FEA of new rotary spark gap

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "MakingLightning" <MakingLightning@xxxxxxxxxxx>

I am designing my new rotary spark gap.
I am using the 1/2" x 1.5" tungsten electrodes that a few people have
ordered form Jim Mora recently.
A fellow engineer at work who specializes in FEA worked through an analyses
of it for me because I wanted to make sure of how safe this gap would be and
exactly what the numbers calculated before hand represented, a good
double-check.

My design may be similar or give insight to others so I thought this info
might be useful for some.

Is there a site or a way of posting this info somewhere for all of us to
access?
How do I do that. I put everything in a MS PowerPoint presentation. Do you
think that is ok? It is what we use a lot to convey this type of thing for
people that can not import the data. Not too many would have this type of
software.

I had it modeled in SDRC IDEAS in 3D and it was made from sketch based
modeling (easy to tie things down and change features of the rotor easily)
The definitions of the properties were taken into the model and links will
be also be posted for you all to see.
Then the FEA was using SDRC Visualize FEA software (in 3D of course, but the
presentation is in 2d)(running on a machine using 2, xenon 3.6G processors,
6 gig of ram, screaming video, SCCI RAID large fast hard drives. We
represented the FEA model using 1/8 of the disc because they would all be
the same. It still took it 6 minutes to crunch the numbers).

My disc is 12"dia G-10. The electrodes were spaced on a 10" center leaving
3/4" of g-10 holding it in.
We crunched the numbers first and they agreed with the FEA results, so I now
have a warm fuzzy feeling that the disk will be safe by a good margin. Even
with all this, I will still shield it with Lexan. Like chaining down
something to a trailer, I have learned to use at least 2 chains, then add a
3rd one. Each one being much much stronger than needed.
A few of us analyzed things and everyone thought if the disk was mill and
drilled, and holed reamed 0.0005" undersize, that should be sufficient to
hold things in. It would take side forces to work it loose and they are
almost non-existent. Huge forces are pushing it towards the outside of the
disk, and press fitting it is very strong. We press brake rotors onto shafts
and no matter how hard you hit those brakes, with heat, with full engine
power applied, they do not come loose. That is direct twisting applied,
which we do not encounter. Yes, I know it is steel on steel and not tungsten
and g-10, but that has huge forces applied as you will see from the FEA
results. One set shows this effect and values with the wheel static. The
disk will be balanced very well. All electrodes are one piece and everything
is precision milled and drilled, and press fit.
Another set will show the results of the disk spinning at 4000 rpm, which
translates to 533 beaks per second. I know the real life situation will be
less, but I have heard reports of people using up to 480 BPS, so I wanted to
capture worse cases, slower will always be better. Smaller diameter, like
the 3/8's will be better too. If there is a common size disc, electrode
number, center, for those, let me know and I can run that scenario for us
fairly easily now, or at least for one more version.
I am not sure what the consensus of the 1/2" rotor designs are either, but I
went with mine.

We did not think the set screws were necessary and could be a bad thing.
They do not publish the specs of that material needed to model how the
layers could separate and be compromised going in on that angle. That
laminate is not designed for that kind of stress. They only publish the
Tensile strength, crosswise and lengthwise, and the flexural modulus
crosswise and lengthwise. It would take a LOT of time to figure out properly
the unpublished data. You are also putting that set screw in the place that
has the most stress applied too. Maybe not the best idea.

Well, someone let me know where I can post this info. You all can give me
your input on this to make sure I am not missing something. Maybe it will
help others.

Kevin