Terry's DRSSTC

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Terry Fritz <teslalist@xxxxxxxxxxxxxxxxxxxxxxx>

Hi All,

Busy Christmas!!  I got up early at 4:00am to work on the DRSSTC since the 
"daytime" was taken ;-))  Santa delivered the boards late on the 24th from 
his UPS truck pulled buy eight tiny cylinders.  A day late, but I'll 
forgive ;-)

I got the Pulse Controller built early Christmas morning that sends the 
coil firing pulses via fiber-optic cable:

http://drsstc.com/Construction/PulseController/PulseController.JPG

The only change was to leave the RESET line open as the Basic Stamp needs 
after all.  I updated the PCB art to fix that and few tiny issues:

http://drsstc.com/Design/PulseContoller.gif

http://drsstc.com/Construction/PulseController/PulseController-01.gif

I cleaned up all the usual software boo boos (just a text file):

http://drsstc.com/Construction/DRSSTC-100.bsp

But now it works great!!

http://drsstc.com/Construction/PulseController/Output.gif

It smoothly goes from 40 - 300 Hz and from 10 - 500uS.  It starts up at low 
power by default and does the other functions as expected.  It seems like a 
perfect solution.  I am still waiting on the heavy aluminum box for it to 
get here.  Below 40Hz requires added software things and above gets kind of 
tricky since the controller needs time to "think".  The pulse width is 
trivial, but I thought it best not to go to "zero" since the thing would 
appear to be "off" when it certainly was not.  The controller is capable of 
all kinds of wild things, but this will get it going for now.  Since it has 
an RS-232 port, it could even be run directly from a laptop...  Fiber 
optics are wonderful for that stuff ;-))  The cable is glass, Arimide 
fiber, and a polypropylene jacket.  Nothing gets through it (well tested!! 
;-))).  There are low power "battery friendly" versions of the Basic Stamp, 
but just the plain BSp24 seems to run fine on a 9V battery for a few 
hours.  I'll probably just usually leave it on over night and kill the 
battery anyway (but it does have an LED to warn...) ;-))


I also cleaned up the main controller schematic and added goodies:

http://drsstc.com/Design/Controller.gif

I hope this is not looking too complex ;-)  But it is pretty easy to 
understand.

On the left, CT+, CTgnd, CT-, and CT+v go to the current transformer on the 
primary output cable (inside the unit):

http://drsstc.com/Design/CurrentPhaseDetector.gif

These input signals are nice square waves that tell if the output current 
is there, the timing, and in what direction.  R1 and R2 limit gate input 
current if something goes wrong.  U1a is null.  U1B and U5 provide a 3.3uS 
trigger pulse on the rising edge of a PULSE signal from the fiber optic 
pulse controller (the receiver electronics are not done yet).  So when the 
PULSE signal tells the mess to turn on, this circuit will turn on the IGBTs 
for 3.3uS to start the oscillation.  U2C could provide an early shutdown if 
something odd happens (like there is not a coil connected).

U3C just takes the PULSE, FAULT, and RUN signals (panel switch) and if all 
is good, lets the IGBTs run.  If anything is not right, the IGBTS are shut 
off.  The RUN switch power signals are provided on the board and they go to 
the main AC power switch.  I hope that switch does not explode since there 
is a heck of a lot of power on the other set of contacts!  But I got the 
"expensive" switch there ;-))

U2A and U2B diss-allow both sets if IGBTs to be on at the same time (a BAD 
thing!).  If one set is on, the other has to be off...

R10,R11 and C12,C13 can provide an RC delay if I wish to provide a "dead 
band" between one set of IGBTs turning off and the other set turning 
on.  If the resistors are just shorted then there is no null time.  But by 
adding various RCs here, I can really forcefully prevent cross conduction 
if I need too.  I thought this was sort of cool ;o))

U1C and U1D prevent a set of IGBTs shutting down until the zero 
crossing.  I thought it might not be a great idea to switch the IGBTs off 
at say a 400 amp peak if the PULSE signal stopped.  So once the half sine 
wave event is started, it will not stop until the current reaches zero 
again.  This is a bit scary if something goes wrong, but if anything bad 
happens, the Lexan will stop the flying parts ;-))  No "electronics" are 
going to stop that anyway.  I though this might help that to never 
happen...  I worry slightly here about "race conditions" and the RC keeping 
something on "too long".  But I can't "foresee" any problems.

U3a and U3B are just the master drive AND gates.  The signals out from 
there are buffered buy U4 and sent the to TLP250 isolated LED drivers.  The 
AND gate output could drive GDT stuff too if one wanted.  At that point, it 
is all pretty universal.

So it does all kinds of cool functions, has less than 40nS delay through 
the mess, runs the axillary stuff (might still need to run the fiber optic 
front panel receiver board), and might actually work ;-))  I was once told 
by and H-drive genius that once you have so many shutdowns, fault lines, 
and protection logic in the drive that it is a miracle if it ever turns on 
at all, you are almost there ;-))

I think I will make PC boards for the CT and it's parts and one for the 
input fiber optic panel receiver.  Probably just stuff both on a PCBexpress 
cheap card and cut it in half...

None of this is actually tested at all yet, but it is looking really 
cool!!  It only has 5 ICs and maybe 16 other parts.  The output 1nF cap and 
150 ohm resistors on the driver outputs should really be near the 
opto-drivers on the other board...

I really like the screw terminals (DK#  281-1435-ND).  They are sort of big 
but they have easy 0.2 inch centers and the screws are not made of soft 
lead that strips out on the first use.  I also have a lot of drivers that 
actually "fit" these screws ;-))  The 2-position ones can just be clipped 
right together as needed...  They take tiny wires up to 12 ga.  In this 
case, the ability to take wires on and off is really needed.  I do tin and 
clean the flux off the wire ends though since plain stranded wire tends to 
squirm out of such connectors.


Everything at:

www.drsstc.com

should be pretty up to date...  Some of the pages are a little "funny" 
since I got tired of the "latest" web page software...  I think I will just 
use StarOffice after all...  Feel free to copy steal and plagiarize 
anything there if it helps your coils out ;-))  I don't mean for the site 
to be all about "ME".  Eventually, it will get more generalized...

Cheers,

	Terry