|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
173.52.246.134
In Reply to: RE: And... posted by MarkJohns on November 14, 2015 at 14:32:47
>In John Atkinson's video on YouTube where he explains how he measures
>loudspeakers he says that if a loudspeaker measures badly it will sound
>badly.
That is correct.>But he recently seemed to say that if an amplifier measures badly but
>sounds good then he'd recommend it.You are confusing what I say as an individual with what I do as the editor
of a magazine that publishes reviews by several different writers. I may
not feel that an amplifier measures well but if one of my team recommends
that amplifier on the basis of his careful and prolonged auditioning, the
amplifier is recommended by the magazine. But if you read our "Recommended
Components" listings you will see that such recommendations are always
tempered by the inclusion of the measured shortcomings.
>If that is what he said am I the only one who finds an inconsistency
>here?No inconsistency.
John Atkinson
Editor, Stereophile
Edits: 11/15/15Follow Ups:
"You are confusing what I say as an individual with what I do as the editor of a magazine"
Well now you've got me even more confused.
"you will see that such recommendations are always
tempered by the inclusion of the measured shortcomings."
Sorry, that still doesn't clarify my inability to reconcile what you said about loudspeaker measurements indicating that poor measurements invariably (or almost invariably) leads to poor sound but poor measurements for amplifiers doesn't. Maybe it's just me but I find it inconsistent.
> that still doesn't clarify my inability to reconcile what you said about
> loudspeaker measurements indicating that poor measurements invariably (or
> almost invariably) leads to poor sound but poor measurements for
> amplifiers doesn't.
With loudspeakers, the criteria for listeners finding the sound to be "good"
are well-established. While it is still possible for a listener to prefer
a loudspeaker that is demonstrably poor, I have found this to be rare.
With amplifiers, the criteria are uncertain. However, where a listener
prefers the sound of an amplifier that in my opinion is poorly engineered,
this is because either:
a) there is a relatively high level of second-harmonic distortion, which,
as long as it is not accompanied by excessive intermodulation distortion,
many listeners will prefer; or
b) the amplifier's output impedance is sufficiently high that the tonal
changes this introduces with a specific loudspeaker compensate for that
speaker's own departures from neutrality.
And as I have repeatedly written in Stereophile, the question is left
begging as to whether the listener likes the sound of a poorly engineered
amplifier because of what it does wrong or despite it. To resolve that
question requires resources beyond the reach of a monthly magazine.
John Atkinson
Editor, Stereophile
With loudspeakers, the criteria for listeners finding the sound to be "good" are well-established. While it is still possible for a listener to prefer a loudspeaker that is demonstrably poor, I have found this to be rare.
If what you said was really true, there wouldn't be such a wide variety of speaker designs and such a wide variety of different listener preferences for one type of presentation vs. another. So much of what a speaker sounds like is determined by how it couples with the room, and the standard suite of measurements doesn't do a good job of revealing that.
In the market, there are some designers who seem to be aiming for a smoothly downward sloping far field response in a typical listening room, and their measurements sometimes reveal tradeoffs that balance deviations in near-field on-axis response against deviations in off-axis response. And other designers who seem to be trying to make the near field on-axis response as ruler flat as possible, and let the far-field and/or off-axis response fall where it may. Few can do both.
The designers who are focusing on power response produce better sounding speakers than the designers who are optimizing for a single microphone position in the near-field. But the latter produce speakers which are more widely praised as being better engineered. In this sense, that one metric for speakers what THD is for amplifiers.
Given the wide diversity of equipment and opinion of such, it would seem abundantly clear that there is no accounting for taste. Is it not true that we see more amp manufacturers and more types of amps now than ever before? The first person who can accurately and reliably predict what the vast majority of people will subjectively prefer will no doubt corner that market, or at least we would see convergence in the market, conversely, it seems we are getting ever more choices, rather than less...Are we to believe no one wants the spoils which would accrue to he who pleases the masses???Given the diversity in tastes, the diversity of associated equipment, and the uniqueness of each combination in its intended space, it would seem naive at best to think that there is one correct and absolute solution which satisfies so many divergent, and many times opposing, needs.
try it! you know you want to!
