The parallel is there, like it or not. The only reason I mentioned it is to illustrate that the audio industry is behaving like any and all other industries.
I have yet to see a car made by anybody which will stand up to its published consumption levels, just as I have yet to be convinced that an amp with a THD of 0,0001% by default sounds better than another amp with a THD of 3%. Tube amps do 3% THD easy, yet many of them sound good and often better than their SS counterparts with a THD of 0,0001%.
Calculating backwards as you said is risky because you have to have a universal platform. Make a boo-boo in defining that platform and you have it wrong across the board. Declare that static THD and IM define the sound of an amplifier and you just made that boo-boo.
The auto industry will always promote a system which makes them look good, just as the audio industry will promote measuremets which make them look good, even at the expense of veracity. In fact, while I cannot prove it, I have an idea that the audio industry is what at least slows down, if not stopping altogether, the development of new measuring methods, which could be more relevant to the sound we hear in the end, just as the auto industry is surpressing alternate fuels which may require significant investment on their part.
The Top Dog is always the mass produced series production and its economics. Examples proving this are gallore.
When the Philips 721 CD player appeared 15 years ago, it was slated as a model to be sold in supermarkets (literally). The nmagazines got hold of it and lo and behold, declared it to be able to box way above its price class (true, in my view, I still have one somewhere). Then the DIY gang discovered that if you swapped its plain vanilla output op amp with a say OP 275, the sound got audibly better still. This begs the question why didn't Philips do the same?
"Complex marketing reasons" was the answer I got from Philips (I had excellent ties with their SE Europe HQ in Vienna). The cost difference would have been around $0.5 per unit, but that works out to $200,000 per a series of 400,000. What do you think, what would most manufacturers rather do, improve their product or invest that money into marketing?
Also, they would end up with a player which sounds better than the models above it, sort shooting themselves in the leg.
OK, why not use tha OP275 or some suc across the board? Ah, but that would make Philips compete with Marantz (at the time still in Philips majority ownership).
If the economic reasons can do that, imagine what they can do in promoting such measutement techniques which will show up their work, rather than show up possible faults or shortcomings.
What I see is that any efforts to really improve our measuring techniques are generally made on the fringes of the general audio industry, in small companies really trying to do better. The big companies are not interested, or if they arem, then it's strictly for in-house use.
I have yet to see a car made by anybody which will stand up to its published consumption levels, just as I have yet to be convinced that an amp with a THD of 0,0001% by default sounds better than another amp with a THD of 3%. Tube amps do 3% THD easy, yet many of them sound good and often better than their SS counterparts with a THD of 0,0001%.
Calculating backwards as you said is risky because you have to have a universal platform. Make a boo-boo in defining that platform and you have it wrong across the board. Declare that static THD and IM define the sound of an amplifier and you just made that boo-boo.
The auto industry will always promote a system which makes them look good, just as the audio industry will promote measuremets which make them look good, even at the expense of veracity. In fact, while I cannot prove it, I have an idea that the audio industry is what at least slows down, if not stopping altogether, the development of new measuring methods, which could be more relevant to the sound we hear in the end, just as the auto industry is surpressing alternate fuels which may require significant investment on their part.
The Top Dog is always the mass produced series production and its economics. Examples proving this are gallore.
When the Philips 721 CD player appeared 15 years ago, it was slated as a model to be sold in supermarkets (literally). The nmagazines got hold of it and lo and behold, declared it to be able to box way above its price class (true, in my view, I still have one somewhere). Then the DIY gang discovered that if you swapped its plain vanilla output op amp with a say OP 275, the sound got audibly better still. This begs the question why didn't Philips do the same?
"Complex marketing reasons" was the answer I got from Philips (I had excellent ties with their SE Europe HQ in Vienna). The cost difference would have been around $0.5 per unit, but that works out to $200,000 per a series of 400,000. What do you think, what would most manufacturers rather do, improve their product or invest that money into marketing?
Also, they would end up with a player which sounds better than the models above it, sort shooting themselves in the leg.
OK, why not use tha OP275 or some suc across the board? Ah, but that would make Philips compete with Marantz (at the time still in Philips majority ownership).
