Memory distortion, crap or serious?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Maybe my memory fails me, but last I read something about
Lavardin I think the point was that the thermal inertia of
the transistors caused the memory distorsion, which may perhaps
make some sense. However, there seemed to be no mentioning
of themal effects at all in those links Pjotr posted, but rather some
obscure talking about electrons leaving traces of flux in
semiconductors!!!!
 
Yet that looks interesting, I vote total crap.

It's not hard to pull your favorite amplifier from the 60's out and give a listen and notice that it sounds very nice. As long as the caps are still good and resistors haven't drifted (also clean volume pots everyone!) I think the sound should be pretty darn close to original.

Also, I've known many people to prefer the sound of a solid state amplifier after its been broken in for a few years.

Aside from all that, look at basic transistor theory: You have electrons and holes. In the N channel of a SS device, you have atomic lattices with extra covalent electrons. These can only be pushed from one end out to the other, like in a water pipe.

In the P region, there are 'holes'; little deficiencies of electrons in the atom lattice. These are where the electrons from the N region may go on their way out of the PN junction, like in a diode. If you try to revers the polarity of the electron flow, the electrons will not be able to flow from the P region to the N region, and the imaginary 'barrier' will widen, causing no useful current flow.

The holes and electron spots here are simply the outer shells of the atoms. And electron may be pushed from one atom to another as they share excess electrons. There isn't much that's going to go on at the electron level to introduce a detecable amount of music distortion as far as I'm concerned.

If anyone more knowledgeable in the field of quamtum physics would like to describe the actual traits of the valence shells that might introduce electron distortion, please do so, I'd like to learn about it if it's known to actually exist and/or cause a sound problem.
 
Actualy, this is a case where I do believe there is some element of truth behind what they are saying (at least from a totaly technical standpoint), but they are applying it in a totaly irrelevent situation.

Basicaly, what they appear to be talking about is the point where you are switching a current through a diode junction. Before you can be totaly satisfied the device is 'off', a small amount of curent must be flushed through the device to restore the depleeted region which acts as the barrier in the semiconductor. So, if you are trying to switch this very quickly, this curent, and time it take to flow through the juction may become significant.

However, as far as I am aware, this is only an issue when going from a forward to reverse bias state on the junction, and will only take a few pico seconds (or nano at the VERY worst) to take place. So, to say that this will effect the sonic character of an amplifier seems to me to be total BS, as I just can't see it even applies in a fully class a circuit, and for anyother type of circuit (with the possible exception of class D perhaps) the frequency range it will effect will be inaudible and hence of now significance what so ever.

So, although what they are saying does appear to have some thechnical basis, it does indeed appear to be total guff in this case.

Long Live The Audiophile, and his never empty wallet.
 
AX tech editor
Joined 2002
Paid Member
I agree that the Lavardin site gives very little true contents and mostly marketing crap. However, I have the original AES paper by Hephaistos (a respected French collegue of Jean Hiraga). This paper is very to the point with measurements and scope photo's to document the effect. It seems real enough to me. However, from the Lavardin site I doubt whether they even understand it, let alone have mastered it.

My vote: real, but implementation probably crap.

Jan Didden
 
Doug Self has investigated thermal distortion here:

http://www.dself.dsl.pipex.com/ampins/thermald/thermald.htm

It's untrue to claim that thermal distortion is not detectable with steady-state measurements, as the graphs show.

I'm also unconvinced that simulations which show the output DC offset drifting at a few mV/sec after a signal peak actually translate into distortion at audio frequencies.

Cheers
IH
 
IanHarvey said:
Doug Self has investigated thermal distortion here:

http://www.dself.dsl.pipex.com/ampins/thermald/thermald.htm

It's untrue to claim that thermal distortion is not detectable with steady-state measurements, as the graphs show.

It's worth noting that Self's measurements are of a monolithic power opamp, where the low current input and VAS stages are built on the same slab of silicon as the high current output stage.

And as Self notes, "There is absolutely no such effect to be seen in discrete-component power amplifiers."

se
 
Hi all,

Thanks for the replies so far. Although I vote the promo-talk on the links I gave as totally nonsense, a Google search on “memory+distortion+amplifiers” brought up some serious links but mostly related to HF amps.

The talk about electrons leaving traces of flux in semiconductors made me thinking of the slugs that sometimes entering the house from the backyard and leaving twinkling traces on the kitchen floor :D

Thanks JF for the peufeu link. There are 2 links there at the end to patents that holds serious stuff regarding thermal modulation. But if that makes really sense to audible artefacts in SS audio amps? I doubt …

Cheers :)
 
However, as far as I am aware, this is only an issue when going from a forward to reverse bias state on the junction, and will only take a few pico seconds (or nano at the VERY worst) to take place. So, to say that this will effect the sonic character of an amplifier seems to me to be total BS, as I just can't see it even applies in a fully class a circuit, and for anyother type of circuit (with the possible exception of class D perhaps) the frequency range it will effect will be inaudible and hence of now significance what so ever.

Although I'm not qualified to comment on memory distortion per se, the above quote refers specifically to reverse recovery in diodes turning off. In diodes of substantial current ratings on the order of an amp or more, recovery time can easily exceed 100ns, and in standard recovery diodes, can be several microseconds. Similar phenomena can occur in BJTs, but I'm not sure of the time-scale there. It is primarily a large-signal issue, but power amps generally deal with large signals.

John
 
A magnet can induce a "memory", type effect in a piece of steel, due to the magnetic domain hysterisis. Although this is totally unrelated, it is at the same atomic "level", as current flow. I don't know enough about current flow through semicondutor junctions to relate it, just thinking aloud really.
Mark
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.