Edits: 11/16/15
"
With amplifiers, the criteria are uncertain. However, where a listener
prefers the sound of an amplifier that in my opinion is poorly engineered,
this is because either:
a) there is a relatively high level of second-harmonic distortion, which,
as long as it is not accompanied by excessive intermodulation distortion,
many listeners will prefer; or
b) the amplifier's output impedance is sufficiently high that the tonal
changes this introduces with a specific loudspeaker compensate for that
speaker's own departures from neutrality.
"
Why should you consider an amp with high level of second-harmonic distortion poorly designed when it is well known that even relatively high levels are inaudible. As long as there is the subsequent avoidance of high order harmonics (known to sound bad) within the normal usage range of the amp then this could be a conscious and good engineering to avoid the nasties.
Why should it be considered poorly engineered because it has a high(ish) output impedance as long as the variations introduced are less than the inherent variations in the speaker? Mathematically it has been shown that a damping factor of 5 is sufficient to not introduce more than 1 db variation and demonstrate adequate electrical damping on a speaker (and if not then it is probably a woefully underdamped speaker).
"likes the sound of a poorly engineered
amplifier because of what it does wrong or despite it."
Again, I challenge your orthodox view of what is a poorly designed amplifier. I would argue that a poorly designed one is one that chooses numbers over psychoacoustic principles.
"To resolve that
question requires resources beyond the reach of a monthly magazine."
I don't think this is true. You could assemble a panel and try to establish your own correlations with the many different amps that come through your shop. Once you have a correlation about sound quality we can see if it matches what some others are saying (like CHeever and his model, for example). This might reach an unpleasant conclusion for the majority of manufacturers though and maybe it is a road you don't want to go down for business reasons??
Taking things to extreme, the SET or single ended vacuum tube triode zero feedback low power output power amplifier with an unregulated power supply strikes me as the worst conceivably engineered general purpose audio design. They measure horribly but some people seem to like them very much. Some are very expensive but I wouldn't own one for free.
I do not like vacuum tube amplifiers of almost any type as they seem to have a peculiar type of distortion solid state amplifiers don't exhibit. I attribute this to the likelihood that the hysteresis loop in the core of the output transformer results in this sound. I think it may be a form of asymmetric harmonic distortion. I left tube amplifiers behind over 40 years ago and never looked back. I was surprised myself at the Vacuum Tube Valley audio show in 2007 that I was able to pick out by sound alone the only solid state amplifier at the show that I heard. It was the only one that didn't have this kind of distortion. To my ears it's even worse than crossover notch distortion. But a lot of people like them too. Generally their harmonic distortion is several orders of magnitude higher than well designed solid state power amplifiers.
I left tube amplifiers behind over 40 years ago and never looked back.
Which is when I find they started getting interesting. Today's best units with stiffer power supplies, balanced operation, better passive components, etc. do not sound like your father's tube amps. :)
"I attribute this to the likelihood that the hysteresis loop in the core of the output transformer results in this sound. "
Care to explain what this sound is that you hear? BTW; you do know that a push/pull output transformer crosses zero (demagnetized effectively) but that a SET output transformer does not (they have DC offset so they are always magnetized).
The alternating magnetic field that couples the primary of an iron core transformer is described by a BH curve that graphs the magnetic inducing field resulting from current in the primary winding to the magnetic field actually induced in the core. The induced field when it's going from positive to negative does not retrace the same curve going from negative to positive. That means that for example if the input waveform is symmetrical such as a sine wave, the output will never be. The impedance of the transformer is also inductive and has a relatively high source resistance meaning it will interact with the speaker crossover network, has a relatively poor electrical damping factor, and it invalidates measurements where the load in the bench test was a pure resistor.
