Re: 48KWATT DRSSTC Warmup

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Steve Ward <steve.ward@xxxxxxxxx>

Mark,

Thank you for bringing that idea up!

I think the .68uF 1600V caps (942C16P68K) look good.  9 in series
gives us a 14.4kV rating with .07555uF per string.  With 33 strings we
would achieve 2.493uF at 14.4kV.  The Irms is 528A and the Ipk is
25773A.  Ipk will never be a concern here, but Irms might become an
issue.  The cost (according to Richardson) would be $4.36*9*33=$1294.

Now, looking at the .15uF 2kv caps...  We would use strings of 7, and
require 117 parallel strings.  Irms is an astounding 1579.5A!!!  Ipk
would be 50.5kA.  But, the cost would be $2.58*117*7=$2113.  Quite a
bit more money, but we would have 3X the Irms rating.

I found that the Irms on my coil producing only 6 foot sparks was
about 70A.  I will have to put together a simulation of the 48kVA coil
to get a prediction of how great the Irms will be,  but i will make a
rough assumption right now.  Say the coil runs with 500uS bursts every
8.33mS.  Thats 6% duty cycle.  Say the peak current is MAX 3500A, that
makes the Irms for the actual burst time to be about 2500A.  So then
the RMS for the whole time might be something like sqrt((2500^2)*.06)=
612Arms.  And that is probably a "worst case" number since the Ipk
will not be 3500A the entire time, and we may not even need 500uS
bursts.  Also, there is reason to believe that larger coils can
operate at even lower break rates and still maintain great spark
lengths, so if we reduced to say 100bps or even 60bps we could expect
perhaps (by the same calculation) 560Arms or 433Arms respectively.
This suggests that the .68uF caps might *just*get us by, and that the
.15uF caps would give a way over-built bank.  Since Chris wants the
absolute highest reliability, and he is forking out the cash, id be
tempted to say that the .15uF caps would be the better choice.

Maybe there is some middle ground to be found with the other 1600V
variety...  but i certainly like the idea of a *very* overbuilt MMC.

Steve

On 1/31/06, Tesla list <tesla@xxxxxxxxxx> wrote:
> Original poster: "Mark Broker" <mbroker@xxxxxxxxxxxxxxxx>
>
> Steve, Chris, et al:
>
> With an MMC this size, I'd think that the 2kV, .15uF caps aren't the
> optimal choice.  Consider using one of the higher capacitance, lower
> voltage caps from CDE.  For instance the .68uF, 1600VDC cap has a
> much larger C*V/$ ratio, and will probably still have an acceptable
> Irms rating (assuming 6x20 array, 320Arms).
>
> As an aside, I note that CDE now specifies 100kHz and 70C for the RMS
> current rating on their spec sheet.  :)
> http://cde.com/catalogs/942C.pdf
>
> Best regards,
>
> Mark Broker
>
>
>
> On Mon, 30 Jan 2006 11:31:28 -0600, Tesla list <tesla@xxxxxxxxxx> wrote:
>
> >Original poster: Steve Ward <steve.ward@xxxxxxxxx>
> >
> >I told Christopher this on the phone as well, but would like to convey
> >my thoughts on the list.  I think that an MMC is probably the best way
> >to go for such a cap.  It seems that MMCs offer very high RMS and peak
> >current ratings for the type of cap arrangements that would be used in
> >DRSSTCs (usually 6-10kV rating, but several times larger capacitance
> >than a similarly sized SGTC).  In fact, for this 48kVA coil i think
> >Jimmy Hynes and myself were thinking something around 2-3uF (thats not
> >a typo).  My largest DRSSTC uses a .5625uF cap for 12' sparks, so id
> >imagine to double that spark length, something around 4X the
> >capacitance would be required.  And if we are going for say 30' max,
> >we might need a bit more.  But really, i think the usable range is
> >very wide and not all that critical aside from affecting other design
> >considerations (but basically, pick a value and work the rest out from
> >that).  You could imagine that an MMC built for a few uF will need
> >many many parallel strings, so the peak and RMS ratings will be *very*
> >high.  Expecting peak currents of only 3-4kA, but the RMS will be very
> >high (havent actually calculated it yet, but in the hundreds of amps
> >easily).  MMC is the way to go!
> >
> >Steve
> >
> >
>
>
>
>
>