Rerurn to Romy the Cat's Site

Hello chaps,

Another option thrown in there by Bud; always enjoy your posts even if I don't follow them 100%...
Let me attempt to barbarically distil what you said into 2 aspects which are

1) The dielectric properties of the insulating materials
2) Magnetic properties and distribution of the core

Now, as regards (1), I can well imagine that mylar and other dielectrics suck, particularly if polar (not sure which are). The ultimate aircore would consist of self-supporting coils stacked vertically above each other, with the minimum of glue or whatever to hold their shape. This would almost eliminate dielectric effects, simply because there aren't any dielectrics there (other than the insulation on the coil wire).
I know you are fond of litz wire and paper etc, but I don't see how any conventional cored construction could compete in this regard.

For (2), air does away with core non-linearity, magnetostriction, eddy current losses, hysteresis losses, memory affects and other distortions. Ok, so as you suggest, you get "non-ideal" inductance with turns behaviour, but I wonder what detrimement to peformance that would translate to, when you gain so much more.

Let's not forget here that Romy is talking about 1kHz and upwards and using (half) a 6C33, which is a very nice load impedance of around 200 ohm, maybe less. So, even 1H primary impedance is plenty, and the half-section use means half the normal standing current.

My aircores, which have only seen use on the test bench so far have a mere 0.25H of inductance, an enormous ratio of 90:1 (for driving a 0.5 ohm ribbon tweeter array, ribbon matching transformers removed) yet measure something like 600 to 40k or so (-3db) when driven from 600 ohms.

I'm just getting some suitably non-metallic, non-magnetic chassis material to try and lash up an actual amp around them, although the ribbon array ain't ready yet.

As an aside:

I've also played with 100% mumetal devices from Sowter: gapped for 135mA, Lpri about 1.5 or 2H, power handling spec'd from 1k up. Despite singing along with the signal, they sound absolutely glorious (interestingly, I noted similar comments from Colloms in his review of the AN UK "Ongaku" which uses permalloy transformers; not that I pay much attention to what them reviewers say).

If my Sowters weren't wound for 3k:8, which is completely wrong for Romy's amps (they'd work but give about 100 time less power), I'd send a pair to Boston for the competition...

Would be no problem for Brian to whip up a design suitable for the S2 Melquiades amp, would be interesting to see how a mumetal core and traditional winding methods would compare with the O-Netics approach.

Still, I'd be tempted to go with the aircore route and that's what I'm pursuing for myself. The difference between M6 and mumetal OPTs in an old lash up amp was not subtle, so I have to wonder how much damage even the mighty mumetal is doing to the signal given that it is sucking a fair amount of energy out of it, and that in a presumably non-linear fashion.

cheers,
cv
(Bud - you may know me as shinebox...)

Rerurn to Romy the Cat's Site