Re: Supersizing SET


"Trevor Wilson" <trevor@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:44d63458$0$31463$88260bb3@xxxxxxxxxxxxxxxxxxxx
:
: "Ruud Broens" <broensr@xxxxxxxxxx> wrote in message
: news:44d5fa8f$0$6999$dbd4d001@xxxxxxxxxxxxxxxxxx
: > ..................
: > you have fallen into the steady state thinking trap. sure, one can measure
: > thd, imd, phase & frequency response, at certain voltage levels, and to
: > some
: > extent, that characterizes the amplifier.
:
: **Indeed. In fact, if an amplifier delivers inaudible levels of faults, then
: it can be considered as possibly faultless. If an amplifier exhibits an
: audible flaw (as characterised by any one measurement) then that amplifier
: can be assumed to be insinuating it's own distortions into the signal chain.
: IOW: Listeners are listnening to AMPLIFIERS and not the music.
:
: >
: > speakers, however, are mean loads, not just complex impedances, but
: > time delayed voltage sources as well (the mechanical system stores energy,
: > then gives it back). the proof of the pudding would be measuring the
: > speaker
: > output, in an anechoic room, driven by the amplifier - that's what comes
: > out, eh
: > ?
:
: **No argument from me. It is a classic mistake to assume that speakers
: represent a simple resistive load. As you say, not only speakers exhibit
: resistive, capacitive and inductive characteristics to the amplifier, but
: they act as an an energy generator as well. That enegery MUST be damped
: effectively by the amplifier.
:
: >
: > i once suggested something along those lines to Stewart Pinkerton, where
: > quasi musical digital files could be compared with measured output in
: > digitised form, then listening panels with those amp-speaker combinations
: > in a listening room stating preferences - could lead to interesting
: > correlations.
:
: **Possibly.
:
: >
: > for most listeners, it is hard to detect less than 0.1% distortion in a
: > 400 Hz
: > sinewave under ideal conditions played on headphones. taking into account
: > that in musical signals, various masking effects in perception operate,
: > those
: > 0.1% are thorougly unproblematic. So, depending on speaker sensitivity,
: > listener preference, and for most - acoustical isolation to neighbours
: > :-),
: > 0.5 to 10 W at or below 0.1% distortion for average listening level is
: > fine.
:
: **Agreed. Except for the power level. You cannot say what power level is
: appropriate for all listeners. Some may require considerably more than 10
: Watts. One of my own amplifiers is equipped with a peak power meter and many
: people who listen to my 89dB/2.83V/M speakers listen at levels where the
: music peak at 100+ Watts. The room is around 5 Metres X 4 Metres. In larger
: rooms, considerbaly more power will be required for realistic listening.
:
: >
: > peaks in the musical signal will for a few ms run up the power required
: > and make the amp create more distortion (not to mention what the
: > speaker does), but of course, we don't perceive it as 'now it's 1%, now
: > 5, now 10', our perceptive analyses is just too slow for that, we get a
: > compounded impression of the distortion profile
: >
: > a trumpet, when blown very soft, creates nearly a sinewave output.
: > as it is blown with increasing force, the amplitude increases, but an
: > ever increasing number of upper partials, overtones, are created also
: > at full force, up to some 50 kHz components
: >
: > with many natural sounds, louder is often also 'having increased
: > noise content',
: > heh
: >
: > so that for loudness perception, that is somewhat exchangeable:
: > you can simulate loudness by increasing overtones level
: > i'd suggest that an amplifier with a steadily increasing level of
: > distortion, but with higher order harmonics staying below
: > lower order distortion can suggest with good approximation
: > a far larger power amp, that _then_ will create far larger speaker
: > distortion levels,
: > so, why the better absoluteS ?
:
: **No. It doesn't work that way. The distortion needs to remain below audible
: limits, to be, well, inaudible. Once distortion levels become audible, the
: amplifier is then insinuating it's own character on the music. I will
: readily acknowledge, however, that when _I_ look at amplifier specs, _I_
: tend to prefer one which exhibits a gradually rising level of distortion,
: with increasing power output. At all intended listening levels, that
: distortion needs to remain below the human threshold, however.
:
uhm, i think you're missing the point. i'm proposing a plausible model for
the observation that SET amplifiers, although often underpowered to
create peak levels at low distortion levels or even -at all-
still manage to produce very listenable audio from the speakers.
substituting the right mix and dynamic onset of distortion components
during the peak signals make the acoustical output appear to
'sound very much like the actual higher power
signal on a blameless speaker would sound'

this would be a form of applied psychoacoustical processing ;-)

knowing that we don't very well hear distortion in the lowest octaves,
the model would predict that for real (sub) bass, there is no substitute
for moving air. a sub with a set recommended

