Odin The All-Fragger: a large DRSSTC / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

Odin The All-Fragger: a large DRSSTC

first previous 13 14 15 16 next last

Move Thread

LAN_403

Steve Conner

Sun May 11 2014, 01:26PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Dummy load test

The load consisted of the largest metal tray I could lay hands on. Even so it doesn't seem to load the tank circuit very well. The current limiter returns almost all of the energy to the DC bus and the coil draws hardly any current from the mains. This is good as it shows we have a nice low-loss primary circuit.

The current limiter is set for 750A, but the actual peak current seems to go to maybe 1100 or 1200A. This is because it only operates in increments of 1 cycle. If it trips on the first half-cycle, the second half-cycle will still be active and will ring it up even further. I expect this to calm down somewhat when loaded by the secondary. Nevertheless, the 200A IGBTs seem to handle it fine.

Thomas W

Sun May 11 2014, 02:25PM

Registered Member #3324 Joined: Sun Oct 17 2010, 06:57PM
Location:
Posts: 1276

Awesome, cant wait to see it when its done..! What happens if you put a magnet in the coil? i wonder how it would react, ahah.

Hydron

Sun May 11 2014, 02:36PM

Registered Member #30656 Joined: Tue Jul 30 2013, 02:40AM
Location: UK
Posts: 208

Looks good! I've used big cast iron pans as dummy loads in the past, though for a smaller coil than that. Even cooked a couple of eggs once (~2kW for >10 min, good thermal test of the primary circuit!).

Do you get much coupling into the aluminium box lid and frame? They seem fairly close to the primary given they're on-axis.

Steve Conner

Sun May 11 2014, 02:44PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Given that no sparks shot out of the corners of the box lid and it didn't catch fire, I think the coupling must be acceptable.

The field strength in this test was similar to what it will be when the coil runs at full power later.

I added the legs to the underside of the primary mainly to reduce coupling, after having problems with it in my previous DRSSTC. Eddy currents in the metal frame caused some impressive arcing and burning at the corners.

Avalanche

Sun May 11 2014, 04:17PM

Registered Member #103 Joined: Thu Feb 09 2006, 08:16PM
Location: Derby, UK
Posts: 845

Is that running with those SKM200GB128 devices? Either way, looking mighty impressive so far

I was wondering if you were going to show a shot of the tray of boiling water towards the end

Steve Conner

Sun May 11 2014, 05:52PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Yes it is the SKM200GB128s. Sadly there was no water in the tray as it was too shallow and badly warped to hold any. Must try again with a dustbin lid.

Artlav

Sun May 18 2014, 06:25AM

Registered Member #8120 Joined: Thu Nov 15 2012, 06:06PM
Location: Moscow, Russia
Posts: 94

About that IGBT brick driver of yours - do you have any write-up on it's design?

Specifically:
-Why +24, -12V? The gates are almost universally +-20V absolute maximum. Is that safe?
-Why GDT on the input, instead of optocoupled logic level signals and on-board driver for the FETs? I.e. HCPL3120 which is two of these in one.

Steve Conner

Sun May 18 2014, 09:56AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

This thread contains details of the gate driver design. You can read some more about it here.

Overdriving the gate voltage isn't exactly "safe", but then neither is running the device at 3x its maximum pulsed current rating.

+24/-12 is safer than the +/-24 used by other DRSSTC builders.

The GDT approach is simpler, cheaper and faster than an optocoupler-based circuit. The total delay through my gate driver is less than 200ns. The only reason to use optocouplers for gate drive is if you need to pass a low frequency or DC component that a GDT couldn't handle.

Artlav

Mon May 19 2014, 08:16PM

Registered Member #8120 Joined: Thu Nov 15 2012, 06:06PM
Location: Moscow, Russia
Posts: 94

Steve Conner wrote ...
Overdriving the gate voltage isn't exactly "safe", but then neither is running the device at 3x its maximum pulsed current rating.

Hm, i thought running over the maximum pulse current was mostly a heat dissipation issue (within fusing reason), while overvoltage on the gates was a breakdown type of issue.
First is radiated away and you're good as new, second is eroding away and you breakdown and go BOOM.
No?

Steve Conner

Mon May 19 2014, 09:15PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Yes and no. The greater the junction temperature rise during the burst, the shorter the life of the device will be. There are fatigue mechanisms that set in far below the 200 degree delta-T that would be required to fuse it in a single shot. For example, in railway traction applications the bond wires often fall off the die.

Overdriving the gate will lower conduction losses and help the device to carry more current before desaturating, so in extreme cases it may reduce the temperature rise during the burst.

The effect is to lengthen the device's thermal fatigue life while shortening the life of the gate dielectric. The optimum gate voltage would be when both lifetimes were the same. I have absolutely no clue what it is, but everyone else seems to use at least 24V.

first previous 13 14 15 16 next last

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob