This reminds me of another conflating issue. Did you do any measurements to try and see what the difference was that you were hearing (or thought you heard?).
In particular, did you check with an oscilloscope? In my experience, this sort of thing is always because some factor with the "bad sounding " cap in the power supply was causing the regulator to oscillate, generally at MHz frequencies. Supersonic oscillations are inaudible, but it "confuses" amplifier circuitry and thus results in audible distortion.
When this sort of thing happens, shunting the alleged bad electro with a film or ceramic cap always cures it. (And Bateman said shunting electros was a waste of time. He didn't think of this one.)
In particular, did you check with an oscilloscope? In my experience, this sort of thing is always because some factor with the "bad sounding " cap in the power supply was causing the regulator to oscillate, generally at MHz frequencies. Supersonic oscillations are inaudible, but it "confuses" amplifier circuitry and thus results in audible distortion.
When this sort of thing happens, shunting the alleged bad electro with a film or ceramic cap always cures it. (And Bateman said shunting electros was a waste of time. He didn't think of this one.)
Last edited:
Plasnu,
No worries. What it comes down to is this: There are lots of unknown conflating issues that I can't know because I wasn't there. But I go by the rule that if you can hear a difference, there will be a reason, and there will be an instrument test that will reveal the reason. The trick is to think of the right test.
Audio is only witchcraft in the hands of witches.
Ah ha! If you had put a scope on the power rail, you may well have found it was oscillating, but still managing close to the correct voltage on average.
Something that I recall after I made my last post: Some years ago a business ascociate who had some electronics training had been a management type for years. He had a stereo - he had owned it for years - and decided to replace all the electrolytics. He said the sound, which he had always thought was good, was now quite noticeably better - and told anyone who would listen that so-and-so brand was much better than the such-and-such the stereo manufacturer had used.
Now, that being true was less likely than winning the big prize in the State lottery. I asked "Did you check the old caps as you pulled each one out."
Answer was "no - don't have a way to do that."
I asked "Do you still have them? Bring them in and I'll test them."
Well, he had thrown them out, but did rescue some from the bin and bring them in. I found all of them a bit low in capacitance, and one had high ESR. Nothing remarkable though.
Then he said to me, as an aside, "Oh yeah, I found the owner's manual. It tells you how to adjust the bias, so I touched that up as well."
Oh, right. Most likely, his amp had deteriorated so slowly over the years he hadn't noticed. The bias adjustment might have been all that was needed, but he changed the caps as well. That cost him money, so that had to the reason for the cure.
No worries. What it comes down to is this: There are lots of unknown conflating issues that I can't know because I wasn't there. But I go by the rule that if you can hear a difference, there will be a reason, and there will be an instrument test that will reveal the reason. The trick is to think of the right test.
Audio is only witchcraft in the hands of witches.
Ah ha! If you had put a scope on the power rail, you may well have found it was oscillating, but still managing close to the correct voltage on average.
Something that I recall after I made my last post: Some years ago a business ascociate who had some electronics training had been a management type for years. He had a stereo - he had owned it for years - and decided to replace all the electrolytics. He said the sound, which he had always thought was good, was now quite noticeably better - and told anyone who would listen that so-and-so brand was much better than the such-and-such the stereo manufacturer had used.
Now, that being true was less likely than winning the big prize in the State lottery. I asked "Did you check the old caps as you pulled each one out."
Answer was "no - don't have a way to do that."
I asked "Do you still have them? Bring them in and I'll test them."
Well, he had thrown them out, but did rescue some from the bin and bring them in. I found all of them a bit low in capacitance, and one had high ESR. Nothing remarkable though.
Then he said to me, as an aside, "Oh yeah, I found the owner's manual. It tells you how to adjust the bias, so I touched that up as well."
Oh, right. Most likely, his amp had deteriorated so slowly over the years he hadn't noticed. The bias adjustment might have been all that was needed, but he changed the caps as well. That cost him money, so that had to the reason for the cure.
