[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Taming the SISG



Original poster: Finn Hammer <f-h@xxxx>

All

The SISG circuit caught my eyes immediately, and I liked it right off.
The simplicity of the combined trigger/crowbar action appealed to me.

One thing put me off, though: I was unable to understand how the voltage doubler worked (later, when Stephen C. had the nerve to admit he didn`t either, Terry broke down and admitted he didn`t understand it himself).

Then of course there was the problem that it worked like a static gap, which is not good if the intent is to use DC resonant charging. I live in Europe, where 3-phase is the norm, so fairly smooth DC is as plentifull as 6 pulse rectifiers.

Therefore, some sort of trigger circuitry had to be applied, and since I appeared to be the one that wanted it the most, I became the first to pick up the glove and try to design it.

What I came up with was this:

http://home5.inet.tele.dk/f-hammer/sttriggedsisg.jpeg

This was heavily inspired by that 40somethingKV SCR stack that I`m sure you all have seen Alltronichs try to sell for years over at EBay.
http://search.ebay.com/scr-stack_W0QQampsspagenameZhQ3ahQ3aadvsearchQ3aUSQQcatrefZC5QQfbdZ1QQfclZ3QQflocZ1QQfromZR6QQfrppZ50QQfsooZ1QQfsopZ1QQfssZ0QQftrtZ1QQftrvZ1QQnojsprZyQQpfidZ0QQsaaffZafdefaultQQsacatZQ2d1QQsacqyopZgeQQsacurZ0QQsadisZ200QQsargnZQ2d1QQsaslcZ0QQsaslopZ1QQsofocusZbs
Of course I was worried that a SCR would not be able to perform the task, but since the SIDAC´s shouldn`t either (too high dI/dT) I gave it a go.

The main reason for trying the gate transformer/SCR solution, however, was this:

Any drive circuitry would have to pass it`s current down the gate drain resistor, and I didn`t want to raise the gate voltage that much.

So, transformer drive it was.

I had selected the IR 25TTS12 1200V/16A SCR for the job, and noted that it could be triggered by 46mA @ 2Volts on the gate, so I figured that if a gate transformer cold be made to deliver 2 volts into a 23ohm rsistor, I`d be on the right track.

I bought a couple of cores, wound as many turns on them as they would take in one layer, and terminated the winding in 23ohms. Then I passed a wire trough them, and started experimentng with discharging capacitors into the wire loop.

I soon found out, that I also needed some zener diodes to keep the voltage low enough, and that the capacitance and voltage across it could be much smaller than I had imagined. I ended up with 22µF charged to 16 volts. This is then shorted into the trigger wire by a IR460 fet for 10µs. A more appropriate solution would be to build a pulse forming network, maby I`l do that in the next iteration.

I found out that the zeners were not needed when the SCR was in connected to the transformer. by then it was impossible to whack the gate higher than 1 volt.

I already had a fiber optic transmitter from a DRSSTC project, so the receiver on the trigger board was used to pull the input pin on a 555 low trough a RC pulser, to create a 10µS pulse, which in turn drives a ucc37722 driver chip, which switches the fet.

This all worked all right in a 2 stage test board, so I built a 10 stage board and applied some power.

Since it still functions, I guess it is safe to declare the design sound and fit for replication, which I hereby invite you all to do.

there are pictures of the SISG board here:
http://home5.inet.tele.dk/f-hammer/TOTSISG.jpg

and a bigger one of the SISG itself here:
http://home5.inet.tele.dk/f-hammer/10KVSISG.jpg

A video is here:
http://drsstc.com/~piranha/PIRANHA/PIRANHA-2/FINN-SISG-MOV00628.MPG
or here:
http://www.sharebigfile.com/file/19656/MOV00628.mpg.html

Cheers, Finn Hammer