products that the ear can barely detect.
Static distortion measurements are limited to one dimension of what we apparently hear with our ears, and we have known this for more than 40 years, so why dote over them?
I have found a few things that are useful for successful audio design:
One is to operate Class A whenever it is possible and practical.
Two is to work with the most linear topology that I can get, or make.
Three is to have fast circuits with slew rates of at least 5V/us for preamps and 50V/us for power amps.
Four is to use as little feedback as I can get away with, and still meet some reasonable specification that looks good (but not necessarily great) on a spec-sheet.
Other factors are optional and even arguable. Works for me!
Now, why do I stick with jfets, avoid tubes, and seldom use IC's.
For two reasons:
Tubes are hard to use with the topologies I am known for developing. All I can do with them is to copy someone else's design, developed throughout the history of vacuum tube design. They require a separate high voltage supply, and they can shock you, if you accidently touch their circuitry in the wrong place.
All else being equal, it would be MORE EXPENSIVE to add tubes to MY designs.
What about IC's, then?
IC's are NOT designed with enough current to be truly Class A, in almost every case. I like Class A operation, because it gives lower order harmonics and IM products that the ear can barely detect.
Static distortion measurements are limited to one dimension of what we apparently hear with our ears, and we have known this for more than 40 years, so why dote over them?
I have found a few things that are useful for successful audio design:
One is to operate Class A whenever it is possible and practical.
Two is to work with the most linear topology that I can get, or make.
Three is to have fast circuits with slew rates of at least 5V/us for preamps and 50V/us for power amps.
Four is to use as little feedback as I can get away with, and still meet some reasonable specification that looks good (but not necessarily great) on a spec-sheet.
Other factors are optional and even arguable. Works for me!
I am with John, except I do not try to use "as little feedback as I can". I try to use either zero, or as much as I can. But practically other factors play the role, like using some parts that have better other parameters, while have excess voltage gain. If compensation does not screw down phase response it is acceptable. If it does, it is "too much". Sometimes it is better to use 2 stages with local feedbacks than one, even if one can provide enough of gain with less feedback. Especially with opamps, when it is better to use dual opamp in place of single one.
Also I am in love with nested feedbacks. They remind me human body: combination of nervous and endocrine systems.
Also I am in love with nested feedbacks. They remind me human body: combination of nervous and endocrine systems.
Last edited:
Where do I sign as an underwirter? Word for word, I agree.
Yet, I was the one who seems to think that op amps are not as bad as they are made out to be.
But I do not deny their shortcomings. I was also the one who claimed, and still do, that they sound a whole lot better with added current boosters.
Overall, I'd say this - op amps are a cost effective way to produce what I'd call "Mid Fi", meaning possibly better than most, but not up to the concept of High End as I know it (meaning the part of audio which pushes limits hard).
The trouble is, John, that what you and I think of when told "High End" is not what it used to be. These days, that mostly means sculptures not cases, showing how far the CNC machinig process has come, and I guess some actually find ways to produce very good circuits inside as well. But if by some misfortune somehow somebody should forget to install a blue LED on the front panel, that's it, sorry, no cigar, you ain't High End, bud.
You know what I mean. And that really, actually, wholly pi**es me off.
- Status
- Not open for further replies.