If the economic reasons can do that, imagine what they can do in promoting such measutement techniques which will show up their work, rather than show up possible faults or shortcomings.
What I see is that any efforts to really improve our measuring techniques are generally made on the fringes of the general audio industry, in small companies really trying to do better. The big companies are not interested, or if they arem, then it's strictly for in-house use.
This was reference kick-back diode . On measurements it did nothing obvious . My dear friend Juan in Argentina who is the local TV repairman with the PHD ( I suspect from his references to Maxwell when problem solving ) said it is analogous to a voltage multiplier . I suspect if so a PP EL 34 amplifier with distributed load must be capable of doing the same ? My brother always said double voltage factors for PP operation . Is it safe to run an EL34 in SE at 450 V with a diode connected forward biased to g2 .
If my spectrum analyzer is right it is harmless . If the author is right that it deals with complex loads better it is a very cheap modification .
My brother said UL s not simple negative feedback . This diode might challenge that belief of his . He said he could see both positive and negative happening . He felt it bootstrapped or had negative feedback depending on the moment in time . For someone not given to conjecture this was a one off . He might be right and this forces it to be negative only . As you say the charge inside the device might commutate where needed when the diode is working ( reverse biased ) . What a nasty thought .
Adding a diode in the g2 connection might have some effect with a genuine tetrode, which can have negative screen current under certain conditions. It should have no effect for a pentode or a beam tetrode, as secondary electrons are kept in their place. There seems to be a modern fad of randomly sprinkling diodes into circuits where they will always be forward biased so they either do nothing or add a little more distortion.
UL is not exactly the same as ordinary NFB, because the feedback is applied to a different electrode (g2) from the input signal (g1) with slightly different curves, but it is sufficiently close to NFB for many purposes.
UL is not exactly the same as ordinary NFB, because the feedback is applied to a different electrode (g2) from the input signal (g1) with slightly different curves, but it is sufficiently close to NFB for many purposes.
Would it be possible that a diode to g2 would help a situation where the back EMF of a speaker is producing different distortion compared with a resistive load ? I can contrive something to test this ? Assume not always using loop feedback . Very interesting about pure tetrodes . I have never seen one I think ? Maybe in a radio or TV front end ?
As for the voltages I will just have to rig up a test . I would have thought if the voltage was rising the measurements should change ?
I have seen signal applied to g2 to better approximate a proper triode . Would be interesting to speculate about what could be applied to g1 if so . I think EL34 even has g3 separate ? If so signal to g3 and what of the rest ? The EL34 is so close to a perfect concept . All it set out to do was be practical suspect ? It is even economical with heater current .
As for the voltages I will just have to rig up a test . I would have thought if the voltage was rising the measurements should change ?
I have seen signal applied to g2 to better approximate a proper triode . Would be interesting to speculate about what could be applied to g1 if so . I think EL34 even has g3 separate ? If so signal to g3 and what of the rest ? The EL34 is so close to a perfect concept . All it set out to do was be practical suspect ? It is even economical with heater current .
Signal at EL 34 g3 , it might resemble a buffer !!
Some used the cathode for a UL type correction . Hafler and Quad had a small dispute over it . I suspect the Quad is best described as cathode feedback as we have always known it . It is possible to connect the secondary of a conventional transformer to the cathode to do that , it doesn't always need a third winding . If Quad are right it means a true triode can be UL connected . It is said by some that both Quad and Hafler invented UL , Hafler is to g2 . Can someone name a present production indirectly heated power triode ? Was there ever one ?
I was told by a quite famous valve man that he used the Quad method as it halved the distortion for a given gain . He was so vague as to convince me he was wrong . He might have a point . I certainly didn't find that when cloning the PYE Mozart . I can see between triode and pentode there are many possibilities and perhaps it does work better ? I recently found that 82% UL and EL 34 g2 triode to be on measurements close to identical except gain . I reason that is because EL 34 g2 triode strapped is not a true triode ? Both look excellent on the scope , pentode is very different .
