SISG Quench Circuit

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Vardan <vardan01@xxxxxxxxxxxxxxxxxxxxxxx>

Hi All,

There has been work going on behind the scenes on an SISG circuit 
that can do quenching.  All "pie-in-the sky" and nothing is "actually 
tested".  The circuit of the "moment" looks like this:

http://drsstc.com/~sisg/files/SISG-Quench-02.gif

Basically, the standard SISG circuit (same reference designators now 
too!) with the following changes:

1.  R4 is just 2.2k now.  This is sort of a "last resort" resistor 
that will shut the SISG down incase something does not go 
right.  Normally, the quench circuit will shut it down far 
earlier.  Without it, the drive voltage could be at say 7 volts 
"forever"!  Like when it is off and on the shelf...

2.  R5 is variable now from the SISG turn on timing thing...  This 
might turn into a fixed resistor if a perfect fixed value can be 
found.  Prolly could be 50 ohms...

3.  R6 and C2  form a voltage ramp that slowly charges C2.  At 12V 
plus the 0.55 volt SCR threshold, The SCR will fire and force the 
gate circuit drive voltage down to ground, turning everything 
off.  D2 was added to give a "nice" mid range voltage trigger 
level.  The quench time is 7.12e-9 x R6.

4.  R7 provides the "override" function.  If the "mess screws up" and 
there again is more than 900V across the SISG, The SIDACS will 
re-fire and the IGBT gate will turn "right back on" even if the SCR 
is firing.  "Nothing can go wrong" :o))  Useful in an "emergency" 
:D  But in this state, the quench will still react "fast" once the 
current/voltage drop, so the quench should "still work".  Sort of 
cool!!  C2 is a 1% timing cap that DK sells cheap ;-))  But there is 
no guarantee that multi sections will decide to shut down at 
"exactly" the same time.

5.  It is a two terminal add-on thing, for testing at least.  It is a 
simple added mod.  If it sucks, it can be easily removed too :o))

6.  There is a tiny danger that if you set R6 to zero ohms that you 
will blow the SCR.  But the SCR is a dumb 40 cent part and not worth 
adding another resistor to protect it.  If you set R6 to zero ohms, 
"you is bad" ;-))

Quenching will probably not make a great spark length difference (but 
maybe a little) it might cut the IGBT heating in half (not a big deal 
though).  But It seems a "popular" thing that seems fairly easy to 
add.  I think this circuit will do it without breaking anything.

If it is "cool", it would be simple to add to my new SISG timing 
program too. :o)

Cheers,

Terry