What I hear from them is a kind of muffled, muted sound with a relatively louder but not very deep bass. Speakers designed with the characteristics of this kind of amplifier in mind will typically sound shrill and thin connected to a solid state amplifier whose source impedance is very low, does not have hysteresis distortion, and has a flat frequency response with many loads and at many loudness levels if well designed. Tube amplifiers also begin to self destruct from the moment you turn them on due to the heat they generate. The tube itself depends on the mechanical relationship of the electrodes which can change with time. The crystal structure of transistors by contrast is locked in. If they change they will fail catastrophically.
First, just for the record, JA's comments above are spot on. I find it odd that they were not perfectly crystal clear to you.
OTLs are a kind of tube amp that has no output transformer. We've been making them for nearly 4 decades now and if they were unreliable we'd have gone out of business years ago.
However your comments about hysteresis and the like are not based on reality. In SETs if the transformer is part of the plate load hysteresis plays no role whatsoever.
The area where the transformer can be a concern as far as distortion goes is in a push pull design wherein the power tubes are biased close to class B operation. When the tubes shut off a spike can occur in the output transformer as the magnetic field collapses. However there really aren't any high end tube amplifiers biased anywhere near Class B (and even AB2 is super rare) so this is not likely to be the basis of your complaint.
More than likely you are concerned about the 2nd harmonic which many tube amps exhibit, which can cause a 'lushness' or 'warmth' to the sound. This is easily avoided even in a transformer coupled design by implementing fully differential balanced topology This BTW is how many transistor amps are designed). BTW, if transistor amplifiers employ single-ended circuits (and many of them did back in the 1960s and early 70s) they too will exhibit a 2nd harmonic as this is a topological issue and has nothing to do with tube or solid state.
You might want to ask yourself " if triodes are some of the most linear amplification known to man" (and they are) "how come tube amps have so much distortion?" The answer has everything to do with how feedback is or is not used and the complexity of the circuit. Tubes are often so linear that no feedback need be applied and the circuit can be quite simple.
But tube amps existed in the 1960s that had THD of only 0.05% at full power; better than any of the transistor amps of the ear and for some years since! So while you may have developed your opinion over years and decades, you might also want to consider that it is in fact opinion and not fact.
Well just for the record, whatever that means, just because you agree with JA doesn't mean it's right.Second of all as soon as tubes heat up they start to change and that is far more rapid by orders of magnitude especially for power output tubes than it is for transistors.
Third of all, hysteresis loss in magnetic core material is real. Here are some BH curves in the link below. They are the ones with the double ESS curves.
The best magnetic material for transformers is still supermalloy. It's still less than perfect.Fourth, I don't ever recall seeing a tube power amplifier that claimed anything less than 0.25% THD and 0.25 % IM. Preamplifiers yes, power amplifiers no. Often they measure far worse.
Finally, even if none of what I said above were true, I still don't like the way they sound. And as I said, even I was surprised that after a whole day of listening to countless tube amplifiers, I was able to pick out the one solid state amplifier at the 2007 VTV show by listening alone. I knew immediately. The exhibitor was North Creek Audio.
Edits: 11/16/15 11/16/15
Those do like like hysteresis curves to me, but they don't apply the way you think they do to output transformers. I used to think they did; IOW I made the same mistake.
Julius Futterman built the amplifier with the low distortion specs I mentioned. It was an OTL. The OTLs I make go to 1Hz-300KHz with full power.
Power triodes tend to drift a lot less than pentodes; sounds like you don't have much experience with them. We can set the bias on our amps and 6 months later they likely will not need adjustment. They are surprisingly stable. It all has to do with design- some designers are better at getting the tubes to behave than others. It does not surprise me at all that you could hear the difference between tubes and transistors- most people can. But FWIW here's something interesting: the more feedback you can put into a tube amp the more it will sound like solid state. Most solid state amps run a lot of feedback! Interestingly though, the solid state amps that don't run feedback are some of the stand-out amps that even tube lovers like...
