IRG4BC20K explosion / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

IRG4BC20K explosion

Move Thread

LAN_403

nrhoades

Fri Aug 17 2007, 06:16AM

Registered Member #610 Joined: Wed Mar 28 2007, 09:44PM
Location: Middletown, RI
Posts: 110

It's late and I'm tired but I wanted to throw this question up on the forum and maybe get an answer by tomorrow...

I built a full bridge using 4xIRG4BC20K IGBTs driven by a GDT from a circuit utilizing a TL494 and double-sided BJT push-pulls (connected to the GDT primary) with a source of 15VDC. The four GDT secondary outputs start looking yucky around 70kHz, so I adjusted the circuit to 50kHz, decent square waves. My bridge has 12V zeners, no gate resistors, and TVS protection diodes, and a 1.8uF bypass cap. I would take pictures but I dropped my Exlim into the ocean when I crashed my sailboat into the rocks

Bridge is mounted on a nice heatsink, is isolated using nylon screws, is electrically correct, yada yada yada...

I was experimenting with the bridge using a source composed of a 20A variac, bridge rectifier, and a 400V 3900uF electrolytic supply cap. I used a huge 120W 465ohm resistor as a test load. I slowly turned the variac up, and when it reached about 90V one of the IGBT's exploded. It was the type of explosion that sounded like a short, and the casing on the TO-220 package is vaporized.

Do you think 50kHz produces too much dv/dt for this IGBT (maybe it latched up)? Could it be because i didn't include gate resistors? (I decided not to thinking that the resistance of the secondary windings were 1 ohm anyway, and would take care of things).

Any ideas of what I should test to find the source of the problem? Again, sorry for lack of pictures, I know they would help.

IGBT datasheet: http://www.irf.com/product-info/datasheets/data/irg4bc20k.pdf

Dr. Dark Current

Fri Aug 17 2007, 06:53AM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

I think the problem is your gate drive, did you make sure the gate waveform doesn't drop below ~10v in on state and doesn't go above ~2v in the off state? Really, I would rather use a gate drive chip instead of a pair of BJT's to drive the gdt and you should really use some gate resistors.
Also the igbts are not exactly fast with total ~400ns turn-off time, try including some dead time (e.g. with a parallel resistor-diode combo in series with the gates).
Hope this helps

Sulaiman

Fri Aug 17 2007, 08:03AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

A few ideas;

1) Does your bridge have diodes across the igbts collector-emitter ?
2) The large resistor is probably wirewound with inductance, to add to the complexity
3) Before re-trying you should find out why the gate waveforms are 'yucky' at 70 kHz
4) You should have some resistance between the GDT drive transistors and the igbts
........could be in the igbt gate connection
........could be on the primary of the GDT
5) assuming the transformer ratio for the GDT is 1:1 from 15 V supply
5.1) your igbt gate zeners are 12V so the gdt, it's drive transistors and the zeners are stressed
5.2) the gate zeners are also diodes so each gdt secondary is short-circuit on negative half-cycles

If you put gate zeners for the igbts then you MUST
either
have a resistor (e.g. 10 Ohm) from each gdt secondary
or
put diodes in series with the zeners to prevent forward conduction of the zeners

nrhoades

Fri Aug 17 2007, 03:43PM

Registered Member #610 Joined: Wed Mar 28 2007, 09:44PM
Location: Middletown, RI
Posts: 110

Thanks guys... I forgot about the IGBT "tail" during turn off, and I think that would cause a significant amount of heating, especially since I did nothing to supress it.

I don't have any more IGBTs, but I ordered a lot of IRF740s (they were cheap, $10 for 20) Things I will change are:

1) Use IRF740 mosfets so frequencies above 50kHz can be achieved
2) Replace 12V zener diodes with 15V zeners to that they are not constantly reverse conducting during half cycles (I noticed that this also amplifies the voltage droop!)

Further questions:
Q1) I didn't mention that I already have 10 ohms in series with the GDT primary, as well as a coupling cap. Should I still add individual small resistors to the gates, and even parallel'd diodes? I'm afraid doing so will add a slope to the edges of the square wave and increase the length of shorted on-time

Sulaiman

Fri Aug 17 2007, 09:40PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

If my understanding of your description of the circuit is correct,
you have four secondaries on the GDT
each secondary has a zener diode directly across it.

If so then I'm surprised your bridge works at all!
The problem of 12V zeners with 15V drive is bad
MUCH worse is that zener diodes are diodes!
The 12V or 15V is the REVERSE (i.e. +ve voltage on the cathode with respect to anode)
breakdown voltage.
In the other direction the zener is a diode (i.e. about 1V)
so two diodes are 'shorting' the GDT on each half-cycle.
This is why I said you MUST either
put a normal diode in series with each zener
OR
put a resistor in series with each GDT secondary

Reaching

Fri Aug 17 2007, 11:53PM

Registered Member #76 Joined: Thu Feb 09 2006, 10:04AM
Location: Hemer, Germany
Posts: 458

i agree with sulaiman, but lol? 70khz from irg4bc20kd???

I used these igbts in my hf sstc at 2,4 and 4,1mhz!!! (Okay, used the irg4b20ud version cause i have 120 of them) they have such a small gate capacitance, around 500pf

nrhoades

Sat Aug 18 2007, 12:16AM

Registered Member #610 Joined: Wed Mar 28 2007, 09:44PM
Location: Middletown, RI
Posts: 110

It's my first attempt with using solid-state power electronics... looks like my chalk mark was a bit too low... I'll take your advise and revise things in a week or so. Thanks

Steve Conner

Sat Aug 18 2007, 10:34AM

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

I can't believe those IGBTs work at 4.1MHz.

I think Richie tried driving the IRG4PF50WD at 4MHz and the tail current was bad enough that it just stayed on all the time and more or less caught fire. W should be a faster grade than U or K, too.

Marko

Sat Aug 18 2007, 04:04PM

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

I can only speculate without schematic, but You absolutely need diodes between collector and emitter. IGBT's without diodes are very sensitive to over voltages and even though you ran them into a resistive load (which could be wire wound and thus with substantial inductance) I wouldn't be surprised if they blew up just from that.

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