Other things being equal, higher voltage caps will have thicker dielectric so for a given signal voltage the electric field across the dielectric will be weaker and so have less effect on any dielectric nonlinearity. This should have little or no effect in a PSU cap, especially a regulated PSU.sumotan said:
Electrolytic capacitors are actually semiconductor devices. The dielectric is an oxide layer. If you look under a microscope you can see that it is not uniform. So, you can draw an equivalent schematic diagram of an aluminium oxide-semiconductor capacitor as a distributed system with multiple parallel diodes in series and parallel with capacitors and resistors, connected through distributed resistors and inductances. I once observed a phenomenon when a push-pull amplifier powered from electrolytic capacitors produced intermods on top frequencies similar to balanced SSB modulator. They were pretty small, but still audible, as a smeared stereo soundstage of cymbals, triangles, etc. Shunting 2x470 uF 450V caps connected in series by 4 uF only film caps I eliminated this effect.
Last edited:
I don't know (haven't read all 21 pages) if it has already been mentioned in this thread, but you can build fully DC coupled tube amps with Norton DC shifters. I use this myself in my amps and can confirm that it works stable.
I found this via Google:
The Audio Signal Path; Minimising Power Supply Interaction | Richard Sears, Vacuum Tube Audio
This is an example of a Norton shifter:
Another option would be to use sweep tubes (EL509..) and drive them on the screen grid (G2)
I found this via Google:
The Audio Signal Path; Minimising Power Supply Interaction | Richard Sears, Vacuum Tube Audio
This is an example of a Norton shifter:
Another option would be to use sweep tubes (EL509..) and drive them on the screen grid (G2)
From the suggested article; "Note; even if the measured performance appears flat, our ears are quite capable of distinguishing micro-effects that limit the potential musical beauty of an amplifier and that is, I assert, the point of this hobby - so long as we are having fun"
Now, how do I make that my "signature"?
Now, how do I make that my "signature"?
Norton Shifter??
Don't need that, it was not a single component device in the 50s as it is now.
But many DC amplifiers were built, just a resister replaces that device. The other end usually goes to anything from -75v to -300V, whatever the designer needed.
I did many, usually differential for DC stability.
The other resister was often replaced by a selected neon bulb like the NE51H.
Or the 991.
Don't need that, it was not a single component device in the 50s as it is now.
But many DC amplifiers were built, just a resister replaces that device. The other end usually goes to anything from -75v to -300V, whatever the designer needed.
I did many, usually differential for DC stability.
The other resister was often replaced by a selected neon bulb like the NE51H.
Or the 991.
I think we all design our circuits do it our own way to please our own ears. I have tried resistor dividers, and found them to be unstable due to temperature. Plus you do not only divide the DC voltage, but also the AC voltage which reduces gain. A Norton shifter is not a single device and can be built with components that where already available in the late 50's, early 60's. It shifts DC voltage while keeping the AC voltage intact. I personally appreciate the voltage stability and the tonal quality of these circuits and I have stopped building C or L coupled tube amplifiers.
This is an example of a Norton shifter:
The circuit as shown is likely to exhibit poor treble response, and if overall neg feedback is used, is very likely to oscillate.
A few picofarad (if next tube is a pentode or tetrode) to 20pF or so (if next tube is a triode) across the level shifting resistor will compensate, so the fix is quite simple
The screen grid used as a control grid is insensitive and usually not as linear as the true control grid. It draws power too, complicating the design of the previous stage.
All you need for the CCS is a small pentode (eg 6AU6 or 6BA6, pentode half of a triode pentode) and three resistors. Nothing could be simpler - except using a neon tube as coupling device (but neons are noisy, have a short life, and can exhibit small voltage jumps in old NOS devices and in neons intended as indicating lamps.)
DC Amplifiers
The DC Amplifiers I built were most all in industrial & scientific applications, no one had to listen to them.Many were in the regulated Power Supplies for both constant voltage & constant current apps.. When good silicon rectifiers became available they were the obvious choice for the rectifier.
The largest was a monster 4.5 KV, One Amp CV thing. The iron was all Hammond, built to order for us. One of the features of the whole thing for which I'm thankful to Prof Anderson for was an SYA setup. Save Your ***, the source relay contacts welded together on contact, they did not release on shut down. There was a very loud explosion as the SYA device operated.
By 1965 there were quite a few multi layer silicon devices on the market. But nothing that would shift something like 100V CC or more. All easy to do now, we are in an age of plenty.
