CCFL transformer

I have looked into that. I cannot find resonable valued transistors at

On top of that, I need a positive pulse, 3.5kV, and also a negative pulse (600V or so). The can be solved with a transformer.

If your application is a CCFL one, then go for it. That's COTS.

If it requires a continuos high voltage, you will need to take extra precautions.

RL

Oh.

Measurement:

Leakage inductance?

Wrong question. What you care about is the coupling between the primary inductance (which you drive) and the secondary inductances (from which you take your output). This can vary from 0.98 to about 0.999 in practical transformers.

The usual way of measuring it is to short one winding and measure the apparent inductance of the other, but this doesn't allow for resistive loses in either winding.

If you know the winding resistances you can correct for it. Sophisticated users presumably do this. I wasn't all that sophisticated when I did it, and didn't bother.

You can document the measured inductance with the other winding shorted as the "leakage inductance" but it creates a misleading impression about what it going on.

And you will write down a much higher leakage inductance if you short a primary with a only a few turns and measure a secondary with a lot of turns than when you go the other way, when both measurements reflect the same mutual coupling.

There's no "reflection" involved. The resonance reflects the oscillating flux in the core, and the parallel capacitances of the primary and secondary windings both get charged up and discharged during the cycle.

The parallel capacitance of the secondary will be higher, and the voltages across it much higher, so it is dominant.

The resonant current is flowing through the capacitances so may not heat the insides of the winding wires.

Measuring the self-heating of a transformer being resonated might be an interesting exercise.

I wrote "reflected", since the inductance on the primary was the measurement. The resonance of the transformer is the same on all windings, if the coupling is reasonable good.

So like you wrote, the secondary is dominant, which is why the primary resonance is due to reflection from the secondary.

I expect that you can't push a decent-looking pulse through a CCFL transformer. Look at the pics on the Coilcraft page. Those things are probably designed to have a huge leakage inductance. That makes sense, since gas tubes are negative resistances.

Of course, we don't know what kind of pulses you need.

Some sort of Marx generator would be fun.

My Pockels Cell driver uses a resonant boost thing with a SiC fet. It goes from 48v to 1400v in one step.

There's no "reflection" involved. Both the primary and the secondary winding take part in the same process, and make their own - more or less independent - contributions.

It's a pedantic point, but getting a proper grip on what going on in transformer does seem to be difficult, and it does take a while. Getting close to precisely the right point of view probably helps.

Of course most of them don't work, and of those that do work quite a few have unexpected gotcha's.

It's two way street. Think first, so that you have some idea where the problems are so, that when you do read you are on the look-out for solutions to those problems, but useful books set you on the track of problems that you hadn't thought of.

The main problem with innovation is that it is a process of exploration, and you need to get as much guidance as you can.

Pasteur's comment was that "chance favours the prepared mind".

I order some similar ones some weeks ago, still waiting for the mail :-)