heh
: >
: > : **Really? Let's discuss facts then.
: > :
: > : An SE amp which delivers 10 Watts @ 8 Ohms, can only deliver 5 Watts @ 4
: > : Ohms. OTOH, a PP amp of 10 Watts MAY be able to deliver as much as 20
: > watts
: > : @ 4 Ohms. Given the fact that speakers are designed to be driven from
: > : Voltage sources, a PP amp is a much better choice.
: > :
: > hmm, you're doing ss thinking, Trevor, with an SE transformer using a
: > different tap you can still do those 10W in 4 Ohms. if you mean dynamic
: > load indifference, that's a function of the flatness of the loadline and
: > the
: > tubes' dynamic output resistance, subject to further nfb if present.
:
: **No, that is not what I am saying. The maximum output of a SE amplifier is
: dictated by the bias current. The maximum output of a PP amp is not so
: constrained. Halving the load impedance connected to a SE amp results in
: less power output, whilst halving the load impedance connected to a PP amp
: results in more power output. IOW: A PP more closely approaches that of an
: ideal pure Voltage source, whilst the SE amp is reverse.

ok, you meant compliance - sure, pp is much better in that respect (can't go
below
0 mA when tilting an SE's loadline :)


: >
: > with a theoretical perfect voltage source, a 0 ohm dynamic output
: > impedance
: > triode, the 'transfer curves' would be vertical lines,
:
: **Nope. Those curves will be straight lines, but not vertical. Otherwise
: amplification cannot occur.

||vertical is exactly what 0 Ohm Ra dyn means.

: so the same voltage
: > swing would be possible with about _any_ load, as tilting the loadline
: > would
: > give the same intersection voltages (for the graphically minded:)
: > - speakers are not pure resistor loads, the resulting real world
: > elliptical
: > loadlines would pose no problem, though.
:
: **For an ideal source that is correct.
:
: >
: > back in the real world, it is a matter of making choices, compromises, to
: > come close to the perfect voltage source output
:
: **Certainly.
:
: >
: > <>some snippents<>
: > : > Only marginally if class AB1, and not any more efficient if pure class
: > A.
: > :
: > : **I suggest you hit the books again. A SE amp can only be a maximum of
: > 25%
: > : efficient (under ideal conditions), whilst a PP (Class A) amp can be 50%
: > : efficient. Therefore, PP can be more efficient than SE.
: >
: > nope, that would be the case for an ss output stage with a current sink to
: > bias it halfway. a tube amp using a transformer would under ideal
: > conditions
: > reach 50 percent.
:
: **Nope. According to RDH4 13.2 (iii) it is 25%. PP Class A amps can, of
: course reach (theoretically) 50% efficiency. SE cannot.

ahah error found - nope, i'm right, so therefore, RDH mustafbeenwrong.

the calculation shows it, donnit?
but for your ease of mind, here the formula in Terman :
*section 10.5, page 337, formula 10.20, 1955 ed. Electronic and Radio Eng
Plate efficiency = 0.5(1 - Uak_min/Uak_bias)( 1- I_min/ I_bias)

yep, with a theoretical 50 % when v and i swing all the way

: > *didn't i give the 4cx350 as a >40% efficiency possible example?*
: > (that wasn't an april 1st joke, you know :-)
: > (of course, this is neglecting fil power and assuming constant load)

:
: **I further cite RDH4 13.2 (v) where the figure of 25% efficiency is quoted
: again. If you have a cite from RDH4 where the 50% figure (for SE output) is
: quoted, I will look it up.

you've earned the boterbabbelaar of the week for finding an error in RDH :-)
Rudy

: --
: Trevor Wilson
: www.rageaudio.com.au


.

Relevant Pages

  • Re: Supersizing SET
    ... elliptical loadlines may be accomodated with less distortion than with ... Negative Voltage feedback makes such an amplifier ... Except for the power level. ...
    (rec.audio.tubes)
  • Re: Another proposal, and 845-SET amps would convince disbelievers.....
    ... It is that distortion that proponents enjoy. ... or that they were not listening to the music. ... The amp has a 55 watt ability, and distortion lower than many PP amps. ... It is interesting that the amplifier with the best ...
    (rec.audio.tubes)
  • Re: Discussing audio amplifier design -- BJT, discrete
    ... and this is, I think, I great reason to go down this design path (what ... estimating what you could expect for distortion. ... circuit drive requirements, ripple on the power supply and transformer ... voltage to get your required voltage output at full load. ...
    (sci.electronics.basics)
  • Re: stromber-carlson amps
    ... the tube-audio experimenter finds various tube type "public ... How to pick a good candidate amplifier: ... Look for amplifiers in which the output transformer is as large ... so I took the G2 voltage from the center of the ...
    (rec.audio.tubes)
  • Re: Supersizing SET
    ... Negative Voltage feedback makes such an amplifier practically ... it is hard to detect less than 0.1% distortion in a 400 Hz ... a far larger power amp, that _then_ will create far larger speaker ... load indifference, that's a function of the flatness of the loadline and the ...
    (rec.audio.tubes)