SG3525 or KA3525 Induction Heater / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

SG3525 or KA3525 Induction Heater

1 2 next

Move Thread

LAN_403

Inducktion

Sat Mar 05 2011, 02:54AM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

Hey all. I've successfully created a breadboarded version of a manual tune SG3525/KA3525 Half Bridge induction heater. If my calculations are right, the frequency range is enormous, going from 166 khz, to 666 khz. Supply voltage was about 13 volts for both the 3525 and the gate drive ic. Speaking of, it utilizes a FAN7392 Gate drive ic, doing away with nasty GDT's and inductive spikes. It was easy to lock on resonance, using a 2 uH work coil and a .5 mf mica cap I have. However, I ran into a massive problem. It seems the IGBT's get excessively hot, and forced me to do an emergency shut down on the induction heater with only about 5 amps in, at 12 volts. I used 22 ohm gate resistors on the IGBT's, and scoped the input signals.
The signal's were crisp, and clean, very very little ringing, and inverted on each of the IGBT's. One turns on, the other turns off. But, there's obviously something wrong here. The IGBT's were excessively warm, and I almost managed to kill them. I have them on a somewhat good sized heatsink, both on the same one. (yes, I have insulating pads).
I don't think they should get that hot, or am I sadly mistaken? Do I need to make them seperate? Need a larger heatsink/fan? Do I need to add a Series inductor, something like LCLR topology? I really want this to work, but other than the excessive igbt heating, everything else works wonderously. I don't want my IGBT's to melt off of the heatsink, so what can I do to fix this 4hv?
Danka.

Once this issue is fixed, I'll try to make a mediocre schematic, and give pics of it working.

Marko

Sat Mar 05 2011, 03:34AM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Hello,

Without the mediocre schematic, I can only guess what are you doing. And my guess is that you connected the parallel LC tank straight to the output of the bridge without the matching inductor. If it is so, you get a large cookie from me if you can figure out why your igbt's get hot.

Marko

Hei

Sat Mar 05 2011, 04:05AM

Registered Member #2808 Joined: Sat Apr 17 2010, 03:25AM
Location: Hong Kong
Posts: 11

well may be is the problem that your IGBTs are running under hard switching, which gives large switching lost and so it heat up.

Inducktion

Sat Mar 05 2011, 04:29AM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

WAIT. I figured this out.

The inductor cancels out the current load to the mosfets/igbt's, and thus creates ZVS switching conditions, and longer lifetime for the fets.

It's impedence matching. ...I can use a flyback transformer core apparently. \o/

Just need to wait for my gate drive IC's now... I killed mine somehow, it was resonanting (measured a sine wave on the scope!) with the breadboard capacitance and heated up, and died.

Marko

Sat Mar 05 2011, 06:38AM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

To stay simple, just remember that output of your inverter is *square wave* and you have your tank capacitor connected right across it. The cap is basically a short circuit for fast rising and falling edges, and huge currents are going to be pulled through your igbt's as it charges and discharges. If it's still not clear to you, think of a gate driver, only here you have like 100x the capacitance and 30x the voltage.

Also, don't bother with high side gate drive chips, they just seem to cause trouble for most people. Use a GDT. Outpus of SG3525 can be used as a gate driver for smaller loads and can also be amplified with a pair bipolar totem poles. Just make sure to put schottky diodes across every transistor.

The first time you build a matching inductor, it will probably melt. You need really heavy wire and a big core for any serious power level. Maybe it would be a good idea to use a spool of thickest wire you have around as it is, for an inductor.

Marko

Inducktion

Sun Mar 06 2011, 09:44PM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

There's my mediocre schematic, I apologize if it's hard to read.

That's NOT what I was using for the gate drive ic when the mosfets got hot. The values were 10k for the resistor and it was a 1 uF cap for both of them on that side, when it WAS working.

And now I'm having issues with my gate driver IC's now too, using the values I posted on the schematic. I already tried switching back to the previous values, and it doesn't work still.

I think I just need to add a matching inductor of the right value...and fix the gate drive ic problem, then hopefully, this'll work.

Marko

Mon Mar 07 2011, 04:59AM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

I'm sorry that I can't be bothered to RTFDS now, but there has to be a connection between the driver chip and high side device's source, doesn't it?

Marko

Inducktion

Mon Mar 07 2011, 09:13PM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

Fixed, working on matching inductor now. I got a nice aluminum enclosure for it too. Hopefully that won't heat up (aluminum has low resistance)!

woo.

Marko

Mon Mar 07 2011, 11:14PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

I wouldn't enclose the matching inductor in metal unless it's on an EI core or something with even less leakage. Air or rod inductor will drop in nductance heavily if you put metal around them. Any why do you want to enclose it anyway? It would inhibit air cooling in any case.
c
Marko

Inducktion

Mon Mar 07 2011, 11:23PM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

Marko wrote ...

I wouldn't enclose the matching inductor in metal unless it's on an EI core or something with even less leakage. Air or rod inductor will drop in nductance heavily if you put metal around them. Any why do you want to enclose it anyway? It would inhibit air cooling in any case.
c
Marko

No no no. Silly. I meant an enclosure for the entire THING. But, I probably won't use it. I need something else to put it on though. Hrm. I tried making an air core inductor, with 23 turns of 10 gauge wire, the inductance was only 4 uH. Some calculator on the internet lied. :/

I need some sort of core. Iron Powder wouldn't work so well, from what I understand. I would use a flyback core but THAT doesn't work well at high frequencies...

Prove me wrong at any time please. I appreciate advice and help. :)

Any idea where I could find a good core for this?

1 2 next

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