Re: PlasmaSonic II Driver Circuit Problems

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

Original poster: humanb-at-chaoticuniverse-dot-com 

Dan, I too have had some warm driver chips, noticed
during testing. Once I installed the FETs the driver
chips would get hot with no air cooling, but once the
muffin fan was installed they run cool to slightly
warm. I used 22 gauge hookup wire for the gate
transformers. I am just about ready to populate my
second PlasmaSonic, so do recommend a smaller gauge
wire? What have you used? I get pretty close to 12V out
to the gates. When run with audio modulation, with air
cooling, everything runs stone cold, with the driver
chips sometimes only getting slightly warm. Overall,
everything has work just fine with the exception of my
initial mishap (the blowing of the rail section and the
two FETs). Many hours of decent audio.

Regards,

David Trimmell

On Tue, 13 Jan 2004 22:03:44 -0700, "Tesla list" wrote:

 >
 > Original poster: dhmccauley-at-spacecatlighting-dot-com
 >
 > Chris,
 >
 > The gate driver chip heating problem has popped up for
 > a few individuals
 > building the PlasmaSonic II board.  I myself have
built
 > a large number of
 > complete PlasmaSonic systems
 > so far using identical parts on the parts list and
have
 > not had the heating
 > problems.  Although, my driver chips have gotten
 > slightly warm on a few on
 > my PlasmaSonic systems,
 > none have gotten hot to the point where one should be
 > concerned.  I do have
 > a few thoughts on this . . .
 >
 > 1.  You should be using at a minimum a normal sized
 > muffin fan mounted above
 > your heatsinks and blowing downward over the
heatsinks.
 >  If positioned
 > correctly, the airflow will also
 > cool the driver chips.  You may find that this maybe
 > all that is required to
 > cool your driver chips.  Try this first.  If after air
 > cooling is provided
 > the chips feel only slightly warm during
 > operation, than you likely don't have a problem.
 > Remember, that depending
 > on the driver frequency and the type FET (especially
 > the larger SOT-227
 > package FETs which have huge
 > gate charge), it will require a lot of power to charge
 > and discharge the
 > associated capacitances seen at the gate terminal of
 > the FETs.
 >
 > 2.  If after air cooling, the FETs *still* feel
smoking
 > hot, then there is a
 > problem somewhere.  This is usually a sign that there
 > is uneven drive on the
 > gate transformers and a bias has developed,
 > or something else is not quite right.
 >
 > 3.  What mode are you running in?  Remember,
 > single-ended and push-pull
 > modes must be configured entirely different.  For
audio
 > modulation, you
 > should be running in push-pull mode.
 >
 > 4.  +/-8V seems too low.  You should be getting at
 > least a 10V at the gate.
 > If you running 12V into your gate transformer, you
 > should ideally get 12V
 > output from the gate transformer
 > provided the turns ratio is 1:1:1.  You are losing 4V
 > which doesn't seem
 > right.  Also, your risetime of 1us is much too slow.
 >
 > 5.  22 AWG is probably too thick of wire.  What might
 > be happening is that
 > due to the insulation thickness of this heavier wire,
 > you may not be getting
 > the best coupling between windings.
 > Because of this, the leakage inductance is much higher
 > than it should be and
 > ultimately will cause the slow risetimes, and other
 > problems you are having.
 > I would try rewinding using either 26AWG or 28AWG
wire.
 >
 > 6.  Also, try varying the number of turns on your
cores
 > and see if your
 > results improve or not.
 >
 > Dan
 >
 >
 >
 >  > Hi All,
 >  >
 >  > I've built and tested a PlasmaSonic II SSTC
recently,
 >  > but it has been plagued by a couple of problems.
 >  >
 >  > 1) The gate driver chips have been getting hot.
I've
 >  > fried one set of driver chips already (4 chips) due
 > to
 >  > excessive heating, but this may be due to the next
 >  > problem.
 >  >
 >  > 2) The voltage at all the FET gates goes only to
+/-
 >  > 8V, and the risetime on the gates is 1us. I don't
 > know
 >  > what's causing this, but I am sure of the gate
 >  > transformer construction and phasing. The GDTs are
 >  > wound on the specified cores with 15 turns of
 > trifilar
 >  > 22AWG magnet wire. The voltage at the GDT inputs
is a
 >  > 0-12V square wave with approximately the same
 >  > risetime. I'm quite sure I used the correct parts
but
 >  > I may have to check again.
 >  >
 >  > Thanks in advance,
 >  >
 >  > Chris Lu
 >  >
 >  >
 >  >
 >  >
 >  >