48V to 2000V DC-DC smallest design

snipped-for-privacy@highlandsniptechnology.com wrote in news: snipped-for-privacy@4ax.com:

Above 1000V it should be potted and no solder mask on the HV section.

Potting boxes can be obtained or spec'd at just over the profile height of your multiplier section, which should be separated from the rest of that board. That minimizes the section size needing a pot shell slipped over it.

How does potting compare to conformal coating with a silicon based coating spray. Not the thick goopy stuff but more like heavy clearcoat.

With healthy clearances, you don't need any of that stuff at 1KV or so. Coating is messy and potting is very messy.

I need it tiny 60Hz transformers are most definately out. :)

think less than a half of a match box,

I'm really pushing the size limits down so voltage clearances are being pushed. I can slit the board in spots but not everywhere.

I've used confirmal coat for tolerance to surface contamination and moisture in traditional voltage circuit boards but never depended on it for HV insulation. Just looking for experience.

In theory, silicon should hold 500V/mil so 10 mil spacing should be good for 5kV. :) RIGHT!! what would be a realistic number?

You do not need a multiplier, 2 kV is not much. Our HV sources go to

5kV directly out of the flyback (that for the HPGe versions, for NaI we make them up to 1.5kV (more on request)/1mA (also more on request). While the transformer for the 5kV takes a lot of details to know how to make, the 1.5 kV version is more forgiving (not infinitely of course). They regulate from something like 100V up to the maximum, driving them with some choice of a waverorm should be OK within some reasonable BW limits (never tried that though on the netMCA it is a DAC driving the analog input setting the HV value).

Dimiter

====================================================== Dimiter Popoff, TGI

======================================================

I was poking around on the Mouser site looking at their ferrites and found a TDK RM8 core - the B65811J0000R608122092 (Mouser part) using the new PC

200 ferrite, which is good for 300kHz.

RM8 is bigger than half a match box, and you'd need two of them for a Baxan dall inverter, but when I went through the transformer design for 45V to 2k V the turns ratios were low enough that it looked as if you could go direct ly from 45V to 2kV while running at 300kHz. The interwinding capacitance on the secondary wasn't as crippling as I'd feared.

A smaller core might work, but I lost interest at that point.

--
Bill Sloman, Sydney