OLTC

Here is some info on my first attempt at a solid state coil. What got me started was Seeing Jimmy's DRSSTC site (see the links section) I wanted to build one , but found Steve Connor's OLTC stuff and decided it would be better to start with an OLTC as they are far easier to implement.

The coil itself was put together using the good old 6" X 22.5" secondary that I have used in many setups for over 6 years. It resonates with the 6 X 24" (approximately) toroid at around 130Khz. This is not great for an OLTC as the whole idea is to use a very large primary capacitor charged to a relatively low voltage (off line with no step up) . A 1.8uf cap charged to 700V holds about .3 Joules of energy so to get any decent power into the coil the break rate was going to have to be very high. I did eventually get 9" sparks with about 600BPS before giving it away to build the DRSSTC.

In the picture you can see most of the components. The iron core inductor is 10mh, I bought it from a local electronics store where they were sold as speaker crossover parts. The 10mh value was choosen based on Steves results using a similar charging choke (sorry no maths here) The two IGBTs (IRG4PF50WD) are both mounted very close on the same heatsink to keep the inductance low , both IGBTs at the same temp and to try and make the inductance between each IGBT and the primary circuit as equal as possible. I did originally start with two one turn primaries , one for each IGBT, but at some point (after much though and reading of the IGBT data sheets) decided that the IGBTs would share the current pretty well regardless so I changed over to a single primary. Each IGBT has 2 400V 1500W transzorbs in series between the collector and emitter. You can get some pretty nasty spikes and after I lost an IGBT for no good reason I installed those (thanks Steve) and had no more problems. The estimated current is about 1100 amps for the first half cycle. Lastly you can see the Coax that disappears to the left of the picture. This goes back to the gate driver. The tank cap is made up of 12 .15uf 2KV Cornell Dublier caps giving 1.8uf with a very high pulse current rating.

I'll add the full schematic here when I get time. Still it is very similar to Steve's except the timing comes from a PIC rather than a 555.

The driver setup ended up being a PIC micro driving a TC4429 FET driver. The 4429 then drove a FET and a bipolar transistor set up as a totem pole. I found the 4429 to be an excellent driver but it is limited to 18volts. Those little IGBTs really needed a bit more which is where the FET and bipolar transistor came in. The final gate drive was 25 volts. I tried up to 30 but found this didn't help much in my circuit. Even though others have run their IGBT gates up to 40 volts I am a little skeptical of the reliability when doing this given that the data sheets usually specify 20 volts max. At the higher gate voltages there is also less head room left if the gates get spikes up, and there are plenty of ways that can happen :-(

This crappy picture shows the thing running and making about 6" sparks. Those little IGBTs do seem to get hot quite fast so I was never game to run it for more than about a minute. If  the IGBTs are being punished with about 2.5 times their rated pulse current it would seem that we should at least try and keep them cool ! My bet is that they will fail if they were run any where neat their normal thermal limits.