Re: 04 Gluing acrylic (fwd)

[Steven Roys <sroys@xxxxxxxx>]

---------- Forwarded message ----------
Date: Thu, 11 Sep 2003 20:49:01 -0300
From: Antonio Carlos M. de Queiroz <acmq@xxxxxxxxxxxxxxxx>
To: High Voltage list <hvlist@xxxxxxxxxx>
Subject: Re: 04 Gluing acrylic (fwd)

High Voltage list wrote:

> From: B2 <bensonbd@xxxxxxx>

> Acetone works for me.  Getting the bubbles out takes a lot of practice.
> Clean both surfaces scrupulously with isopropyl alcohol (reagent grade
> ethanol if oven stress relieved).  Bond strength gets better the longer the
> acetone is in contact with the acrylic.  Practice a lot on scraps before
> attempting the big pieces.  You may have to "soak" the inner discs in the
> acetone to get enough surface penetration for a good bond.  A good bond is
> present when the stack appears optically as one disk with no or very few
> bubbles.  Weights may be required to get the bubbles out.  Bonding is almost
> instantaneous.  You only get one chance to do it right!

A problem with solvents for this case is that there are metal sectors
between the plates (aluminum foil). They are just a small fraction of mm
thick, but something must fill the gaps between them. A quite messy
possibility would be to dissolve some acrylic in acetone and use the
resulting "glue" to glue the disks filling the spaces. I wonder then if
the glue will dry, and will not dry -while- I am applying it to the
large surfaces. I made a test some time ago using common paint thinner,
that also dissolves acrylic, but the results were not good. I had only
seconds to apply the glue before it started to dry. The result was
electrically unsound. I will try to find what happened now, months
after.

> Another method is to heat the assembled acrylic pieces together in an oven
> just below the melting point.  They must be lying on an absolutely flat and
> level surface.  Use weights on the top piece.  This requires great skill to
> be done successfully.

I did this with three small disks (that fit in my oven), but with some
hot glue between them. I made tests without glue, but the adherence
was weak, unless I actually melt the acrylic. There is the same problem
with voids, that sparks find very easily, and acrylic deforms when
heated (usually shrinks), causing severe deformation.
I pressed the disks between two flat glass plates with weights. This
also caused another phenomenon. A liquid, apparently just water, leaves
the acrylic when it is heated, forming pools between the plates and the
glass, that leave marks (easy to polish away).

> Another method is to hold them together with many nylon screws.  Use RTV to
> insulate the holes.

I tried to joint the disks with screws (metal screws at the edge and at
the center don't cause serious losses), but the voids between the
sectors
light up when the machine operates, indicating heavy spark tracking,
and the performance is poor.

> Another way is to use a compression hub with two halves that screw together
> and tighten the center part together.  Can be as simple as a piece of half
> threaded rod, two nuts, two lock washers, and two fender washers.

It's necessary to join the borders too, unless I deform the plates
first, so they ate slightly conical. Insufficient insulation again.

> There are other methods using lasers, microwaves, induction (dielectric)
> heating, or ultrasonic welding, but I am not familiar with them.

Thank you (and the others that suggested solvents) for the suggestions.

This is a picture of the three disks with sectors, before assembly:
http://www.coe.ufrj.br/~acmq/wehrsdisk.jpg
The assembled disk in place, fixed with screws:
http://www.coe.ufrj.br/~acmq/wehraug1.jpg
The result was poor (this machine should produce at least 25 cm sparks):
http://www.coe.ufrj.br/~acmq/wehrtest.jpg

Eventually I damaged one of the disks in the stack, and can't test a
method for sealing the stack until I make another.

I installed a single Wimshurst-style disk, and got 20 cm sparks:
http://www.coe.ufrj.br/~acmq/shslpk2.jpg (this is a 15 cm spark)
But the current (short-circuit) is at one half of what it was.

Antonio Carlos M. de Queiroz