[Prev][Next][Index][Thread]

RE: CAP DESIGN




 * Original msg to: Kukkonen-at-snakemail.hut.fi


 KU> On Sat, 31 Dec 1994, Richard Quick wrote:
 
 RQ> Yes this will work. But you are better off building a flat   
 RQ> plate capacitor construction with this thinner material. Use 
 RQ> maybe two thicknesses of plastic between sheets of AL foil.  
 RQ> Build up stacks of very large capacitance (like .15 uF) and  
 RQ> wire a bunch of the stacks in series to get around .015 at   
 RQ> 15kVAC pulse.

 KU> I just made a calculation about the cap - do these specs     
 KU> sound reasonable to you? :
 
 KU> ..using about 30m^2 of .2 mm (.0078") thick ld-PE sheet I'll 
 KU> make 4 caps that will be connected in series, each 69nF. 

I work this out at .0172 uF (microFarads) when four units are
placed in series. So far so good.

 KU> These individual caps will have 3 layers of ld-PE between    
 KU> each cap-plate (Al-foil, 35um [cm assumed? rq], 30cm*20cm)   
 KU> and using 40 plates I should get about 69 nF.. All together  
 KU> I should get 17nF at 43kVDC using the voltage-rating of      
 KU> 18kV/mm and thus about 14kVAC for Tesla-use..

Not bad, not bad at all. You have left yourself a reasonable
voltage margin to spare, your capacitance calculations are very
reasonable for a homemade unit of this construction. Properly
constructed this will take a bit of time, but should be quite
servicable once completed. Make everthing clean, clean, clean. 
Do not allow lint, dirt, defective plates or defective dielectric
into the construction.

 KU> The unit size will be about 36cm(h)*30cm*12cm minimum +      
 KU> oil-immersion tank.. 

What plans, steps, etc. were you going to take to remove air from
this capacitor? Natural bubble-out break-in by mounting the
plates vertically in the tank, letting it sit for a few days,
then pulsing the cap at lower voltage, gradully bringing it up to
full power? Pumping the tank down with a vacuum pump? Both?

 KU> Thanks for your advice in advance!

Not a problem. I just wished I had had someone to bounce ideas
off of and get suggestions from before I spent weeks of my time
designing and building failures. Besides, it has been an
excercise in analog-digital numerical conversion. One day I might
get the hang of it. |-)

Richard Quick
... If all else fails... Throw another megavolt across it!    
___ Blue Wave/QWK v2.12