high voltage inductors

I asked the exact same question years ago and got the exact same answer :-(

You'll have to roll your own. Depending on the app that can go quite far. Many times when I needed inductive parts for class-III medical and such I had to line up a manufacturer that could furnish 100% tested and certified military-grade or aircraft wire.

Carlisle-Tensolite is an excellent source for such wire.

--
Regards, Joerg 

http://www.analogconsultants.com/

Did you check any (former) CRT flyback transformer makers, what they could offer ?

If you could still find one you would have to be able to speak Mandarin :-)

--
Regards, Joerg 

http://www.analogconsultants.com/

I once asked Bourns about the SDR0805

We used it for 450VDC and they changed the standard coil just slightly to conform that voltage

Cheers

Klaus

If you are a big enough customer they'll go through a lot of hoops. But for any lower volumes there's usually only the datasheet which is paltry for most inductors.

I have found XFMRS.com to be very good when it comes to providing engineering data. They even sent me samples from Hong Kong within a day. No idea how they did that.

--
Regards, Joerg 

http://www.analogconsultants.com/

I think they were primarily concerned about corona discharge, allthough I do not know the excact design challenges in high voltage inductors

We were talking 500k/year, so no problem getting the needed attention :-)

Cheers

Klaus

RF folks don't usually use inductors in the mH range.

Minimum PCB trace spacing for 5000V is about one inch. If you series-connect surface-mount parts, like 1812 size, you only get 3mm spacing at each element: it'll take ten in series. 500V rating is plenty, for terminal separation purposes.

But, that assumes you have 5000V terminal-to-terminal (circa 3000 VAC) signal applied? Why? How?

Corona is another interesting limit: if there's actually grounded elements nearby, like a ferrite core, or PCB planes, and you don't want to wind with fat wire, it takes potting or oil-filled construction. One might want to leave the printed circuit board entirely, and mount the HV parts on standoffs. HV standoffs were quite common years back, don't know what parts are still available.

Ten in series might be OK, as opposed to some weird axial part. I can pick-and-place ten 1812s with zero labor. So, I'd need some 1812s good for 500 volts.

It's a high voltage pulse generator. Typical pulse widths will be around 1 microsecond.

We'd keep HV stuff away from ground to keep the gradients down. We migh conformal coat, too. Potting and oil are both nasty.

5KV isn't too bad for corona.

--

John Larkin         Highland Technology, Inc 

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

that they don't rate their inductors for voltage breakdown. I was hoping t o string together a few surface-mount parts.

nnect surface-mount parts, like 1812 size, you only get 3mm spacing at each element: it'll take ten in series. 500V rating is plenty, for terminal s eparation purposes.

ignal applied? Why? How?

s nearby, like a ferrite core, or PCB planes, and you don't want to wind wi th fat wire, it takes potting or oil-filled construction. One might want to leave the printed circuit board entirely, and mount the HV parts on stan doffs. HV standoffs were quite common years back, don't know what parts ar e still available.

Farnell still stock a line of "Press-fit PTFE components" which consist of a pin and a PTFE/Teflon bush, which are what we used to use for high-volta ge/low leakage circuit nodes at Cambridge Instruments.

47797

is one of the cheaper ones. It's a labour-intensive low-volume solution.

--
Bill Sloman, Sydney

Wurth has a line designed specifically designed for high voltage (and rated for that):

Cheers

Klaus

It looks like it will be based upon the inductors mechanical design. From what i can see formvar and other coatings are good for >=5kv @

150c and I think thats single coat.

The series string idea seems to be the best option, as long as you can balance the voltages.

Cheers

It's rated for 400 volts DC. I'm not sure what that means.

I asked them for the WE-STD 1516 doc, which might explain it.

Thanks

John

--

John Larkin         Highland Technology, Inc 

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

It means it is rated for applying 400VDC as long as you keep the other constraints like saturation peak limits etc

Regards

Klaus

Probably, that you can rely on 400VDC insulation between the external surfaces (exclusive of the endcaps) and any loose hookup wire or PCB element that brushes against it. Like, an underpass trace. It sure DOESN'T mean you can safely apply 400VDC between the terminals!

500VDC used to be the nominal minimum magnet-wire insulation. Belden #8058 (Beldsol) datasheet says breakdown is over 1000V. And, if the PCB can withstand applied voltage between the two pads, the presence of intermediate voltages along the length of the inductor means no 'new' stress (assuming, of course, that there's only one winding layer, solenoid-fashion). I suspect that there's lots of us with SMPS needs for voltage-doubled 120VAC (i.e. about 340VDC). So, 400V is a kinda 'sweet spot' for folk working from checklists. Who don't want explanations.

Well, I am quite sure you are wrong on this one. It makes no sense that it's to external surfaces. It's between the terminals, the SDR0805 is an example of a 400V part, voltage between the terminals (the Wurth part also)

Cheers

Klaus

** He is only stating the bleeding obvious. ** The published data ( see the PDFs) says the part is suitable for DC-DC converters that have up to 400VDC input.

Rectified and filtered 230VAC would be one example.

It should be automatic that the insulation will not fail to core, shielding or any external part either under the same condition.

.... Phil

it can't be between the terminals, there's practically a short between the terminals at DC

-Lasse

Not when the inductor is used in a switching application such as a buck or boost converter. And that's then problem, the datasheet does not say if, for example, running this on rectified European mains is going to be kosher.

The boost converter is often the worst case. Assume you run it at 320VDC in order to generate 2000VDC. With the usual incomplete datasheets that would be a white-knuckle ride.

--
Regards, Joerg 

http://www.analogconsultants.com/