[JN] magik ???

Philip Yates toobman57 at gmail.com
Mon Feb 1 19:16:53 CST 2021

Oh, this is both interesting and scary, from a recording engineer.



On Mon, Feb 1, 2021 at 7:14 PM Philip Yates <toobman57 at gmail.com> wrote:

> On Mon, Feb 1, 2021 at 4:59 PM Christian Rintelen <christian at rintelen.ch>
> wrote:
>> AFAIK peerless used the pinstripe design only for large signal i.e.
>> output transformers (i’m not even sure whether they only used it for SE
>> OPTs). if you have an airgap to avoid saturating the core, the inductance
>> drops like a rock. to make up for some of that loss, the pinstripe makes
>> perfectly sense — the airgap prevents saturating the permalloy (aka
>> mu-metal) too fast, and the SiFe gets a jump start so to say. but applying
>> the same principle to a microphone or phono cartridge transformer doesn’t
>> make sense. first, because there is no DC to saturate the core. second, the
>> signal voltage is so small that it will never saturate even a pure nickel
>> core. third, these types of transformers were the first toroids as soon as
>> these became available. and why would anybody roll SiFe and permalloy into
>> a toroid??
> Well the question was why do the Magik trannys sound so good, and my WAG
> was that they used two materials, not in a "mix" configuration (like
> bypassing a cap) but in a "crossover" configuration. With or without a gap
> -- and almost everyone in the past used a gap to linearize the mu, although
> Vitroperm 500f and a few other new materials have very linear mus --
> reversing the current by a small amount puts the core into a "hysteresis
> zone" where the mu drops to almost nothing. You say that Peerless only used
> pinstriping in their output trannys, and that it doesn't make sense to use
> Si-Fe + permalloy in small signal trannies. True. But before drawing
> conclusions, we should ALWAYS look at a situation from as many viewpoints
> as we can see. In this case, they used Si-Fe for output trannys in part
> because a permalloy tranny would have been ungodly expensive (and bigger).
> From that point of view, we might come to the conclusion that they used
> pinstriping because an affordable output tranny had to use a magnetically
> inferior core, unlike a small signal tranny, not because of the presence of
> DC or an air-gap. And I'm pretty certain that I never suggested that the
> Magik trannys either used Si-Fe cores, or should use them.
>> core material tecchnology has advanced quickly and with big leaps fueled
>> by (among others) the demand for «faster» core material in switching power
>> supplies and regulators. i think the peerless engineers would be pinstriped
>> with envy if they saw what materials are available today. and i’m equally
>> convinced they would have switched to toroids instead of taking the scenic
>> detour with pinstriped EI or M cores.
> True, but again, irrelevant to the (possibly incorrect and/or useless)
> point I was aiming at. None of the amorphous/nano stuff to my knowledge has
> the small hysteresis of even Sendust, lat alone mumetal, permalloy, or the
> best ever (to my knowledge), supermalloy. This means that very low level
> transients will fall into the hysteresis zone, just like with Si-Fe. With
> luck and a lot of work, at least *one* of the good materials, meaning
> materials with very low hysteresis, ranging from the amorphous/nano stuff
> to Sendust to the permalloys (and maybe something else), will have a curve
> leading *out* of their hysteresis zone that is at least an adequate
> inverse of the saturation curve of supermalloy, thereby allowing the
> supermalloy to add another 10 dB to 20 dB of resolution to the tranny. This
> is technically equivalent to a boost in the dynamic range, although that
> assumes that the trannys are always being used such that the peaks just
> begin to saturate the tranny.
> You're also missing another point, which is that if this trick works with
> one of the Sendust alloys, then we could have *very* good, very
> INEXPENSIVE signal trannys that would not only make DIY amps and preamps a
> lot easier to design and make in general, but would also allow the use of
> circuits that are all but impossible without a signal tranny. Remember, I
> actually TRY to think "outside the box." ;-) Also, at some point, if a
> signal tranny's dynamic range ever exceeds what we can hear, then except
> for the "character" added by the tranny (perhaps from the wire), a tranny
> that has a larger dynamic range, and is MUCH more expensive, will add no
> audible improvement whatsoever. Well, unless several are used in a row
> (this obviously assumes "all else being equal," e.g., frequency response).
>> (btw i remember reading a paper by sowter senior (brian sowter’s father)
>> advocating airgapped cores even for PP-OPTs (and we’re talking post-war
>> cheap english steel, not peremalloy or that kind of stuff). his reasoning
>> was that the airgap flattened the inductance and this the impedance jumps
>> of the transformer; he’d rather use a bigger core and put more wire on it
>> than dispose of the airgap. which of course makes perfect sense as there is
>> (or was at the time) no PP-amp that had perfectly symmetrical current draw
>> through both windings at all levels and frequencies. even small AC or DC
>> imbalancies can drive an ungapped PP OPT into saturation so you needed
>> either an oversized core (expensive after WWII!) or a smaller core with an
>> airgap. so why not take the big core and apply a small airgap? it will give
>> you headroom re: DC saturation *and* a flatter inductance curve)
>> on the other hand, there is absolutely no reason to use any of these high
>> tech materials for a mains transformer. quite the contrary — you want to
>> keep the bandwith of a 50/60 Hz core as limited as possible so it acts as a
>> filter.
> Now *that's* a good point!
>> ((hold the press: i just localized sowter’s AES paper. enjoy.))
> Excellent, thank you, I have already read part of it, and will definitely
> keep this!
> Phil
>> ©
>> _ _ _
>> aka christian rintelen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.soundlist.org/pipermail/sound/attachments/20210201/a6fe98e6/attachment.htm>

More information about the Sound mailing list