Some used the cathode for a UL type correction . Hafler and Quad had a small dispute over it . I suspect the Quad is best described as cathode feedback as we have always known it . It is possible to connect the secondary of a conventional transformer to the cathode to do that , it doesn't always need a third winding . If Quad are right it means a true triode can be UL connected . It is said by some that both Quad and Hafler invented UL , Hafler is to g2 . Can someone name a present production indirectly heated power triode ? Was there ever one ?
I was told by a quite famous valve man that he used the Quad method as it halved the distortion for a given gain . He was so vague as to convince me he was wrong . He might have a point . I certainly didn't find that when cloning the PYE Mozart . I can see between triode and pentode there are many possibilities and perhaps it does work better ? I recently found that 82% UL and EL 34 g2 triode to be on measurements close to identical except gain . I reason that is because EL 34 g2 triode strapped is not a true triode ? Both look excellent on the scope , pentode is very different .
Nothing I tried seemed to better it all things considered . I have no great interest in PP feedback valve amps . Doubtless a different world .
I have to be honest I prefer transistor amplifiers over most PP valve designs . It is very possible that the ones I tried were not very good ( TVA ) . I like the Dynaco very much , valves has nothing much to do with it . Did anyone use KT 66's or the like in one ? The Marantz 9 is very good . Both are EL 34 designs . Quad is not my favorite . I love the Quad 303 .
These images borrowed from another thread . It shows how g3 grid of an EL 34 can be strapped to the anode . This seems a perfectly reasonable thing to do if reading other peoples accounts . The EL 34 is not quite a 300 B in conventional fake triode mode , it looks very reasonable . With g3 strapped up it looks rather poor . If nothing else it says the baby EL 84 in it's triode mode is not loosing out because g3 is fixed .
A nice curve set if wanting to see UL compared .
The EL34 Tube
http://www.diyaudio.com/forums/tubes-valves/67422-el34-triode-mode-g3-where.html
And as you know, there are far too many fedayeen (holy believers) regarding specs in general, and especially static distortion measurements (THD and IM). I am not saying they are meaningless and/or irrelevant, but I think much like you I believe, I see them as a starting point only.
John, we both know that you and others like you (James Bongiorno, Dan D'Agostino and the rest of the Big Boys) are what you are because of a combination of two key factors: 1) you understand the circuits you design, you know exactly what's going on in there, and 2) because of your hearing, natural and trained.
In the end, each and every amp ever made aiming for any above normal sound quality has been "voiced" according to its designer's hearing. You can know all the electronics in the world, but if your hearing isn't clear, your product won't be either.
Too many people think that because they know electronics, they are automatically slated for greatness in design and sound; unfortunately, it turns out this is simply not so. Blind faith gets you nowhere except into the ditch.
If I had to make a choice between one man with superlative knowledge of electonics but just a so-so hearing, and another with just a so-so knowledge of electronics but a good hearing, my first choice would be both working together, and failing that, the man with the better hearing.
I do not claim that I have any special hearing, all I can claim is that what sounds good to me seems to be pleasing to a large circle of people I am in touch with. This leads me to believe, perhaps erroniously, that I have a shall we say solid hearing. And I can't wait to get the first prototype boards, solder the parts in and get down to it. No hurry, months if necessary, and with a bit o' luck, I'll have Nigel working separately in ye olde England, because I am deeply convinced that he can tell good sound when he hears it. Then we compare notes. I do trust myself, but not completely.
That's how I managed to have my current speakers - worked with a friend and a panel for 6 months on just the crossover. But I have yet to hear another speaker which pulls off the disappearing act - no box, just music. That one trait in itself is priceless - he who believes not, pray ask anyone who has ever developed a serious speaker.
How difficult is doing the same with an amp, ask John. Don't forget to ask how long it took him to do it.
who's got the Blind (testing) Faith?
Stereophile's Golden Ear reviewers had faith - but when put to a blind test - in their own listening rooms, their choice of speakers, "SOTA" tube amp and music - they failed the Carver Challenge against his modded by measurements $600 SS amp
Stereophile's Golden Ear reviewers had faith - but when put to a blind test - in their own listening rooms, their choice of speakers, "SOTA" tube amp and music - they failed the Carver Challenge against his modded by measurements $600 SS amp
Last edited:
Yes, but that does not HAVE to be so. The reason is that the basic model is wrong, it has few links to reality. Would you drive 90 km/h when you can legally and actually drive 120 km/h?
If so, then why was the unrealistic 90 km/h chosen for measurement?
Eventually, we need to be MOST careful not in reading, but in fully understanding/interpreting the specs.
For example, when you see an amp claiming and very possibly managing a THD factor at rated power of say 0,001% across the band, how do you react? Do you say to yourself wow, this is one clean amp, or are you more like me, and say to yourself, wow here's a guy who's using global NFB without prejudice, like it was for free.
Bliding specs should always make you wonder how much of them is natural and needed, and how much is plain and simple rape.
Last edited:
@JCX
I remember that, but it seems insane that about 99,99999999% of those who read it appearently missed Bob Carver's EXACT wording at the time. And he was pretty clear.
He said he would rely on measurements only, but what most missed was the rest: to produce a solid state amp which will almost perfectly mimic THE TRANSFER FUNCTION of tubes. I do not recall him ever mentioning THD, IM and the rest of it, except to say that his device would probably be better than the one he was mimicing.
There's nothing to stop you from trying that yourself, JCX. If that's what you want, of course.
But do note that not being able to tell the two apart does not automatically mean both produced what you or I might consider to be good sound. For example, many people consider Mark Levinson's products to be the first next thing after perfection; personally, while appreciating their many good aspects, they sound a bit cold, a bit detached to me. Just as Krell product sound a little brash and a touch too warm to me.
Many think of NAD as a cut above the rest company, yet to me, they are just another wanna-be. I loved dearly the old "Designed in the U.S.A., manufactured in Japan" Marantz products; now I hear them as just one of a myriad of Japanese products of that kind. Nothing to write home about.
All would be wise to note the sentence which explains how Bob didn't give a damn about astronomically expensive parts. That's anothe folly pushed upon us, it's as if the parts account for sound quality. Not saying the don't help, but sound quality they DO NOT make, only the circuit and the designer's hearing do.
I remember that, but it seems insane that about 99,99999999% of those who read it appearently missed Bob Carver's EXACT wording at the time. And he was pretty clear.
He said he would rely on measurements only, but what most missed was the rest: to produce a solid state amp which will almost perfectly mimic THE TRANSFER FUNCTION of tubes. I do not recall him ever mentioning THD, IM and the rest of it, except to say that his device would probably be better than the one he was mimicing.
There's nothing to stop you from trying that yourself, JCX. If that's what you want, of course.
But do note that not being able to tell the two apart does not automatically mean both produced what you or I might consider to be good sound. For example, many people consider Mark Levinson's products to be the first next thing after perfection; personally, while appreciating their many good aspects, they sound a bit cold, a bit detached to me. Just as Krell product sound a little brash and a touch too warm to me.
Many think of NAD as a cut above the rest company, yet to me, they are just another wanna-be. I loved dearly the old "Designed in the U.S.A., manufactured in Japan" Marantz products; now I hear them as just one of a myriad of Japanese products of that kind. Nothing to write home about.
All would be wise to note the sentence which explains how Bob didn't give a damn about astronomically expensive parts. That's anothe folly pushed upon us, it's as if the parts account for sound quality. Not saying the don't help, but sound quality they DO NOT make, only the circuit and the designer's hearing do.
Last edited:
I thought he used a null test setup and fiddled with things to make the null as deep as he could.
But he did follow up later, with another company (was it Sunfire?) and, showing his rarely-rivaled talent for marketing-newspeak, featured a "current source output" which amounted to putting a big fat power resistor in series.
When I used to go to AES local meetings, once a presenter was gushing on about how his company's "digital" amplifier sounded tube-like to various golden-ear types. As was so often the case, fortified by a slight surfeit of wine from the preceding dinner, I barked out something about really wasn't the "tube-like sound" due to the highish output impedance of the open-loop output stage and its L-C lowpass filter? And that this naturally produced frequency response changes as it interacted with the loudspeaker load?
- Status
- Not open for further replies.