Suggestions for sources of gate transformers, rated for 5 to 10kV continuous DC. Or actually, small 1:1 or 1:2 transformers to make simple dc-dc converters for
Will need 100 to 200 pieces, or more. Parts test rated for 5kV for one second, or even for 1 minute, are not acceptable. Maybe parts rated for 15kV for one minute would be good enough. I can design and hand-wind transformers to meet this spec, but sheesh, I don't want to wind 200 or more transformers!
Digikey has cheap DC/DC converters in the 6 to 10 KV range. But they are mostly tested for one minute. You can't expect a Digikey part to be tested for a year.
You might make the transformer as a PC board thing. That is sort of free.
FWIW, I've hi-potted one small common mode choke that failed at 15kV -- it jumped pin to pin, not over the bobbin separator fin.
Slow for gate drive, but maybe that's okay, I don't know. :)
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website:
You got my curiousity....
What is a gate transformer used for? Those are some high voltages you have there. Either you're powering some hefty transmitter tubes or powering a neon sign. But it's probably some commercial device that I never heard of.
It's for an AC bidirectional 10 kV MOSFET on/off switch, with about 14 ohms of ON resistance, and low off leakage. Each switch uses seven stages of 1.7kV SiC MOSFETs, and we need 10 switches, that's 70 high-voltage gate drivers.
Spiral turns on top and bottom layer of PCB. Prepeg can handle 10's of kV
Cheers
Klaus
On a sunny day (18 May 2017 10:07:39 -0700) it happened Winfield Hill wrote in :
The Schaffner pulse trafos I have used (to drive HV thyristors) go to only 3 kV. A couple of month ago I was playing with optical fibers. You know LED on one side, there do exist photodetector high output chips. Depends also on how fast, and if you have any voltage on the other side to drive a special detector. If not, then [for example] a HUGE light bulb and small photo-cells to power the optical RX chips. There are many ways to skin a cat / polar bear, what not.
Have you considered fibre optics.?
If some power is required at elevated potentials, a laser and a fibre is sometimes used to power small devices sitting on a 400 kV AC line. An other fiber is used to bring any measurement results back to ground potential.
Do the SiC fets have equivalent reverse substrate diodes? Is there an avalanche damage hazard in the series string?
If it doesn't have to be fast, maybe some sort of PV opto gate driver would work. Fiber or free-space.
I still want to do something with a PCB transformer. Resonate a planar pcb coil on one side, drive hard; resonate another coil on the other side, rectify.
Or maybe use surface-mount unshielded ferrite drum inductors on opposite sides of a PCB, as a dc/dc converter or as the actual coupler. That would be fun.
Yes, I had been evaluating CM chokes, which often have quite impressive insulating distances. 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 and charge the bulk cap) is a modest 66mA p-p.
My next step, play with potting and a month-long 2x overvoltage test at 10kV.
I haven't seen any above the sometimes-seen 3750 volts. Examples?
HV parts AC tested for the standard one minute have to be dramatically derated for continuous DC. To account for the breakdown paths that develop over time.
Recom has 2012 whitepaper "isolated DC-DC Converters" with a derating table, they knock off up to 50%.
In all cases where I needed something similar we ended up making our own. Starting with aerospace wire that has certified insulation properties, where every foot of it is tested, and so on. Tensolite-Carlisle was usually the supplier. If it absolutely has to be COTS maybe talk to these guys:
What does that you mean DC insulation refering a magnetic based transformer ? Please do not be offended about my trivial question ... not so trivial indeed !
Habib
Den onsdag den 31. maj 2017 kl. 21.50.48 UTC+2 skrev habib:
nuous DC.
one side of the transformer floating on 5-10kV DC with respect to the other I assume
Klaus,
Are you sure PCB prepreg could sustain "10's of kV" ? I'm sure not ! My experience of that sort of things (10 or 15KV/AC 60 or 50Hz for 1 minute) is very far from a standard ... very far from !!
Habib.
Ok then ! Please consider standard materials to sustain 10kV DC for 1 minute. My experience on that topic is that only custom transformers with Vaccum impregnation and Teflon insulators could do the trick ... may be.
Habib.
Den onsdag den 31. maj 2017 kl. 22.04.11 UTC+2 skrev habib:
kV
haven't read it
Klaus,
No I have not read it yet, I will of course tomorrow morning at work. Mat be I'll learn something useful. Thx.
Habib.
He says that 800 v/mil is a typical spec for FR4, and 300 v/mil is conservative for long-term use. That would be 300*62 = 18.6 KV side-to-side on a standard 62 mil board.
Sounds good to me, a free transformer, no winding required.
Have something to add? Share your thoughts — no account required.
Ask the community — no account required