Klopfenstein taper

Also "kino" is occasionally used as slang in the US to mean light sexual foreplay that occurs in public rather than private, not sure how that happened.

The visual equivalent of the 'earworm'?

My go-to solution for wideband impedance transformation would be Guanella transformers. A bandwidth of 5 decades of frequency should be quite easy, six if you're careful.

Jeroen Belleman

It occurs to me that one can use scaling* of an exponential taper to vary smoothly between ~ linear to sharply exponential, so calculus of variations could be used to find the optimal taper for a specified length.

Joe

"Transformer" is probably the wrong word, the "binomial transformer" in this context is a stepped-impedance microstrip structure on a PCB.

Easier manufacturability than something with magnetics, but I get the sense with Mr. Larkin's products they have to think less about unit costs than I do.

The steep taper is non-ideal. And a real PCB would certainly be better.

I got some new 32 mil copperclad

I measured surface resistivity and got about 1.6 mohms/square, which works out to about 1/3 oz copper. I didn't know that anybody made 1/3 oz copper.

Reminds me of the tinned "copper" wire that Amazon sells.

I doubt that rough edges make any difference at this time scale. My ultimate target will be maybe 2 or 3 GHz equivalent bandwidth, which is pretty slow.

I sometimes x-acto trim rough edges for better cosmetics. It's really easy on this 1/3 oz copper.

I need some gold plated 32-mil 1/2 oz copperclad.

We just bought a small brake/shear for our little shop, for slicing up bulk FR4 and aluminum.

There's certainly no reason to design or wind transformers when a PCB pattern will do.

We're now designing some PCBs as planar transformer windings, sort of the same idea.

That's been around for years. It's hard to get enough copper into the winding space with conventional printed circuit boards, but there are people who specialise in the business who can put very thick copper plating onto very thin substrates and get close to the sort of copper fill factor you can get with wound coils.

Some of them can get very narrow deep gaps between adjacent tracks.

About fifteen years ago I went to an industry exhibition in Utrecht in the Netherlands and talked to people at a couple of booths. It was educational.

Think they have manufacturing problems, check out this planar delay line they stuffed inside an HP182C:

(It seems to have had a manufacturing defect)

Anyway, the point was a stepped "ziggurat" taper can probably work for a (bipolar) pulse, the math is simpler and you don't really need a high pass response for that

Thin copper is fine. Even 1 oz copper will be severely skin depth limited.

What did you wind up building?

Skin depth is defined by frequency. Go to a Litz wire supplier for a plot.

Transformers operate at all sort of frequencies, often limited by the core materials available. Manganese zinc ferrites tend to limit you to below 100kHz, nickel zinc ferrite are used up to a couple of MHz.

Nothing. I just wanted to know what was commercially practical. If I had actually gotten stuck with having to get something built I'd have done a lot more work.

What if you reverse in and out?