Re: SSTC - Q's

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

Original poster: biomed-at-miseri.winnipeg.mb.ca 


 >Date: Fri, 23 Jan 2004 08:10:23 -0700
 >From: "Tesla list" <tesla-at-pupman-dot-com>
 >To: tesla-at-pupman-dot-com
 >Subject: Re: SSTC - Q's
 >Message-Id: <6.0.0.22.2.20040123081020.02593c58-at-twfpowerelectronics-dot-com>
 >Content-Type: text/plain; charset="us-ascii"; format=flowed
 >MIME-Version: 1.0

 >Original poster: herwig.roscher-at-gmx.de



 >Shaun,

 > > The MOSFET's that are used on the plasmasonic have a "fast body diode"
 >- The IXFH13N50 I am using have a "fast intrinsic rectifier" also.
 >
 > > the internal diode takes a long time to stop conducting when the
voltage on
 > > the gate goes low.
 >- In my case the t rr of the diode is 300 ns, while td (off) of the
 >FET is 100 ns. Therefore I have disabled the body diode by connecting
 >a large Schottky diode in series with the drain and replaced it with
 >a very fast MUR 1560.
 >
 >Some figures:
 >
 >Taking the output signal of the TL 494 as a reference, the turn-on of
 >the FET is delayed by 30 ns and the turn-off is delayed by 200 ns. I
 >will run the system at about 370 kHz. Do these figures sound solid?
 >

You'd better ask the maker of these Boards, Dan McCauley about this. The
Trr of the IXYS mosfet, speced out for the plasmasonic board is 200 nS and
that doesn't present a problem with turn off time,(no short across the
power supply).  The MUR diode across the mosfet conducts reverse current
faster than the intrinsic rectifier could.  With some other SSTC coilers
the shottkey diode they were using shorted out, probably due to too high a
current through them, so the coils driver boards worked fine without them.

 > > What happens is the other MOSFET turns on while this body diode is
still
 > > conducting and then you have a short.
 >- Unfortunately I did not realize the large difference between turn-
 >on and turn-off time (it is the first time I am using power MOSFETs).
 >
 > > The turn ons are delayed abit the the resistor and parallel shottky
 > > combination.
 >- So the goal is to make the turn-on time more or less equal to the
 >turn-off time?

Looks like that would be ideal.
 >
 >Reading that zener diodes connected to the gate directly might cause
 >oscillations, I have connected a series resistor of 15 Ohms in
 >between. This resistor should increase the turn-on time as well. In
 >order not to make the turn-off time longer, I have connected a
 >Schottky diode across.
 >
I had lots of ringing on my gates too.  I put a snubber network across each
gate transformer output.  51 ohms, 1/4 watt in series with a 1nF cap.  This
snubber worked great to cut out ringing.  I was getting so much ringing
that it would false trigger the power mosfets.  This was the original
reason I made a snubber.  These values work very well.

 >BTW: Instead of two back-to-back zener diodes I have used a
 >bidirectional suppressor diode.

As long as it prevents the gates from seeing too high a voltage,  I don't
remember what the maximum is but you can download the specs from IXYS.

 >
 > > I thought of the voltage spikes too, so I made four snubbers to go
 > > across the Source to Drain of each MOSFET.  51ohms resistor in series
 > > with 1nF cap.

This second set of snubbers was an after thought to protect the mosfets.

 >- I have tested your solution and found that the ringing caused by
 >the wire wound resistor, I used as a load, nearly disappeared! Thank
 >you!

Cool,
 >
 > > The plasmasonic MOSFETS are 10 nF capacitance each
 >- C iss of my power MOSFETs is 2800 pf typ.
 >
 >Regards,
 >
 >Herwig
 >
 >

Shaun Epp