WB,
I agree. Everything you say here is confronted and argued, often with nastiness thrown in.
The tube forum is far better. I wonder why this is, though?
Hugh
I agree. Everything you say here is confronted and argued, often with nastiness thrown in.
The tube forum is far better. I wonder why this is, though?
Hugh
IMHO, tube people are more close to a position that best sound is more an art, experience and empirical knowledge from the world of components, rather than information from textbooks. SS people tend more to compete in educational background, and still believe that simulation tools and textbook receipies provide solid basis for success. Since listening impressions are subjective, in the real world and in this forum wins the one who better argues on the basis of education.
Last edited:
Aha, so Vladimir, it's about winning?
Always knew there was something insidious about education.....
Hugh
Always knew there was something insidious about education.....
Hugh
I don't think this is correct. 'SS people' are usually the first ones to point out that the simulator is NOT the real world, that sims have their limitations, that the models have their limitations, and that you really need to build that amp to know how it behaves and sounds. I think you misrepresent this.
And 'winning in the real world'? I would argue that 'winning in the real world' has only a very weak link to the arguments. There are successful ss products that have very high feedback, or zero feedback, or tube products with low feedback, or without feedback etc etc. 'Winning in the real world' seems more stronly linked to marketing. Unless you agree that 'all amps sound the same anyway' but I don't think you'll do that 😉 . Nor would I.
jan didden
Without any minor intention to upset somebody, I guess that any field of activity needs some basis for competition, like todays technical knowledge, and this basis predefines what will be achieved at the end and who is considered as a leader. Competiton is needed and inherent to any activity, but at the same time it dictates a bit aggressive style, which people learn starting from school and university. I would say SS forum is normal, and the tube forum seems a bit marginal, since people do not struggle there and think only about their DIY. Everybody choose his own happiness.
I agree with Jan that success is more often linked to marketing, and therefore, regardless of the technical merits, products succeed or fail for other reasons.
Therefore, I concur that the SS forum, where it is possible to achieve very good measurements which do not correlate well with sonic excellence, is based more on technical argument based on credentialism, and not upon subjective issues like sound quality, which it can be argued are spurious and too subjective to be trusted.
Since the world is now resoundingly solid state, then the best brains evidently work in this area, and their knowledge is both contemporary and commercially valued. Therefore these guys will be naturally competitive, and all too often their aggression will result in intergalactic urinary contests.
Ergo, the tube forum is more polite.......
Hugh
Therefore, I concur that the SS forum, where it is possible to achieve very good measurements which do not correlate well with sonic excellence, is based more on technical argument based on credentialism, and not upon subjective issues like sound quality, which it can be argued are spurious and too subjective to be trusted.
Since the world is now resoundingly solid state, then the best brains evidently work in this area, and their knowledge is both contemporary and commercially valued. Therefore these guys will be naturally competitive, and all too often their aggression will result in intergalactic urinary contests.
Ergo, the tube forum is more polite.......
Hugh
Last edited:
Maybe we a bit too hot discussing obvious things. Success - among whom, among zillion consumers buying components 300...1000 USD ?
Best brains are not needed there, just educated medium engineers.
Success among SS forum participants? Here one would definitely need extensive scholarship, ability to demonstrate it and to press opponents, what is also important for success in any field. It would be silly to take such a style close to hart, and no need to change forum becuase of this.
Success before the God (or among some super experienced listeners from other planet)? Approaching thruth is just an individual and personal way.
Best brains are not needed there, just educated medium engineers.
Success among SS forum participants? Here one would definitely need extensive scholarship, ability to demonstrate it and to press opponents, what is also important for success in any field. It would be silly to take such a style close to hart, and no need to change forum becuase of this.
Success before the God (or among some super experienced listeners from other planet)? Approaching thruth is just an individual and personal way.
Interesting points Hugh. Just a few observations:
'Sonic excellence' is not an objective measure either; it's subjective, in the eyes (ears?) of the beholder and varies all over the place. A useless term 😉
'The best (ss) minds' are definitely not in audio; from a technical point of view, audio 'is done'. No money or fame there, I'm afraid. Audio is one of the very few areas where you can sell a 'breakthrought' with just a change in packaging and a good story 😎
jan didden
What about MJE15035 - 4 ?
I was thinking to use them together with TO3 Finals to get some kind of P3A, but I had bad feedback from the forum without technical explanation. Too big junction? BD139/40 defenetly better?
Thanks
effebi
I was thinking to use them together with TO3 Finals to get some kind of P3A, but I had bad feedback from the forum without technical explanation. Too big junction? BD139/40 defenetly better?
Thanks
effebi
Last edited:
What a great thread!
When the "defined process" is sticking to the "popular test" results then the test becomes the standard no matter how horrible the amplifier sounds. This is seen in a lot of operational amplifiers also as edified in the LM833. When "unpopular test" are suggested for tube or solid state then a storm of rejection of the test occurs because it causes the "highly regarded" circuit to fail miserably which is taken to mean the test is wrong and not the circuit. There are many examples by many authors in this forum of suggested test rejected by the "all knowing."
I do not note who said feedback eliminates S21 affects but that person is dreaming. The reflected impedance is reflected no matter what else happens. Feedback can provide a correction signal to the reflected impedance (error) in the last gain stage but the effect still exist. And of course sufficiently low impedance of the correcting stage (at the summing node) is required or the (total) correction signal will not occur because the drive impedance is to high. Said another way, one cannot pull a train on a bicycle. To pull the train requires a powerful (low Z) source. To correct using feedback (or any other method) the "correcting force" must be sufficient to overcome the error. Do not forget that feedback always has delay so what happens between the time the error occurs and the feedback signal is applied? Another big issue.
