5 to 10kV rated gate transformers

May 17, 2017 79 Replies

On a sunny day (Wed, 31 May 2017 15:57:39 -0700) it happened John Larkin wrote in :

Surface size depends on number of turns, current, frequency, do you need some core material?, board surface outside needs some cover (moisture in air high relative humidity) else it is just a big capacitor to air, would not trust it at 18 kV for a microsecond. You would need to cover it with several mm of some isolating stuff, not just 'solder screen'. Yes I have read that paper, not much new we did not know.

Go optical.

Also take a look at two sided Teflon (PTFE) boards, these can also sustain a lot of punishment.

?:

V

Yup agreed. Not so expensive nowadays regarding insulation performance. FR4 becomes a sponge with age ...

Habib

I assume that Win needs to couple the gate drive power as well as the signal. Optical is fine for signal, but an expensive way to couple much power.

Anything that's 10KV can pick up dust and make corona, PCB or whatever.

I think one could transfer ballpark one watt through a PCB in roughly a square inch without a core. That's a guess.

John Larkin Highland Technology, Inc lunatic fringe electronics

Kapton is rated something outrageous like 7KV per mil. Coilcraft and others make cool kapton-pcb transformers.

John Larkin Highland Technology, Inc lunatic fringe electronics

On a sunny day (Thu, 01 Jun 2017 07:13:56 -0700) it happened John Larkin wrote in :

True, like I said, big Edison style bulb, small solar cells.

Oh much more,

100 W should be no problem, use a planar coil for the top one that works even better, like this:
formatting link

Your traces need to be thick enough and the coil area large enough, that is all.

of kV

My

to air,

Use a 4 layer board. Primary side on layer 2, secondary side on layer 3. No copper in layer 1 and 4 (isolation to surrondings). Big Z distance from 2 to 3, small distance from 2 to 1 and 3 to 4

Cheers

Klaus

I have seen a coreless spiral PCB transformer rated at 50W. But, it's kind off big (about 40mm diameter). Operation at 8MHz

Cheers

Klaus

One example:

formatting link

These days two fibres, a 1 W laser feeding power up to a device sitting on a 400 kV power pole conductor and an other fiber carrying the data down.

On a sunny day (Thu, 1 Jun 2017 07:33:55 -0700 (PDT)) it happened Klaus Kragelund wrote in :

Would not a solid copper layer short your magnetic field? It sticks at a right angle out of the board!

On a sunny day (Thu, 01 Jun 2017 17:44:57 +0300) it happened snipped-for-privacy@downunder.com wrote in :

Cool, way to go!

What kind of PD is generating DC power from laser's one watt? Need p/n please!

Thanks, - Win

One of these one each side of a board

formatting link

formatting link

would be interesting. Resonate both.

You could also use a couple of pot core halves.

The other fun thing would be a flex kapton transformer. The thin dielectric would allow pretty tight mag coupling between sides. A flat ferrite disk, on one side or both, should help but not require punching any holes.

John Larkin Highland Technology, Inc lunatic fringe electronics

The vias are tricky in a high-voltage application, getting the winding ends out. A wire jumper, or some long surface-mount equivalent, would help. Like, a really long zero-ohm jumper.

John Larkin Highland Technology, Inc lunatic fringe electronics

On a sunny day (Thu, 01 Jun 2017 08:45:10 -0700) it happened John Larkin wrote in :

Did we not discus that here years? ago? I think I still have one of the bottom ones.. Maybe I'l give it a try if I can find 2.

That is good stuff, used for rotary transformers, the air gap (PCB thickness) should be low though.

It does not have to be Teflon. Remember the old CRT TV sets? Their flyback line transformer secondary was often insulated to an operating voltage of more than 10kV. Operating voltage, not just breakdown test voltage. The cable used for the output was also typically rated at that because it could come to rest on the plated and grounded back of the CRT or the chassis, or both.

There are many ways to build this gate transformer, it's just that there aren't many companies still able to do such stuff. One that comes to mind that might be is Vacuumschmelze in Germany but it would be a custom part.

Regards, Joerg http://www.analogconsultants.com/

Joerg,

What does not have to be Teflon or Kapton ? We are speaking about PCB.

The transfomer you are reffering to was for accelerating electron beam in the ancient vaccum tube. 10Kv did not transit on PCB. AFAIK.

Yeah right ! here Atesys

formatting link
has a huge experience on high or very high voltage Transformers/Capacitors (No partial discharge, no Tang Delta derivation over years !!).

I designed many high voltage instruments with their top class products with high competitive costs ;-)

Habib

I think Win meant regular transformers although he'll probably be open to PCB versions.

In the color units it sometimes did. They had Cockroft-Walton multipliers on the HV-side of the transformer and those were usually PCB-mounted just like all the primary side stuff. The voltage into the cascade was often "only" 10kV, on the PCB, and then stepped up to 30kV. The step-up circuitry past 10kV did not have PCB contact.

They had long air slots in the board, something I do on HV designs a lot as well. I have even seen this done on phenolic board.

They don't have pulse transformers in the standard repertoire but a company like that probably has the know-how to also build those.

Regards, Joerg http://www.analogconsultants.com/

Ever heard of burried vias?

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required