The DC Amplifiers I built were most all in industrial & scientific applications, no one had to listen to them.
Many were in the regulated Power Supplies for both constant voltage & constant current apps.. When good silicon rectifiers became available they were the obvious choice for the rectifier.The largest was a monster 4.5 KV, One Amp CV thing. The iron was all Hammond, built to order for us. One of the features of the whole thing for which I'm thankful to Prof Anderson for was an SYA setup. Save Your ***, the source relay contacts welded together on contact, they did not release on shut down. There was a very loud explosion as the SYA device operated.
By 1965 there were quite a few multi layer silicon devices on the market. But nothing that would shift something like 100V CC or more. All easy to do now, we are in an age of plenty.
Right, and that micro-effects are dynamic, while measurements are static and averaged.
If its your hobby, that's fine - it's your money and it's not hurting you or anyone else.
But any engineer who designs audio equipment with a claim to quality sound professionally or for home and friends should read Eberhard Zwicker's book "Psych-Acoustics" (Springer Verlag).
Zwicker spent a lifetime researching just what our ears can and cannot do. The book (3rd edition) is about 450 pages of data, and is effectively a very detailed technical specification of the human hearing system. And when you have read and digested Zwicker's book, you will be well aware that the statement quoted by jjasniew is complete nonsense.
And if you are not convinced by the book that all aspects of human hearing can be measured and reduced to data, you can obtain the CD of test signals and try it on your self, if you have a high quality reproduction system. Or, if you have sufficient electronics knowledge and test equipment, you can reproduce Zwicker's tests.
There is NO defect in electronic audio reproduction that cannot be measured. None. But some of the measurement/test techniques required are unknown in much of the audio community.
This has been a very entertaining read, thanks to all the well informed contributors.
What is the concept that we call a 'soundstage'? One of the things that I hear differently with my tube amplifier, over my quite reasonable solid state setup, is the virtualisation of the placement of the instruments.
This is dependent on the recordings, but in the case of an orchestra, or a miked recording studio, that information must be part of the signal, as phase shifts and wave reinforcements. Our ears are tuned to this, since we have one on each side of the head.
Is it just that I am not comparing like with like? I.e. my solid state amplifier is not on the same level as my tune amplifier?
What is the concept that we call a 'soundstage'? One of the things that I hear differently with my tube amplifier, over my quite reasonable solid state setup, is the virtualisation of the placement of the instruments.
This is dependent on the recordings, but in the case of an orchestra, or a miked recording studio, that information must be part of the signal, as phase shifts and wave reinforcements. Our ears are tuned to this, since we have one on each side of the head.
Is it just that I am not comparing like with like? I.e. my solid state amplifier is not on the same level as my tune amplifier?
OldHector,
Is the difference necessarily the tube amp vs the solid state amp?
Are you using the exact same loudspeakers in both cases? Or different speakers? Or are the speakers on your tube amp in a different room or location within the room, as the speakers for your solid state system?
Is the music source the same or different? I've got a turntable on one system, but my other system is fed from a CD drive. Naturally they sound quite different.
Speakers and cartridges have far more effect on sound stage definition than do amplifiers.
Is the difference necessarily the tube amp vs the solid state amp?
Are you using the exact same loudspeakers in both cases? Or different speakers? Or are the speakers on your tube amp in a different room or location within the room, as the speakers for your solid state system?
Is the music source the same or different? I've got a turntable on one system, but my other system is fed from a CD drive. Naturally they sound quite different.
Speakers and cartridges have far more effect on sound stage definition than do amplifiers.
30+ years ago, I worked for Digital, who had a product similar to the platform diyAudio runs on - but this was on ASCII terminals. They opened it to all employees and one of the 1000+ topics was "Audio" which hosted many lively discussions about what someone claimed to be able to hear; cables with braided twists, capacitors, speaker wire, etc. One of the most popular contributors once claimed they could hear the effect of a pair of Radio Shack RCA right angle adapters he was forced to use temporarily to manage his rack cabling.
Of course all the measurement people (myself included) were floored. Many tried to convince him in reply that it was not possible for a 2" piece of metal to have any effect at audio frequencies, but he wouldnt back down. Since then I've learned that to tell another they're not hearing what they claim to hear - or its not possible - isnt the best road to travel down in a "I showed him" context.