The result is this, running 100ma on the last gain stage (VAS in this thread) on a ±50 volt SS amp means the intrinsic stage impedance is in the range or 250-500 ohms or so. Running 1ma on the last gain stage means the impedance is 25,000-50,000 ohms. Which do you suppose will drive a bad impedance reflected load better? This is covered in "watt sucking fire ball" thread.
Last gain stage (VAS stage, our subject) current is set here by what seems safe for the device used. 10-20 watt 150MHz devices in a ±50V amp usually end up class "A" running 4-8 watts depending on several factors. That is 40-80ma in this example. Testing here suggest this is a good trade off point. Using tiny devices and running 1ma does not do well on the very unpopular test which are always used here.
When the "defined process" is sticking to the "popular test" results then the test becomes the standard no matter how horrible the amplifier sounds. This is seen in a lot of operational amplifiers also as edified in the LM833. When "unpopular test" are suggested for tube or solid state then a storm of rejection of the test occurs because it causes the "highly regarded" circuit to fail miserably which is taken to mean the test is wrong and not the circuit. There are many examples by many authors in this forum of suggested test rejected by the "all knowing."
I do not note who said feedback eliminates S21 affects but that person is dreaming. The reflected impedance is reflected no matter what else happens. Feedback can provide a correction signal to the reflected impedance (error) in the last gain stage but the effect still exist. And of course sufficiently low impedance of the correcting stage (at the summing node) is required or the (total) correction signal will not occur because the drive impedance is to high. Said another way, one cannot pull a train on a bicycle. To pull the train requires a powerful (low Z) source. To correct using feedback (or any other method) the "correcting force" must be sufficient to overcome the error. Do not forget that feedback always has delay so what happens between the time the error occurs and the feedback signal is applied? Another big issue.
The result is this, running 100ma on the last gain stage (VAS in this thread) on a ±50 volt SS amp means the intrinsic stage impedance is in the range or 250-500 ohms or so. Running 1ma on the last gain stage means the impedance is 25,000-50,000 ohms. Which do you suppose will drive a bad impedance reflected load better? This is covered in "watt sucking fire ball" thread.
Last gain stage (VAS stage, our subject) current is set here by what seems safe for the device used. 10-20 watt 150MHz devices in a ±50V amp usually end up class "A" running 4-8 watts depending on several factors. That is 40-80ma in this example. Testing here suggest this is a good trade off point. Using tiny devices and running 1ma does not do well on the very unpopular test which are always used here.
Last edited:
PFB by current in VAS looks like a good approach; but in real world it is still hard to find the proper equilibrium: while in simulation universe some parameters are easily fixed, in the real one they are variable.
In what context of feedback are you refering too? I am talking about local nested feedback, use of an error amplifier for the driving stage after the VAS. This 'linearizes' the impeadance that the VAS stage will see. If the CB stage sees a constant impedance, then it will not be affected by the final follower output stage. You may still have a global loop, but the local FB/FF minimizes the distortion of the output stage, which is load dependent, from entering into the global loop, as important IMO as providing the CB VAS with a constant load Z. This is not the same as adding more emitter following stages to increase gain, or increasing the global FB, in this case, the load Z would still be reflected to some level back to the VAS.
Last edited:
These popular test are just voltage measurements, not sound based test, and I know that nothing of them using any sound based test. Also they are close to failures than correct results, even for measuring rms and peak power, voltage only based test are worse thing.
My driver stages are small transistors, some using parallel 2SC3355, 2SC4308, some others using PN2222A-2907A, 3906-3904, PN200, they are working with floating voltage rails. 6 of 2222A in parallel could deliver 1Ampere.
That "feedback" is all very nice for steady state but feedback takes time and during that time the load is bad. Every amp I have tested using these schemes illuminated above oscillates easily during the "waiting for the feedback to arrive" interval. Therefore, for audio it does not really work.
Why not just make the last gain stage so it will drive the bad Z rather than trying lots of schemes to get around the situation with solutions that do not work dynamically? The simple and direct solution is lowering the Z at the last gain stage sufficiently to drive the reflected load.
Why not just make the last gain stage so it will drive the bad Z rather than trying lots of schemes to get around the situation with solutions that do not work dynamically? The simple and direct solution is lowering the Z at the last gain stage sufficiently to drive the reflected load.
I thought (maybe incorrectly) that the last gain stage of an amplifier was the output stage, where current gain happens.
I also thought that the typical single ended VAS ran in "class A" regardless, as it swings from rail to rail, and that the only real concerns for selecting its idle current was linearity and sufficiently supplying the drivers of the output stage.
The thread is very long enough; did anybody mention that emitter follower may be easily turned into VAS by a single autoformer? 😉
CBS240,
any kind of feedback worsen the isolation between stages. The common collector amplifier provides poor isolation because of the high inherent feedback. The (varying) load has a great impact on linearity, but the distortion due to an added stage can be worse. What kind of error amplifier do you mean? The driving stage after the VAS should be a FET.
any kind of feedback worsen the isolation between stages. The common collector amplifier provides poor isolation because of the high inherent feedback. The (varying) load has a great impact on linearity, but the distortion due to an added stage can be worse. What kind of error amplifier do you mean? The driving stage after the VAS should be a FET.
EF and transformer may good idea. This mine, small, high current, also fast enough.
Attachments
No Offence
Some folks think that V-FET stages can function well with EC only😀
Trying other approach isn't worth for the jumping zombies.🙄
Some folks think that V-FET stages can function well with EC only😀
Trying other approach isn't worth for the jumping zombies.🙄
Wavebourn,
If this is so then what is the point of running the VAS at 1A idle, if a triple or quad EF stage at output cannot get away with reflection.