Polarity Question.

[KJNW]

I have attached a couple of pictures of a capacitor. One covers the specifications. 500 NFD @ 100 KV. My question is, would this particular capacitor be sensitive to polarity as we are rather a "negative" bunch here. I can find nothing on the Maxwell Web Site, or anywhere else for that matter.

Carl G.

[Richard Hester]

Can't imagine why - that's your standard oil filled cap, though with special construction for low self-inductance. Having said that, it looks as thought one of the terminals might be connected to the case, so it would be hazardous to float the cap.

[Doug Coulter]

I've got some similar caps, but bigger. No, they are not polarized, and yes, one side is connected to the case. A bit of fishing around on the Maxwell data site (don't have a link handy, sorry) shows that two things can kill these - higher than rated peak currents, and hard ringing at high currents (due to making a tuned circuit with the inevitable inductance of your connections). Each cap will have a rating for life vs peak current during discharge. For example, my maxwells, 120 uf at 10kv, are rated at 100 kiloamps peak, yet have internal series R that would allow much higher peak currents, so you have to be careful how you put them into use.

At super high peak currents, there is significant danger of the wiring exploding from the magnetic fields that produces, and I found out the hard way in college - 8 pieces of RG8 in parallel on a 120u 15kv cap simply exploded, and it took hours to pick all the little pieces of braid out of my backside. Had I been facing the setup at the time, I'd be blind now. You can see more recent demos of the effect with the modern day coin crushing setups that use this type of capacitor - 1/4" pipe used as the coil simply turns to plasma on each shot.

I've got extra of 120uf at 10kv if you want to trade...mine are more suitable for things like "focus fusion" or Z pinch, due to the lower volts but higher currents allowed.

[Richard Hull]

The case connection on this type capacitor is always at the ground potential be it + or -! Floating these is a nasty business, at best.

Richard Hull

[Rich Feldman]

Carl, here's more about what Doug just said.

We know your setup is enclosed to keep fingers away from high voltage,
and to protect people from x-rays (and maybe from vacuum implosion fragments).
But with that capacitor comes a new explosion hazard.

Suppose you are not -planning- to use the capacitor for pulse discharges at more than, say, a few kA.

You still need to be extra careful with HV insulation, air gaps, and sharp edges everywhere in the path between capacitor and current-limiting element. An unexpected discharge through air could create a God-awful bang -- 2.5 kJ is similar to the energy of a proper M-80 firecracker.
Even if it leaves the capacitor undamaged, and even if there are no high-velocity fragments of metal wire.

[edit] took a while for me to dredge up this account of an amateur's accident.
http://www.brianb.org/images/Backyard%2 ... cident.htm
-Rich

[David Housley]

I believe Carl's main intent is to use this as a high voltage filter cap to reduce ripple (yes, I know the various arguments pro and con regarding ripple ). I know this type of pulse cap is often used at negative potentials, case grounded, and we don't expect high discharge currents in a filter cap configuration. To the best of my knowledge, no very high current discharge experiments are planned in the near future. I am going to reccomend an insulated blast sheild for the cap, just to be on the safe side. I believe 3/4" thick HDPE secured with 1/4-20 nylon hardware should do the trick, but welcome any input from those with specific experience. Thanks for the input, everyone!

[David D Speck MD]

Carl,

The nameplate looks to me like it's 5 uF, not 500 nF. Either way, you will need to supply some sort of discharge mechanism to bleed off the considerable stored energy after you are done with your experiment.

I would suggest a normally closed vacuum relay that connects a string of appropriately sized (i. e. Big) power resistors across the cap that will dump the charge in a few seconds when the driving supply is turned off. I'd pick a value so that five RC time constants work out to 10 - 15 seconds or less.

Dave

[Rich Feldman]

> The nameplate looks to me like it's 5 uF, not 500 nF.
Don't you see a typed decimal point to the left of the nominal and measured capacitance values in uF, and the ESL value in uH ? 220 uH would be ridiculous. Wish they had included a zero to the left of DP.

I think the case is not big enough for a 25000 joule capacitor of this type.
That's using the nameplate and the fluted insulator as approximate references to determine the scale of the pictures.

p.s. David, thank you for the hint about how to remove plug from McLeod gauge.

[David Housley]

We are using a Ross engineering high voltage relay under dialax HV transformer oil, with an appropriately sized dump resistor, also under the oil, such that when the high voltage is shut off, it will automatically dump the charge on this cap and those in the C-W. This has been a safety system in the works for awhile----things just come in stages. THanks for the input!