Re: SET v. PP, the big fight tonight


"Andre Jute" <fiultra@xxxxxxxxx> wrote in message
news:1134964421.604342.162010@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>
> Trevor Wilson wrote:
>> "Andre Jute" <fiultra@xxxxxxxxx> wrote in message
>
> Snipped, a barney of misunderstandings deliberately created by Arny
> Krueger in an effort to win a petty and spurious point.
>
> Out of that confusion Trevor offered to supply some page numbers from
> the RDH to prove that SET amps are less desirable than PP amps. That is
> the subject here. Let's keep keep it clean, fellers and come out
> swinging.
>
> SET v. PP, THE BIG FIGHT TONICHT
>
> First Trevor hauls out with his Langford-Smith, the voice of God, From
> the RDH 4.
>
>> Section 13.1
>> Section 13.2
>> Section 13.4 (Not entirely necessary for this discussion)
>> Section 13.5 (note the following words on page 574: "Push-pull operation
>> tends always to reduce the effects of hum in either the grid bias or
>> plate
>> supply voltage.")
>> Section 13.6 (Not strictly for discussion, because they speak of
>> non-Triode
>> amplifiers)
>
> Andre is on the canvas. You can't argue with a series of hammerblows
> like that. The ref makes him take a count of eight.
>
> Snips, which can be seen other threads. Trevor next asks us to
>> address the following points I made about SE(T) amplifiers:
>
>> ---
>> * ALL SE amps suffer from even order harmonic distortion, which is
>> automatically reduced by using push pull topology. IOW: All things being
>> approximately equal (same output valves, high quality iron, good power
>> supply, same bias current, etc) push pull will outperform SE.
>
> True.

**Good to see you are displaying the kind of honesty not present in some
others I've discussed this topic with.

>
>> * ALL SE amps suffer appallingly bad load tolerance. IOW: A 20 SE amp (at
>> or
>> near clipping) will deliver 10 Watts @ 4 Ohms, 5 Watts @ 2 Ohms and so
>> on.
>> Unless the user has an almost resistive load, then severe power problems
>> can
>> be expected. This problem can be eliminated by using push pull topology.
>
> True.

**Good to see you are displaying the kind of honesty not present in some
others I've discussed this topic with.

>
>> * SE amps are MUCH less efficient that a similar power PP amp.
>
> Do we even need to say it?

**Apparently we do. Patrick Turner has proclaimed that none of what I wrote
is correct.

>
>> * SE amps, generally, exhibit higher levels of hum and noise than PP
>> amps.
>
> Possibly true for the SE amps built by those you know, Trevor, or even
> some kind of a majority which can at a stretch permit you use the
> description "generally".

**I used the term "generally", since it is possible to build a crappy PP amp
and a relatively good SE(T) amp, where the PP amp is inferior. However, all
things being equal (or as equal as can be, PP will be superior).

>
>> * SE amps have a lower damping factor than a similar PP amp. This may
>> lead
>> to audible frequency response problems, within the audio range.
>
> Also true, given that you add word generally to the first sentence.
> "Similar" is really a copout. Nobody who starts work on a SET amp has
> the constraints (usually of costs and a high power level) which
> generally hogtie PP design.

**One must compare apples with apples.

>
>> --
>
> Andre is staggering. His corner fingers the towel nervously. Wilson is
> prancing around, his fists already rising to take the victory on a
> technical (hee-hee) knockout.
>
> The truth is, Trevor, that my personal taste and the winner in blind
> tests with the musicians (still unspecified, if you don't mind) I like
> to use,

**I'll tackle this point, right here and now. The mere fact that a person is
a musician cuts no mustard with anyone. A musician, depending on his/her
particular talent, has some excellent abilities. The ability to judge how
realistic a sound reproduction system is (the goal of high fidelity) not one
of them. Dexterity, the ability to follow music sheets and other factors are
relevant for musicians.

Anecdote:

A few years ago, I attended the sound system owned by one of Australia's
foremost, most respected classical musicians. The man was not old. He is
also one of the nation's best conductors. That puts him way ahead of the
vast majority of musos, given he actually hears the entire orchestra, rather
than listening only to the (say) squeaking of the first violins. Most musos
only get to listen to their own instruments clearly. Anyway, I walked in,
put on an unfamiliar disk and was instantly struck by the obvious problems
in the system. First and foremost, one of the speakers was out of phase.
After rectifying this fault, I asked the man what he thought of the
(considerable) improvement. His response: "I don't hear any difference".

Hmmmmm..

One of my old school mates played in the Sydney Symphony Orchestra's First
Violins. He listens to his system with the tone controls (??!!!) turned all
the way up.

Don't talk to me about musicians.

as well as other qualified persons, is for small Class A
> push-pull trioded pentodes operated at low power with zero negative
> feedback.

**Good for them. I'm not disucssing preference. I'm discussing high
fidelity. IE: What sounds closer to the real event.

The best amp I ever designed is my T113 Class A PP EL34 in
> triode mode (its switchable, actually) with zero or very little
> negative feedback (also tunable in my own copy).

**I have no issue with PP, of any persuasion. PP eliminates or reduces most
of the problems associated with SE(T).

That doesn't mean I
> abhor SET or solid state. I have and use both, too. Nor does the choice
> necessarily have anything to do with your reasons above. We shall
> return to those in a moment.
>
> At this point Trevor starts taking blows.

**Oh, dream on.

In another thread Trevor
> wrote:
>
>> **Why else would any sane designer choose SE(T) over PP? SE(T) offers no
>> advantages. None whatsoever. PP is superior in every way.
>
> Aha. That's a different story:
>
> 1. Commercial reasons. A Japanese manufacturer I have worked with
> several times wanted a PSE 300B for about 16W. I wanted to give him
> trioded pentode PP for a better sound. His wife killed the discussion:
> We could never sell the better amp for half of what we can charge for
> the double-300B. I do as I am told. Herself has spoken.
>
> 2. Just because Andy Cowley is an unnecessarily abrasive ***
> doesn't mean he is always wrong. He gave us a perfect reason for SET
> amps the other day: their transfer curve and interaction with the right
> speakers play right into the aural condition of the men who can afford
> them, who are largely middle-aged.
>
> 3. Well, actually, PP isn't superior in every way for every purpose.
> There are niches, and SET suits some of them very well indeed, thank
> you. Let's look again at your list of the shortcomings of SET, which I
> have admitted above, and see why none of those factors really matter in
> the end result.
>
> Now Andre is off the ropes, swinging. This is the same set of facts
> Trevor presented above, and Andre agreed to, seen in a new light:
>
> Andre offers Trevor a handicap: You can assume the worst condition for
> SE (zero negative feedback) and the best condition for PP (all the
> negative feedback even an incompetent designer may want), and I'll
> still whip your ass.

**Nope. If your hypothetical SE(T) amp uses no Global NFB, then so too does
the hypothetical PP amp. Same output tubes, same Class A bias, similarly
sized power supply, etc, etc. Let's compare apples with apples, not apples
with bricks.

Remember, Trevor, this isn't about which is better
> on the bench, this is about the rational right of audiophiles as
> listeners to choose their own topology.

**You mean: "The irrational right of some nut-jobs to promote dodgy
technology to those who know no better."

>
>> ---
>> * ALL SE amps suffer from even order harmonic distortion, which is
>> automatically reduced by using push pull topology. IOW: All things being
>> approximately equal (same output valves, high quality iron, good power
>> supply, same bias current, etc) push pull will outperform SE [on
>> distortion].
>
> There is no problem making an SE amp as flat as necessary even without
> NFB.

**Yes, there is.

> You just choose sensitive speakers

**Hang on a sec. You've just shifted the goalposts. You're saying that SE(T)
amps are OK, but you must use certain speakers. I presume you're going to
tell me that they need to provide a perfectly resistive load too? Don't
forget, that pesky power problem associated with SE(T) amps.

and then load up the impedance
> on the plate until its response is flat; all it costs is power turned
> into heat. At this point you discover that the PP amp with its feedback
> has disturbing odd harmonics. The argument isn't about added euphonics
> in the SE amp, it is that the wretched NFB has drained some glee from
> the PP amp.

**There is no necessity for a PP amp to use Global NFB. If you want to
compare non-Global NFB SE(T) amps, then compare them with a similarly
configured PP amp.


>
>> * ALL SE amps suffer appallingly bad load tolerance. IOW: A 20 SE amp (at
>> or
>> near clipping) will deliver 10 Watts @ 4 Ohms, 5 Watts @ 2 Ohms and so
>> on.
>> Unless the user has an almost resistive load, then severe power problems
>> can
>> be expected. This problem can be eliminated by using push pull topology.
>
> True for those so thick they apply solid state paradigms to tube amp
> design.

**Nope. True for ALL SE amps. Every one. It's just that the distortion
levels of SET amps (and most others which eschew the use of Global NFB)
increases distortion somewhat gradually. The end result is still the same.
Halve the load impedance and halve the power. Excatly the reverse of a
theoretically 'perfect amplifier' and far from what almost any PP amp (even
those which use no Global NFB) will do. IOW: SE(T) amps are about as
anti-high fidelity as it is possible to get.

The solution is sensitive speakers and to design the amp so
> that it never comes within a hundred nautical miles of clipping.

**Cop out. It is desirable to prevent any amplifier from clipping.

A 40W
> Pdmax 300B loaded to give under 4W of very flat power with a signal

**Which becomes 2 Watts, when the load is 4 Ohms and 1 Watt, when the load
falls to 2 Ohms.


> swing that never comes within some tens of volts from grid current,
> driving a speaker that will never demand a whole watt, is so far
> removed from this theoretical problem, we don't even discuss it much.
> It is only when a solid designer transferring in, or a newbie who still
> considers power a benefit blows in that we need to explain something
> this ingrained.
>
>> * SE amps are MUCH less efficient that a similar power PP amp.
>
> Extravagantly true. So what? Hedonism has a price.

**I was thinking that ignorance has it's price.

>
>> * SE amps, generally, exhibit higher levels of hum and noise than PP
>> amps.
>
> Quite untrue in my experience.

**Go read the RDH4. It's all there.

There is no intrinsic reason an SE amp
> built by someone who knows that it won't benefit from the noise
> rejection PP amps have as birthright, should not be so designed as to
> minimize the disadvantage. I have reviewed several American and Far
> Eastern PP amps that have more him and general *** than any of my SE
> amps, and the price difference wasn't so huge in all cases (I admit, in
> most the price difference was large)

**There you go again: Comparing apples with bricks.

>
>> * SE amps have a lower damping factor than a similar PP amp. This may
>> lead
>> to audible frequency response problems, within the audio range.
>
> I repeat: Also true, given that you add word generally to the first
> sentence. "Similar" is really a copout. Nobody who starts work on a SET
> amp has the constraints (usually of costs and a high power level) which
> generally hogtie PP design.

**All one needs to do is to put as much effort into a PP design that one
puts into the SET design and the superiority will be blindingly obvious.

>
> I'm going to leave Patrick argue this one with you. Didn't we have some
> private correspondence about this late last year, Patrick? Look it up:
> I thought you were admirably concise and to the point.

**Patrick lacks the balls to argue with me.

>
>> --
>
> There are some unique advantages to SET amps. One important one is that
> they naturally operate in Class A

**Er, no. They MUST operate in Class A. MUST, MUST MUST. BIG diference. A PP
deisgner can choose any Class of operation, from Class B all the way to pure
Class A. Moreover, since the efficiency level is significantly higher, more
Class A can be run, so that operating into lower impedance loads still
affords some Class A operation, if that is desired. The SE(T) designer is
stuck with whatever he/she can manage.


with only natural feedback (internal
> to tube as contrasted with added NFB), whereas the temptation in PP
> amps is always more power in Class A/B which never sounds quite as
> pleasing.

**Again: If you want to compare zero Global NFB designs, then do so. Don't
compare apples with bricks.

>
> I hope I have demonstrated that SET amplifiers are a perfectly
> legitimate choice

**Only for the ignorant.

even though I agree with you that a PP amp (strictly
> of the Class A1, ZNFB provenance, of course) will always be a
> technically better amp.

**Thank you. THAT is my point.

>
> The scorers are at work...

**I suspect I know what the result will be.


--
Trevor Wilson
www.rageaudio.com.au


.

Loading