These folks would have get-togethers, one of which I attended. The host did something to their system, which caused everyone else to literally jump from their seats. I was like, "huh? What? I dont get it!". So from my own experience, there's attributes to be heard and other people who can hear them which I cannot. Unless it's simply a matter of ear training...
Now, wind 30 years forward and my hearing is far worse than before, but I'm still interested in all this (hence the post). I can still hear attributes - enough - to make it interesting. I'm more of a believer in measurement and simulation gets you in the ballpark - the rest of any positive refinement is empirically acquired. That's in all electrical and mechanical contexts BTW - not just audio.
So we're back to the ancient debate "Why does the amplifier that measures better sometimes not sound as good as one that measures worse". I think we all know the answer to that one. Whether all perceptible sonic attributes are quantifiable will, I'm sure, be debated long after I'm gone. While I'm still around, I'll choose not to fight that one. Thanks.
Last edited:
Provided the amp used is operating properly I've always thought the reproducing part of the system, the speaker in its enclosure & the environment in which the whole thing is operating is the determining factor.
And a glass of wine.
Never worried about what resister or cap was in the cct, I was too busy listening to the Rolling Stones & Neil Young! And may others.
And a glass of wine.
Never worried about what resister or cap was in the cct, I was too busy listening to the Rolling Stones & Neil Young! And may others.
And long after I've gone. too. I expect.
There seems to be a number of camps in this debate, leaving aside for the moment that some people confuse music origination (where effects are exploited for creative use) and music reproduction (where effects are not wanted), and that many people enjoy a particular sound (which is not accuracy in reproduction, but something they have every right to enjoy).
The major camps are:
1. The camp that believes that commonly used measurements and engineering theory tells all. I am not in this camp. This camp is not very vocal, and they are wrong.
2. The camp that believes that no measurement or engineering care can ensure quality sound. I'm not in this camp either. This camp is very vocal, perhaps there is a sense of insecurity amongst them. They are wrong too.
3. The camp that believes that all defects that can occur can be measured, but some of the measurements are not in common use. I'm in this camp.
It is interesting that the more a professional engineer is involved, career-wise, actually in audio reproduction, and the more knowledge and experience he acquires, that he typically starts off in Camp 1 or Camp 2, and ends up in Camp 3.
And then there are the snake oil merchants, who sell directionally oriented speaker cables, gold infused solder, and other stuff that in the 1960's would have had them prosecuted for fraud. Lots of ignorant folk buy the snake oil stuff. Unfortunately, the view of government authorities in English-speaking countries (USA and Australia being prime examples) is now that the economy is helped by not having restrictions, and let the buyer beware.
Hence the rise in electronic components with magic properties, neutraceuticals, and a certain product, a little box, well known, that the promoter claims that if you install it in your car, will repeal moose in North America, and kangaroos in Australia, at ranges that basic physics says that this would require the power levels of a locomotive. And people buy all this stuff, and swear it works.
I'm in Camp 3, after a career in audio electronics. But due to government policy resulting in the demise of manufacturing in Australia, I moved to electric power generation - it was either that or become unemployed. If I was to move back into audio, I would have to pretend to be in Camp 2 - that's were the money is.
There is a lot of truth in that. A heck of a lot. But it is not the complete truth.
For instance, loudspeaker non-linear distortion is essentially proportional to cone displacement and the square of displacement. A solid-sate amplifier can suffer from cross-over distortion, which is a maximum at a particular (usually quite low) level, where speaker distortion has become very small.
Intermod in an amp is an important defect, as then the bass can muddy up the mid-band, where human hearing is most acute. Intermod from bass is relatively unimportant in quality loudspeakers due to the use of separate bass drivers and sub-woofers.
Intermod is also important in an amp, as supersonic components can thereby be folded back into the audible range. But speakers, due to the physics of voice coils and cones, just ignore supersonic frequencies.
But in practice loudspeaker imperfections dominate, and completely mask any defects in a competently designed modern amplifier. It always amuses me that amplifiers are specified for THD down to 0.001% and even lower, when typical loudspeakers are seldom better than 0.5% or so, and only that good mid-band at moderate levels.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.