by Doug Coulter » Fri Jul 01, 2011 2:08 pm
If you're only close to arc over, glyptal (the real thing, accept no substitutes) is a good insulator that can be dipped or brushed on -- a couple coats will hold off an extra few kv well. Maybe 10 if you do a nice job. It also cuts down sub-arc corona losses a lot. Back in the day, we always had some around the TV repair shop for this. It can stand pretty good heat to -- it thermosets (cross links) the first time it gets hot and becomes even better. Some of the two part potting compounds are good too, but not cheap at all and hard to work with. I got some stuff like Spellman uses and it's not easy to work with, as you have to put it under vacuum after mixing to get all the bubbles out, then you can pour it and pressurize it into the nooks and crannies. Regular silicone isn't so good. The hardware store stuff gives off ammonia to cure, the "GE type 2" is just a failed formula for most uses (including the intended one -- it's either hard when bought, or never gets hard). The only stuff I've had luck with is the blue Permatex gasket maker, which is pretty expensive -- and not as good as glyptal for either insulation or vapor pressure in vacuum uses.
The low losses on the pump caps is why people mostly are well satisfied with a plain CW multiplier -- they work, and you just move on to doing the experiments. The ripple on the main stack is because during one half cycle, the charge for the coupling caps is supplied by the filter caps below it....which is why most all supplies use split phase and as square a wave as they can -- then the ripple contributions of the two sides cancel. And you only need one big driver instead (much cheaper usually).
Somewhere above 30kv -- say 50kv and up for sure, the relationship between voltage and sparking length completely breaks down for all the but biggest, smoothest balls (and no dust or insects). What happens is "streamer formation" and once that begins, arc length is more or less unlimited. A little channel of ions forms (could be started by a cosmic ray) and it's conductive -- and has a very pointy end. The pointy end of course concentrates field around it, so makes more ions (drawing enough current to keep the channel ionized too) and the process can continue almost forever. I've had a 10 foot arc here on 100kv, was all too exciting -- it went out from the supply about 5 feet, made a u turn (right near a ground too!) and then came back about 5 more feet to hit about 6" from where it started. I've seen 300 foot arcs in movies on the 'net from a mere couple hundred kv when a breaker popped and the quench didn't work.
This is one of several reasons for corona rings on things like this -- to ensure that no single drop is large enough to cause this to happen. By dividing any big drop into definite, controlled, smaller drops, it's all much more manageable.
It's also how the voltage of mere lightning can make an arc to the ground....which is a good place to look for more information on the phenomenon.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.