5 to 10kV rated gate transformers

..

Cockroft-Walton mulplier ; we call it "Latour multplier". Ok a voltage mult iplier on PCB but creepage and clearance distances were considered only on surface and not through the PCB thickness ! (1.6mm for 10KV) E(v/m) = V/d ; V voltage and d distance ... for sure it's painful for a aged FR4 (altho ugh some are consider this comfortable ...).

Right ! not on standard transfomer products but they can do it on custom sp ec.

Habib

Sure. How would you use them to make coils on both sides of a board? Seems to me that you'd give up insulation clearances.

I wrote earlier that I'm evaluating CM chokes. "I settled on a Panasonic ELF-21C023A. It has Lm = 1.7mH, Ll = 40uH and Rcoil = 0.08 ohms. An 100kHz H-bridge running from 24V, with 1N5819 Schottky bridge rectifier, makes a 24.5V output, into 5mA load, 23.6V at 25mA, according to SPICE, for +20,-4V SiC drive. The magnetizing current (discharge / charge the bulk cap) is 66mA p-p. [not bad] My next step, play with potting and a month-long 2x overvoltage test at 10kV."

I'm preparing the potting fixture now. That will make this a custom transformer. :-) The Panasonic transformer looks well suited, with thick bobbin insulating plastic, and generous room for further insulation through potting. Only need to handle 2.5kV to the ferrite core.

Another parameter, the coil-coil capacitance is under 2pF. Amazing, I had budgeted to 10pF.

That does not guarantee that miniscule destruction won't be creeping in. Meaning the transformer is going to be compromised after that month and there is no telling by how much. This is why hipot tests at the end of production are usually limited to a minute or less.

Says "AC 250 Vrms max" which would be the rated maximum working voltage.

Even if you assume the low end of your spec (5kV) you can't assume that the core will always sit at exactly half between the coils. At least not for low frequency AC or DC. Creepage, contamination, capacitance difference and all that can make this voltage divider lopsided.

It might all "work" but in a design review for a commercial product this would be flagged or at least should be.

It can help to bury the coils or at least the HV-side coil one layer down into the board to avoid St.Elmo's fire.

For very good reason. Applying hipot level voltages continuously to a part that is only breakdown-rated at that level will degrade the isolation properties over time.

Pot cores are also very nice devices for HV apps. You can insert gap material that is highly isolating but all across the core halves. Meaning no screw going through them. Just like rotational couplers are made except not air.

With pads, that will have about 15mm of surface creepage distance; on your PCB and on the part itself, unless you bring wires out of your potting or something. Seems labor intensive.

Potting is a sticky, ugly mess. You need to measure, mix, degass, pour, wait, clean up.

I like the idea of drum cores or pot cores on opposite sides of a PCB, but nobody else seems interested. Imagine just a couple of cheap pick-and-place inductors doing the job.

The max dielectric spacing would be with the coils on the outer layers. All the HV-side circuitry is subject to corona, with the flat secondary coil the least likely component to glow in the dark.

AFAIK you can't buy an open half pot core as a standard part. I'd be all ears if you could.

Upright unshielded rod inductors, yes, but their spew around EMI to no end. They are like a dirt bike with the muffler sawed off.

On a sunny day (2 Jun 2017 08:44:20 -0700) it happened Winfield Hill wrote in :

Did you test for flash-over between windings and core at 20 kV?

Blind via from layer 1 to 2, that allows access to coil no 1 Blind via from layer 3 to 4, that allows access to coil no 2

Insulation prepeg from layer 2 to 3. Placement of the vias needs to be considered. Creepage requirements in internal layers are a lot lower than outside. If you cannot life with the design, go to 6 layers, or make one coil as a add-on PCB

Cheers

Klaus

Depends on where it goes. If the FET that it drives is all potted up that would just leave the coil. Unless you also pot that which is always an option.

Then there is the risk of touching. On one of my HV designs I had to be the first one bitten. The Chinese engineer who was with me laughed and said "See, you should have included a warning in German, not just English and Spanish".

When I was doing ignition system designs (~1968), I got blown off of a lab stool onto my butt... producing great laughter from the technicians ;-) ...Jim Thompson

-- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | STV, Queen Creek, AZ 85142 Skype: skypeanalog | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

Thinking outside the box... producing elegant solutions.

"It is not in doing what you like, but in liking what you do that is the secret of happiness." -James Barrie

I put a long crack into a sturdy wood cabinet door (the good stuff, custom-made by a carpenter) when flying off the chair after touching HV.

Seems like a surface-mount jumper, or several, would be an easy way to get the inner end of a planar coil out without complicating the board or compromising insulation.

On a sunny day (Fri, 02 Jun 2017 10:46:06 -0700) it happened Joerg wrote in :

Next one up could .. but you must have read Baron Von Munchausen...

As noted, I hate potting. Sticky, tedious mess.

I did this design, 1400 volts max,

without getting zapped even once. That itself was shocking.

Groan ;-) ...Jim Thompson

Yeah, FR4 would force a bigger magnetic gap than something serious, like kapton or teflon. Still, it wouldn't be hard to push a few watts through 50 or 62 mil thick FR4, good for 15KV or so.

Maybe someone would make a rigid-flex sort of pcb stackup, mostly FR4 but with an inner layer of kapton. That has interesting possibilities for vias and such.