Why am I burning my IGBTs ? / Electromagnetic Projectile Accelerators / Forums

Forums

4hv.org :: Forums :: Electromagnetic Projectile Accelerators

« Previous topic | Next topic »

Why am I burning my IGBTs ?

1 2 next

Move Thread

LAN_403

magictomcat

Tue Apr 29 2008, 09:23AM

Registered Member #1456 Joined: Tue Apr 29 2008, 09:02AM
Location:
Posts: 9

Hi,

I am building a coilgun, using an IGBT to close the circuit, and to empty the capacitor into the coil.
What I don't understand is that I have burnt three IGBT so far...
Here is the circuit I use, does somebodyknow what is wrong in it ?

I guess I burnt two of them with a current that was over their limits. However, I really don't understand why I burnt the third one, which is said to be a N-channel working at 600 V, 300 A at peak current. (According to the usual RLC equations, the current flowing in my circuit doesn't exceed 200 A during 1 ms)

NB : I don't want to use an SCR, I really want to use an IGBT. Could someone explain what doesn't work in this circuit ?
Thanks a lot

Steve Conner

Tue Apr 29 2008, 09:31AM

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

If the battery really is 3.5V, then 3.5V is too low. You want 9V or 12V to turn the IGBT on fully.

The switch isn't debounced. If you don't know what contact bounce is, maybe now would be a good time to find out.

The 47k gate resistor is too high to allow quick turn-on of the IGBT.

magictomcat

Tue Apr 29 2008, 09:44AM

Registered Member #1456 Joined: Tue Apr 29 2008, 09:02AM
Location:
Posts: 9

Hi,

(The battery is 9.5 V, this should be all right)
Thanks for your advice. Indeed, I could enhance my circuit.
Yet, do you think the strong resistor and the bounce effect could burn my IGBT ? Or does it "only" reduce the efficiency of my circuit ?

OZZY

Tue Apr 29 2008, 11:50AM

Registered Member #511 Joined: Sat Feb 10 2007, 11:36AM
Location: Somerset UK
Posts: 55

Hi magictomcat

The switch SW1 could cause you problems.

After charging when you change the switch over to the fire position you are suddenly changing the voltage accross the IGBT from 0V to 330V, this high dV/dt signal could be a problem, It is very unwise to do this to SCR`s I`m not sure about IGBT`s.

Switch SW1 is in the high current circuit, can this switch handle 200A? It may be better to have the capacitor hard wired into the circuit.

OZZY

GeordieBoy

Tue Apr 29 2008, 01:13PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

you are suddenly changing the voltage accross the IGBT from 0V to 330V, this high dV/dt signal could be a problem, It is very unwise to do this to SCR`s I`m not sure about IGBT`s.

dv/dt induced turn-on is certainly possible with IGBT's if you use a 47k gate resistor!

-Richie,

magictomcat

Wed Apr 30 2008, 07:31AM

Registered Member #1456 Joined: Tue Apr 29 2008, 09:02AM
Location:
Posts: 9

what do you think about those:

what is a driver for?
thx

sofi

Sat May 03 2008, 12:07PM

Registered Member #1464 Joined: Sat May 03 2008, 10:56AM
Location:
Posts: 9

Hi, I'm new here, and as first I want to apologize for my english - I'm working on it.
I'm from Slovakia, my hobby is electronics, robotics and all around of coilguns. Last 3 months I try to develop a low power and effective coilgun (kinetic energy 2J, eff. at least 8%), and I focused on multistage IGBT halfbridge design. I spend lot of time while my design start work (partially) and I must react on this problem.
At first, is necessary to understand how IGBT's working - this device combine characteristics of BJT and FET transistors, is controled by voltage betwen G and E.
To switch this device on, the Gate must charge to min. 9 volt, but to fully open is better 12 - 20 volt. As soon it is charged, the switching loses will be smaller. For 100nC charge (common value - see your datasheet) 9.5V and 47k resistor, it will charge after 0.5ms and this is terrible. On my design time of charging on is 1us and charging off 0.5us (this values aren't best). Your IGBT is melt inside (cause switching loses) after you try to turn on SW2.
Gate resistor must be smallest as possible (at other side very fast switching means very large dV/dt and this may cause interference in switching logic) and you need strong current source to control Gate of IGBT. For this purpose is good to use driver (IC, for example IR2101 from IRF), or simple "push-pull" circuit.
Gate is very vulnerable to voltage over 20V - is good to use 18V transil diode between gate and emitor. If you use IC don't forget to decoupling capacitors.
Aplications with IGBT modules are not simple and easy (like SCR).
If you don't understand what I wrote, internet is the good place to improve your knowledge.

magictomcat

Wed May 14 2008, 09:18PM

Registered Member #1456 Joined: Tue Apr 29 2008, 09:02AM
Location:
Posts: 9

thbnk you
Is 18v good enough with 1kohm gate resistor?and can you explain what dv/dt,switching loser and transil diode?
To turn on we can apply minus 18 v and plus 18 for off?
Thx

sofi

Thu May 15 2008, 04:27PM

Registered Member #1464 Joined: Sat May 03 2008, 10:56AM
Location:
Posts: 9

To calculate switching time I using a simple formulas, like I=U/R and t=C/I where U is voltage on gate, R is gate resistor, t is switching time and C is gate charge of IGBT (this is only for estimating, not for exact values). In your case (18V, 1kohm, charge 100nC?) t=5usec - it is slow.

Why you want use 1kohm gate resistor? My IGBT driver deliver max 200mA current (in 12V, therefore my gate resistor is 47ohm), if your driving circuit can supply more current...

The dV/dt is amonut of change of voltage per time unit. Very large changes means very large EM field and thereafter possibility of induced currents in others electronic circuits.

Transil diode is protective component. It is similar to zener diode, but is faster and can absorb more energy.

About switching loses - characteristics of transistor (as switch) belongs to basic electronics knowledge, is a lot of it.

I hope that my post will help to save life some IGBTs

magictomcat

Thu May 15 2008, 05:54PM

Registered Member #1456 Joined: Tue Apr 29 2008, 09:02AM
Location:
Posts: 9

Hi,

thanks for your reply.
I am planning to modify my circuit, so that I wouldn't have this dV/dt problem anymore :
2494545909 1b77bf5245

(if the image doesn't show, try here)

The problem of the current flowing into the gate still remains.
The current flowing through the coil peaks at 400 A (for 1ms). When I toggle the switch from A to B, is there any risk that a high current flows into the gate ?
To to be sure not to harm my IGBT, I am now using a strong gate resistor (R on the sketch), but it implies a high switching time...
To protect the gate, would putting a diode on the "B" wire be efficient (if it cuts the contact between the gate and the emittor, what would the IGBT work like ?).

You wrote about your driving circuit. Mine is the obviously the problem in my circuit.
What is yours like ? Is ther a chance I could use it to control my IGBT ?

Thanks a lot

1 2 next

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