Was:SS MOT's Now: IGBT / FET differences

["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>

> You can get IGBTs for free as samples from Fairchild-dot-com
> I dont know how well they will work, but they go up to 1500volts and are
> labled as high power SMPS and induction heating use.I ordered a
> few.SGL40N150DTU
> SGL40N150TU, you get 10 at a time.Should be good for experimenting?Would
> these work like a MOSFET?Like, can I replace MOSFETs with IGBTs in a
> SSTC?What kinda results can i expect?
>
> Thanks,
> Matt G

Matt
This can get off topic of TC fast so I will make some general statements
applicable
to both devices:
1.  FET's are usually superior at switching frequencies above ~200Khz,
power limit

     is usually <5kW unless multiple modules are employed in parallel or
multiple
     switch converter subassemblies are likewise use in parallel.  RdsON rises
radically
     at high voltages, so at high frequencies, conduction losses predominate.
FET's are
     not available for voltages above about 1kV, and at these levels,
currents are

     restricted (i.e. lower power output).  Sweet spot for Vds, Rds On
processed
     power is in the 500 - 600 VDC link operation range.  Realize also that if
device
     is 600V rated, reliability suggests derating device V and I to 40% of
manufacturers
     "absolute maximum" ratings.  600V devices would be suitable to operate
off of

     US / European 240VAC.
2.  IGBT's are superior power switches from DC to ~100Khz, power levels of
     MEGAWATTS are routinely being controlled daily.  As an example, EUPUC
     devices are available 3.3kV at 800A, and are used on 1.5kV DC links in 3ph
     AC Variable frequency drive railroad traction service (GE locomotives use
     GE inverters, GM {Electromotive] use Siemens technology).  The NWS
     Doppler Radar uses a 160kV IGBT pulse power modulators, I saw one
     (4-5kV) oil immersed IGBT switch out of this system at the IEEE
Applied Power

     Electronics Conference (APEC)...     in 1993!!!  Recently 6.9kV devices
     are now available for use in railroad, ship traction service; I don't know
cost
     of devices, but be assured you WON'T find them at Radio Shack...   :^C
3.  Can't use solid state for a solid state spark gap eh?  Check under
Google the
     following thread:  "pulse power modulator  IGBT".  You will be greated
with
upwards
     of 1000 hits.  One vendor is in fact running a series pulse width
modulator
using
     80 (!!!!) IGBT's in series to build a 160kV PWM / Pulse power modulator
     switch handling megawatts,  series with an additional 80 unit module
for hard

     pulse modulating the load (Klystron power tube for linear accelerator
service).
4.  IGBT's have conduction loss advantages at lower frequencies at high power
     (<= 25Khz) but you have to deal with tailing turn off currents,
however this
     phenomenon is being improved with each succeeding generation of device.
     This is on reason why IGBT's are not suitable and will burn up if you
try to
force them
     to run >200Khz in hard switch service, due to higher switching losses.
5.  Between 100Khz and 200Khz is a "gray area" where either technology could
     be used, but advantages become application, circuit topology, overload
tolerance,
     and serviceability (i.e. reliability) driven.
6.  Generally drive circuits between IGBT's and FET's are similar... BUT
there are

     issues to consider.  IGBT's have such high power gain that if a slow down
resistor
     is not used in the gate lead, many failures are assured.  Check Marco
Denicolai's
     work on 50Khz, 5kW IGBT SMDC power supply for THOR.  My experience
     (and humble opinion) suggests that IGBT gate drive circuitry is,
regardless
of what
     vendors tell you, _significantly_  more complicated then FET's,
especially if
overload
     short circuit tolerance/protection is being provided.  I have applied
FET's
for nearly
     20 years now with few failures, but my applications have been pretty
vanilla
(power
     switching  at Fo <= 200Hz, <24VDC, up to 300A peak).

     But generally IGBT's hold a significant power advantage of FET's,
especially
with
     lower frequency, high voltage applications.

7.  With either IGBT or FET's; at high power and high frequencies, the ugly
expression of
     VL = L * di/dt rises to bite you.  What this means is if you try to switch
high
     currents (say 500A for a hypothetical SSSG) in microseconds, means your
     circuit designs can't be just "wired together", the wire will present an
inductance
     of about 200nH / foot (if my memory is right).  You will get 100-200V
overshoots
     and wander why you're not happy (no sparks ...  :^C) in very short
order, due
to
     exceeding dV/dt and dI/dt limits on devices.  Stripline techniques, low
inductance
     busbars/ PCB's, surface mount snubbers and drive circuitry to try to limit
signal
     delays become critical in high power / high voltage circuitry as mentioned
above
     and in SSTC's.  The only thing helping you in SSTC is if you are in
tune, the
SS
     devices are effectively switching at near zero voltage, zero current
transitions,
     which help drammatically limit losses (heat).  But how tuned are you
when you

     add a load to the resonator (like streamers, breakout versus non
breakout, or
you
     walk near an operating SSTC resonator)?

In short, good question but no simple answers!  For additional information,
suggest

International Rectifier:    www.irf-dot-com
IXYS                      :    www.ixys-dot-com
Fairchild                    :   www.fairchildsemi-dot-com
On Semiconductor    :   www.onsemi-dot-com

or type in FET or IGBT into www.goggle-dot-com; over 70K worth of hits await you!

Regards
Dave Sharpe, TCBOR
Chestefield, VA. USA