Hi Ken,I expect that the orientation [co-planar, co-axial, co-incident] of the primaries shouldn't matter; ultimately it will come down to the actual coupling values between all of the windings involved.
In your simulations of the parallel primary windings, you'll probably arrive at optimal values for these inter-winding couplings, which in turn would suggest how far off-axis or out-of-plane to place the windings.
Ease of construction would also direct whether to offset the parallel windings axially, laterally or radially.
Cheers, Greg On 4/15/2015 8:10 AM, Ken Herrick wrote:
Hi, Greg (et al)- Thanks for the comments. A question remains, though: Your comments seem to refer to, and the SLAC 76-primary design shows, a coaxial set of primaries. But I am referring to a set of primaries that are co-planar and not coaxial. That is, in my case, each of the 4 primaries is to occupy 1/4 of the cross-section of the adjacent secondary. And, as I indicated, all 4 are to be driven synchronously, resulting in 4 synchronized flux paths. My reasoning is this (but see next par.): if the primaries were to be of nominally identical shape, coaxial and adjacent, with each driven by the same amount of p-p voltage, the result would deliver no more ampere-turns than if just a single driving electronics were to be used, with a single primary having 1/4 the dc resistance. That's because of transformer action between the primaries. Whereas with, say, 4 primaries that are co-planar but not coaxial, there would be minimal (or at least, markedly reduced) magnetic coupling between them. Each would generate its own flux path through the secondary and it would be the flux paths that would add. The only other way I can see to do it would be as I had done before: daisy-chain multiple driving sources into one equivalent primary coil. But that is more complex since all but 1 of them must float with respect to the one, so the h.v. sources of the other 3 must be choke-isolated. (In any case, of course, transformer-coupling for each l.v. drive is required, to galvanically-isolate the l.v. from the mains.) So...it has just belatedly occurred to me to computer-simulate it; piece of cake! Driving a simple 3-coil transformer from 2 identical sources--one to each of 2 primaries which have respective ("perfect") k of 1 and with the secondary:primary k of 0.95, I get only about 15% difference in the rms secondary currents with one of the drives connected, then unconnected (proving my surmise). But when I reduce the primary:primary k only to 0.9 rather than the 0.95, I get a 2:1 difference, within a few %. Interesting...it takes only a _small reduction_ in that k to bring the current gain right up to 2! So...(once again): which would be best? Stacked, identical and coaxial primaries or co-planar, identical and non-coaxial ones? If the former, it appears necessary that they not be too closely situated. But perhaps just a rather short spacing would result in a respective k less than 0.95. Although...would there not be resultant flux leakage out from between the primaries in that case? I'm getting way too old for this... Ken
_______________________________________________ Tesla mailing list Tesla@xxxxxxxxxx http://www.pupman.com/mailman/listinfo/tesla