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

MMC (multi mini cap) (was Re: Tube coil capacitors)

Hello Steve, Dave, Terry, all MMC builders & users,

My comments below.

>Original Poster: "Steve Young" <youngs-at-konnections-dot-com>
>
>These caps are probably fine (economical) for tube coilers operating at
>much lower voltages than typical disruptive TCs.  But they don't appear
>very economical for the latter.
>
>Based on Terry's experience, let's assume the caps can actually
withstand
>twice their rated voltage.  So each cap is .0056 uF at 3200 VDC.  If we
>string 16 in series, we get .00035 uF at about 53 KV DC.  So a hundred
of
>them would give us 6 times that or .0021 uF.  For most coils we would
need
>an order of magnitude more capacitance, or about 1000 caps in
>series-parallel.  So the cost would be probably $600 per 1000, based on
the
>$69.75 per hundred price in the posts below.  That's expensive, and a
lot of soldering!
>Bottom line, probably more economical to buy a Maxwell or other
commercial pulse cap.  Comments?


Sorry Steve, but I canīt agree with you at all.

My 8" coil is presently "nursing on" MKP (polyprop) caps.
Each cap has 62nF and is rated at 1000Vdc. I started out
with a string of 22 (thinking similiar to you and not wanting
to kill them right from the start). My power source is a
7.5kV (10.6kV peak) 75mA NST. The coil ran great.


Then I started shorting out caps in the string and retuning
the coil for each new run. Do you know where I stopped
shorting out caps? At cap number 12!! So, I am presently
running PeakACV ~ rated DC cap voltage. With 60 caps
(5 strings of 12 each) I presently have a 25.5nF cap.
Total cost (for the 60 caps) was 9 bucks!!. The caps
have seen a total real "on time" of at least 30 min.
They stay ice cold and the 41" output supports my,
Terryīs, Richard Hullīs, etc, (all who have ever done
this) idea, that this WILL work. Output arcs are white
hot. I am NOT using any safety gaps or equalizing
resisitors. Just how evil can you be to a cap?


Interestingly enough, I added a small PE cap (~2nF)
and guess what? Yeah, the arc length actually went
(I DID retune the coil) down!!! This PE cap was built
without great care, etc. It was one of my test caps
(that had survived my tortures).

My final setup will be 15 MKP caps in a row (for a
higher safety margin). I am presently working on
a 100-120nF cap. Total cost for this cap will be
roughly around $150. This is all inclusive (caps,
resistors, safety gap, box, etc). I donīt think you
will find a commercial pulse cap for this price.....


The secret behind this construction is the fact, that
you are using commericial caps and doing nothing
different, than a pulse cap manufacturer does
(lots of series strings in parallel). So, you get all
the benefits of a pulse cap without the heavy
price tag.

Plus, if I compare the multi mini to the PE cap, I
donīt think the MMC is more work. Sure it IS a lot
of soldering, but on the other hand, the PE cap
involves measuring, cutting, folding, rolling, etc, etc.
In order to get your PE caps bullet proof it requires lots
of work, too (getting out all the air is a good example).


Another benefit of the MMC vs. R-PE-C is the simple
fact, that if your PE cap is rated at 15kV youīd better
not use more than 15kV (safety margin). The caps used
in the MMC are conservatively rated. If the cap has a
PRINTED rating of 1kV, you can bet your socks, that
you can feed it 1.5kV WITHOUT doing any harm. No
commercial cap manfacturer will print the MAXIMUM
voltage the cap can take on the label (one reason is
possible variations in different batches).

Of course buying a commercial pulse cap is
easier, but.... ;o)) $$$$$

A while ago, Richard Hull wrote he wouldnīt trust
the MMC for more than 2kVA input. I canīt agree
with this either, because the total load is processed
by more than one string of caps. This means the
A-load per string (depending on the number of caps
per string times the number of rows) is divided up
evenly. If you use a little thought during
construction each string will never be
overloaded (amps*ĩsecs).

My present setup runs 320A of primary current.
I am using 5 strings, so each string of 12 caps
"only" sees 64A effective current. My final setup
will be running ~1.5kA of primary current, but I
will be using 24 strings, so this amounts to
~62.5A per string of 15 (!). I canīt see any
survival problems here.


Coiler greets from germany,
Reinhard

P.S:
Terry wrote: "... (I am a failure analysis engineer)...."
So, Terry, you like killing electrical things?  :o)) !

-o-o-
   o
####