In 1983 I attended a Seminar and demonstration at the WQXR auditorium during the AES convention in NYC presented by New York Audio Labs, Harvey Rosenberg's company. Their engineers had taken a lot of death bed testimony from Futterman and his widow gave them all of his technical papers. They used the same principles but provided a modern solid state power supply. They also built a rig for adjusting the bias. It used to take Futterman an entire day to get this right. They could do it in five minutes.
Among the many shortcomings of vacuum tubes is that thermionic emission of the cathode varies with the square of the temperature. That means a small change in temperature results in a much larger change in the rate electrons boiled off. Negative feedback properly used mitigates this problem. Improperly used it creates more distortion than it is intended to solve. In fact used very wrong and you've built an oscillator. The rigorous study of negative feedback and servo systems which operate on exactly the same principle entails understanding mind blowing mathematical equations. This is why tyros invariably get it wrong and should stay away from it. Like any powerful tool it is dangerous. You can build a house with a hammer or smash your thumb.
If I had to buy a tube amplifier on sound alone, it would be an OTL type. But I still prefer solid state for practical reasons.
see the link.
I'm not against feedback, but sometimes you are better off without it given some of the compromise that develops.
One problem that is well known (see Norman Crowhurst) is that feedback enhances higher ordered harmonic production that might not have existed open loop. Since the ear uses the higher ordered harmonics as loudness cues, the feedback can and does cause bifurcation of the amplifier output; IOW increased loudness cues due to the increased higher ordered harmonics, even though they might be in trace amounts.
This phenomena is old enough that Crowhurst was writing about it in the 1950s. Its not gone away- the ability of feedback (in sufficient quantity) to cause an amplifier to behave as a voltage source has been considered a desirable trait for taming speaker resonance in box speakers. So this issue is still around over half a century on.
Less quantify , how much is too much feedback ....
Go Rossi ......
Here's the problem: feedback is often inappropriately applied.
Here's a link that describes the problem quite well.
Its probably best if you read through all the parts.
Ralph , you missed my question, how much feedback is too much for Ralph , 1db, 10db, 20db, 30db ..? Would 10db be perfectly acceptable , no need to get into the academic stuff like stability and OL bandwidth , just the amount you would be comfortable with...
Go Rossi ......
All designs are different and what works for one may not do so for another.
I like a gentle overload characteristic and there is almost no way to do that if you are running loop feedback- the overload onset is rather sudden.
In addition I like lots of detail and tonal neutrality, which means not only no 2nd harmonic but also no artificial brightness. Both are colorations IMO. To accomplish 'no brightness', using feedback is tricky indeed (as it will generate higher ordered harmonics where they may not otherwise have existed at all) so I tend to avoid it.
It does not matter to me if this limits the speakers that I can use. If a speaker will not sound right unless the amplifier employs feedback; well, put another way its always going to sound bright due to higher ordered harmonic generation if feedback is used. Since that is an annoying artifact and a coloration, IMO/IME if the speaker needs the amp to have feedback the result is that it might sound like a great hifi but it will never sound like real music. IOW I like the speakers that have the ability to sound real- and that can only occur if the amp is low distortion and lacking feedback. Its a narrow path for sure, but it works.
...like I have said, some prefer solid state distortion over tube distortion.
Neither are perfect.
They are different - I can't listen to solid state amps very long so I'm your opposite in this regard.
You guys must have a storage room full of bad SS amps ...:)
Go Rossi ......
...the way a ss amp adds a character to the music sounds different than the way the music sounds through tubes.
It's not a measured distortion thing - I hear it on even the even best ss amps I've heard.
Same for toobs, so it's all the same , choose your poison i presume..
Go Rossi ......
...or find your passion.
Which sounds more like real music and irritates you less.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: