[Halld-cal] 2500 volt capacitors

[Fernando J. Barbosa]

Hi Paul,

As promised, I performed some tests on HV caps. These are also from AVX 
but smaller than yours, size 1825 and 220 pF. I have two types in reels: 
one is rated 4kV and was purchased about 2 years ago; the second is 
rated 3kV and I just purchased them.

I tested both of them with a Bertan HV supply with wires soldered to the 
metalizations of the caps, ~60% RH. I increased the HV to 3kV and had no 
sparking or leakages of any sort. At 3kV the current was zero.

Best regards,
Fernando



Paul Smith wrote:
> Hi Fernando,
>
> I've tried several of these capacitors, AVX #2225WC103KAT1A, .01 uF @ 2500 volts.  I solder wires to the metallization, then connect to a high voltage power supply.
>
> The ones off of a cut tape I got from DigiKey typically spark at ~1500 volts, then stop.  This is independent of humidity from 40% to 55% RH.  If you gradually raise the voltage, they will spark at higher voltages, but eventually will hold 3 kV without sparking.
>
> I tried one from a reel I got from another distributor; it holds 3 kV at 53% RH without any "conditioning".  If I breath on it, it will start sparking, but then stops.
>
> If you look at the metallization on the ends of these caps under a microscope, it isn't anything close to smooth.  The 2nd batch is better than the first, but still not great.
>
> I'm pretty sure that the sparks burn off the "points" of the metallization.
>
> It's hardly practical to "burn-in" each of nearly 10,000 capacitors.  They will work with just the 1B31 acrylic conformal coating (and no epoxy), but I think it needs to wick under the capacitor as well so that the capacitor is surrounded by the coating.  I've been using some coating in a syringe which I squirt along the sides.  Removing a capacitor verifies that the coating does indeed wick underneath.
>
> I don't see how we can rely on these capacitors without some sort of coating surrounding them on all 4 sides.
>
> Paul
>   
-------------- next part --------------
A non-text attachment was scrubbed...
Name: barbosa.vcf
Type: text/x-vcard
Size: 188 bytes
Desc: not available
Url : https://mailman.jlab.org/pipermail/halld-cal/attachments/20100811/aec03e43/attachment.vcf