Re: IGBT's, pulsed power (was Re: Waveguide TC)

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "David Sharpe by way of Terry Fritz <twftesla-at-qwest-dot-net>" <sccr4us-at-erols-dot-com>

Hi Greg,

Outstanding work!  One comment; if you are using a IGBT + copack diode
as a freewheel switch replacement for a spark gap, your di/dt I  think
would be limited by the Zo of the tank circuit, and current starting in a 
sinusoidal
waveform from 0. dV/dT stress would clearly be a much bigger problem, where
you are dropping from capacitor charged voltage to effectively 0 in
microseconds.  Turn off is not a particularly severe problem, especially if 
using a
resonant charging L, and switching off occurs when all energy has been
transferred to secondary or you're at a natural notch quench.

With the SLAC modulator however, your dealing with a square wave impulse
versus a sinusoid, so your stresses in both dI/dT and dV/dT are extreme.  I'd
rather try to tackle a SSSG replacement then what you are trying to accomplish
in a Z-transform pulse power regime.  :o)

The bond wire placement and uneven current distribution stresses have been
discussed here, AMAZING that OEM's have not taken the analytic work to
the level that you are performing.  These devices are not cheap, and these
devices are normally used for railroad and ship traction service.  IR has
started to offer improved devices that are limited by silicon not bond wires
with their Magna-Tab Program for FET's for 42V/500A automotive applications.
Unfortunately, I've heard from several sources that this technology won't be
expanded to IGBT's.  That's too bad, there is much work that could be done
is this arena of solid state pulse power, and I'm sure market opportunities
would develop as outcomes from successfully solving these problems.  At a
minimum, a more robust and reliable device under any field application should
evolve from this work.

Good luck, maybe a new device technology may evolve from this work that
may eventually find its way to industry, and eventually Tesla coils...  :o)

Best Regards
Dave Sharpe, TCBOR

Tesla list wrote:

 > Original poster: "Greg Leyh by way of Terry Fritz <twftesla-at-qwest-dot-net>" 
<lod-at-pacbell-dot-net>
 >
 > >Original poster: "David Sharpe <sccr4us-at-erols-dot-com>
 > >
 > >[snip]
 > >Stanford Linear Accelerator (SLAC) is one of the highest power
 > >pulsed linear accelerators in this area, Greg Leyh may have some additional
 > >comments concerning that system.  I'be been watching their ground breaking
 > >work with IGBT's and pulsed power klystron drivers with great 
interest... and
 > >hoping for several technical breakthroughs that can be utilized at our end
 > >of the power spectrum.
 >
 > SLAC is engaged in the machine design of many subsystems
 > for the Next Linear Collider, including the pulsed power
 > modulators that will feed the 8000 or so 75MW x-band klystrons.
 >
 > Each of the 1000 or so modulator stations needs to generate a
 > 500kV, 2000A, 3uS pulse, at 120BPS.  Originally based on hydrogen
 > thyratrons, the current modulator design now incorporates a
 > novel 'fractional turn' transformer geometry driven by an array
 > of 156 3300V, 2000A IGBT's.
 >
 > The IGBT's have been somewhat problematic, given that they are
 > being operated in a very un-spec'd area of their parameter space.
 >
 > Here's a summary of the IGBT development that's been done to date:
 > http://www-group.slac.stanford.edu/esd/default.htm
 >
 > IGBT's hold a great deal of promise for other pulsed power
 > applications, as well.  They might be the right answer for
 > medium-scale coils; it will depend largely on the requirements
 > for dI/dt and the primary voltage.  For larger coils, where dI/dt
 > is not an issue and where primary voltage and RMS current
 > capability is critical, it appears that laser-triggered silicon
 > might very well be the answer.