I have done some experiments on the transformer, with the initial design, and several different winding arrangements. I didn't have any suitable copper foil in stock, but found some fine expanded-mesh screening that actually worked better for construction, because of its see-through nature. A 110 W soldering iron was just barely enough to bond it to the bushing, by soldering through the mesh to the rim. The turns count was 69, as I thought before, but I managed to squeeze on two more, so got 71 T and about 45 uH. I changed the winding lead access to be parallel to the main axis instead of radial, so assembly was easier. I added some layers of 1 mil Kapton film, in sloppy fashion, just in case I needed to solder on the outside of the mesh wrapping - the film would keep solder from flowing through and sticking to everything. I'll be putting up some photos later showing the results.
I set up a low Z divider with the TG feeding 47 ohms going into four or five 10 ohms in parallel to get around 2 for the primary. A 49.9 ohm on the output, and another on the primary, are used for monitoring those levels with the SA. The transformer exhibited SRF around 50 MHz, so as-was, so could be workable for a 30-40 MHz top end. The problem is that I'd like to get to at least 100, so the inductance was too much. I'd give up some at the low end to get a higher top limit. So, I had to trash the nice looking transformer by having to unwrap and redo it a number of times, for experimenting. Remember, winding this thing was a PITA, so I tried to get as much use as possible. I first took off a bunch of turns to get down to 50, and spread them out and re-wrapped. The SRF went to around 90 MHz, not bad. At an earlier point I tried to make about 33 bifilar turns from the whole winding, but they did not balance. So,after the 50 turn round, I ended up stripping the winding and doing a true bifilar, and managed to get 2x36 turns and a full core. I'm getting better at winding this thing, so it wasn't as bad as the first time. The SRF was right about 100 MHz, still not quite there. I also tried various schemes to cancel displacement currents and such, since there are two "identical" windings, but still didn't move it up much. Finally, I deleted one of the windings, leaving the other nicely spaced out on the core. This time it was around 110 MHz, as long as the wrapping isn't too tight.
During all the experimenting I could see how significantly the "foil wrapping" that I've been advocating for low secondary leakage inductance, affects the primary winding capacitance and SRF. So, the foil tends to foil the high frequency performance that I want - always another trade-off. I have some more experimenting to do yet, but I think I'm close to the right turns count, and will figure out a way to not wrap it too tightly, so I can get a good margin above 100 MHz operation. Since the now lower turns count has less impedance reduction, I plan to make a very low Z driver amplifier, somewhere around a half ohm, if things work out. More about that later.
So anyway, I still like the foil wrap, if the goal is in the medium frequency range, but it gets tricky when you want to push higher. My problem here is that I have only this one core type and size, that happens to fit just right on the post as-is, so the winding on the inside layer is very tight, adding quite a bit of C to begin with, then aggravated by being quite a large core, and the nice tight foil wrap that would keep the secondary leakage down. I've seen a number of designs others have done, in the discussions - there are plenty of ways to go, and use whatever works.
Next time, transformer pictures and amplifier stuff.
Ed
