Rolled cap plate connections

From:  Chip Atkinson [SMTP:chip-at-bolix-dot-com]
Sent:  Thursday, June 25, 1998 5:02 PM
To:  tesla-at-pupman-dot-com
Subject:  Re: Rolled cap plate connections 


I've done some construction experiments along just these lines.  I 
haven't yet had the time to fill them and test them out, but I'm sure 
they'll work :-)

There are two issues that must be dealt with.  The first is what you 
mentioned, a smooth way to fasten the leads to the plates, and the second 
is that you need to have these leads aligned such that all the leads from 
one plate are on one side (not end) of the roll, and the leads from the 
other plate are on the opposite side of the roll.  

I chose to not have the leads from each plate go to opposite ends because 
of problems sealing the lead while allowing access for construction, and 
also because vertical cylinders are easier to store and use than horizontal.

Here is how I solved both problems.  I suppose if I had a spotwelder, it 
would have helped.  Soldering aluminum that thin is too dificult for me, 
so I used copper tape.  I don't have any mechanical attachment such as 
welding or screws, so there is some risk, but my hope is that by rolling 
the roll very tightly the pressure and tape will combine to make the 
connection reliable.  For the leads, I used strips of the same flashing 
that makes up the plates.  One thing that you should do is anneal the 
flashing strips to make them easier to bend, which you will have to do as 
you assemble and connect them to the output terminal.

The next problem is alignment.  It is necessary to align the strips so 
that they don't cross and thus present a high risk of shorting.  To do 
this, assemble the sandwich of plates and plastic without the leads 
attached.  Put the inside end of the sandwich into a piece of heavy PVC 
pipe with a slot cut in it at an angle to reduce the bending of the 
sandwich end.  The PVC pipe is about 1.5" od, and about 20" long or so 
(long enough to provide good handles to grab onto on each end).  Attach 
hose clamps or some other method of clamping the end inside the PVC pipe 
and turn the pipe so that the cap is wound up.  (I used a lathe because I 
have one)  Once the bundle is tight, bind it up in a temporary manner so 
that you don't have to keep holding the thing.  Now use a magic marker to 
marke a radial band where the leads go.  Use another colored marker to 
mark another radial band opposite the first.  For the first few times, 
unwind the bundle almost entirely and rewind it.  Double check that the 
colored bands line up.  Once you can wind the bundle up and align the 
bands, you are ready to attach the leads.  The leads are attached in 
the position on the flashing that is shown by the band on the insulation.
Stop winding when the leads line up.

I will have pictures of this process on my web site soon.

 -- Only 47 days 'till the PGI Fireworks Convention --
        !!!! 6 full days of fireworks overload !!!!

On Wed, 24 Jun 1998, Tesla List wrote:

> ----------
> From:  Gary Lau  24-Jun-1998 0803 [SMTP:lau-at-hdecad.ENET.dec-dot-com]
> Sent:  Wednesday, June 24, 1998 7:44 AM
> To:  tesla-at-pupman-dot-com
> Subject:  Rolled cap plate connections
> >> is it bad to just make the connections, one on the end of one plate and one
> >> the oppostie end of the other plate and roll it once?
>  From:  Malcolm Watts [SMTP:MALCOLM-at-directorate.wnp.ac.nz]
> >The best caps have foils extending out each side and simultaneously
> >connected. Otherwise you have a rather large distributed series
> >inductance and resistance path the current has to traverse.
> Looks like a tradeoff between the simplicity of a single connection to
> each plate and high ESL vs the complexity of an (unrevealed) method of
> connecting all layers, approximating the low ESL's of plate and
> commercial caps.  How is this accomplished?
> As a compromise between these two extremes, how about fastening some
> number of tap points, 3-4,  to each plate, with taps from one plate going
> upwards, and downwards for the other, and tying them all together at the
> top and bottom.  Obviously we don't want screw heads and nuts rolled up
> in our nice smooth roll to fasten the tap strips.  Is it possible to
> spot-weld strips of thin flashing (or thick foil) to each plate?  This
> is the issue that will make or break this idea, I don't know much about
> welding.  If it is possible to spot-weld thin Al-Al or Al-Cu, it should
> significantly reduce the inductance and make for a neat roll.  Unknown is
> exactly how much difference, performance-wise, the multiple taps would make.
> Gary Lau
> Waltham, MA USA