HV insulation test

Winfield Hill wrote in news: snipped-for-privacy@drn.newsguy.com:

Looks like it is partially potted in hard "Stycast" epoxy.

You should encapsulate the rest (windings) as best you can.

Ideally Vacuum encapsulate it.

However, you can simply spray on some varnish and pull a partial vac on it and it will soak in and possibly cover the leak site.

Crank the voltage up until it arcs.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Doesn't even have to be that fancy. I hi-potted a one like this,

breaking down at ~10kV. Sparked over the middle divider. ("Like" as in, it was actually a Panasonic version, now obsolete; this one looks very similar.)

Should have much lower leakage than the kind you've shown, so you aren't forced into a resonant type supply.

I wonder at what point the bobbin breaks down (through the divider, or to the core and back).

Can always cascade things for more voltage standoff...

Have you looked at planar (PCB based) transformers? Should be good for a lot of voltage (you can get the board basically as thick as you like), and low leakage. May be on the expensive side; definitely slow to prototype.

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Design Website:

Could have been surface moisture or just organics (finger oils, lunch, etc). You can also throw that into a oven and bake at ~85C to burn off organics.

Cheers

The Sylgard 184 is transparent, so you don't see the 2mm of extra potting around the entire transformer. After pouring, it was placed in a vacuum to set, and then baked overnight. Dow Corning Sylgard 184 is a flexible silicone elastomer, meant for HV insulation.

--
 Thanks, 
    - Win

10kV sounds good, I'm thinking of 8kV tests for a few days.

--
 Thanks, 
    - Win

Then I'd have to throw it away. I finished the SHV cable at the end of work Friday, and took home a 5kV supply. Tuesday I'll push to higher voltages, thinking of stopping at 10kV.

I need to rebuild the SHV cable, forgot to flair the shield away from the RG-58/U center dielectric, before soldering a ground wire.

--
 Thanks, 
    - Win

Are you familiar with HV putty?

Winfield Hill wrote in news:qnvakb01i21 @drn.newsguy.com:

We made ours in a way we called a "stab" termination.

The entire center, insulator and conductor is out about 6 full inches on our 50kV RG-8 version.

On RG-58, I would have a stab of about 2 or 3 inches.

We used a PVC pipe and filled the thing with dielectric paste.

The 'stab' gets a big solder ball placed onto the node protrusion (which should only be slightly less than the ball height). Down at the bottom of the pipe, we epoxied the HV out of the supply inside it. That makes the distance between those two nodes a couple inches at least.

It is like one of those unmentionable appendages with the sheath pulled back. :-)

Connector GND Coax ____________|-------| Stab | |------------@ ____________| |------------@ |-------| ^place solder ball here (in center of stab)

Box Wall Entry | | | PVC Pipe ____________|-------||__________________ | || XX ____________| ||________________XX

snipped-for-privacy@decadence.org wrote in news:qnvcpr$19am$ snipped-for-privacy@gioia.aioe.org:

It did not display right on my client.

I was thinking that you could make a HV transformer by

Winding a ferrite toroid with a few turns of well-insulated wire, maybe HV wire on secondary only

or

Separating the halves of a pot core with a sheet of really good insulator.

You could also make a capacitive or inductive coupler as a PC board. Maybe use unshielded surface-mount drum core inductors on opposite sides.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

They do make HV BNC connector.

But you also need the HV 58 cable, most are only good to 5kv.

Cheers

No, I wasn't, thanks!!

--
 Thanks, 
    - Win

That's interesting, assuming I understand it. The issue I was talking about is creating a high electric field on an abrupt edge of the shield braid, which can cause the dielectric to fail. Flairing the braid, and filling the space with insulator (heat-shrink layers, etc.), before doing any operations on it, avoids creating high E-fields.

--
 Thanks, 
    - Win

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

It is 715 V/mil dielectric strength.

Double strength transformer wire has more. Some makers go up to quad strength.

But I found the ultimate HV wire...

WOW! 500 degrees! 25kV!

board.

Winfield Hill wrote in news: snipped-for-privacy@drn.newsguy.com:

Ahhh... We also placed HV shrink on things. You are averting corona failure or a plain and simple insulation failure breach. I see.

We hired an engineer from a competitor HV company once, and this guy was really smart. He did a demo for us once while we were developing a 50kV supply. We test in fluorinert bath and that is like 460V/mil dielectric strength. We have an HV probe that can read that voltage, so he immerses the output node in a corner of the bath away from the circuit board, and then an HV probe tip can be put on it. OK so far. Then, he takes about 4 feet of an old broomstick handle, and runs an HV insulated ground wire up it to the tip where he places a nail with the tip rounded off. Standing back from the bath, as he approaches it with the tip, the entire bath started shaking and as he entered the bath, the HV arc'd right through the fluorinert to the nail. I know about gradient and such, so i knew it would do it, I just didn't think it would jump that far through the fluid, and we had never seen the dancing bath thing before either. That was cool. THEN he changed the round tip nal for a nail with a very sharp tip. Talk about far out! It started dancing from several feet away, and a LOT as he approached the fluorinert pan, and WOW, it jumped up through the fluorinert and air to the nail before he even got to it. Thankfully, each of these demonstrations were also a test of the supplies shut down features, so it was OK. Only a single SNAP arc each time. But that high gradient, sharp nail show was COOL!

snipped-for-privacy@decadence.org wrote in news:qnvgc3$1pb6$1 @gioia.aioe.org:

Not good. That gap has efficiency costs.

Two drums, same side, Nomex paper separator. Still need a good core in there.

Which one was that, I couldn't find it.

--
 Thanks, 
    - Win

Winfield Hill wrote in news:qnvi6b01uq5 @drn.newsguy.com: