Power Switching Problems

I don't know the specs for these devices you're using but maybe the 380V supply has some spikes that are exceeding the voltage rating of the MOSFET???

John.

thanks for your comment. But it's all in order before 2 MOSFET broken dow. I don't know how to upload the schematic diagram.

Over heat and reactive load may cause these effect. If you wish can send me an schematic at tecnologia@arnet.com.ar.

At a quick glance the device looks OK for what you are doing but we really need the circuit diagram and details of the load. If you wish to post the circuit you can by saving a copy of the diagram as an image (BMP, JPG, GIF) on your system then when you add a comment click the icon that looks like a camera (6th from the left) on the comments toolbar and use the box that come up to insert the image in the coment..

Years ago, when I was using SCRs and Triacs in 380Volt circuitry, I always used devices good for at least 1000Volts as anything less broke down before long. Even the 1000Volt devices very occasionally were affected.... As a previous comment mentioned, the spikes can be killers of semiconductor devices, although I must admit that I personally have no MOSFET experience in such areas.

Ask yourself, have you really got MOSFET devices that can handle such reverse spikes of 1000Volts or more?

According to the data sheet (see image below) the device you are using has an 800V zenner diode between the source and drain to protect it from spikes.

But it also means that if the source goes positive with respect to the drain the zenner will be forward biased and will go on. I do however agree that you may be cuting it a bit fine with 380V AC and the 800V limit for the device and is the 380V you speak of RMS or peak becacue if its RMS then you are tlaking about 432V Peak which will almost certainly produce spikes over 800V.

We really need to see how you are using this device to understand what is going on.

Maximum gate source voltage for this MOS transistor 17N80C3 is only 30V. If you exceed by few extra volts then the insulation will be punctured. Perhaps it fails when you are starting from zero. Use a riding pulse with slow start and then increase the voltage. This means when output is already increasing at the same time you increase in the input voltage and do not try full 31V pulse at the gate when Source is at zero voltage. I think you have not taken care of this slow start limitation of this transistor.

I do not see any other reason for it to fail except for gate damage, which is very easy to do. PWM will not do. You need increasing pulse height in the pulse train. See the scheme below. R2 will reduce the input voltage to gate only when output is low. You can also replace R2 by a zener to clip the pulse voltage to only 5.1V to the source. Gate source voltage is usually kept at constant control limit and never used at Maximum limit to blow the parts, which you have tried now.

thanks for commens. It's very useful for me.This is the schematic diagram.

If you want maximum current from FET to be about 50A then limit the gate source voltage to 10V using a zener between gate and source and not gate and ground. Do not exceed this limit.

Another thing is that this FET does not require negative gate voltage to shut OFF and goes to uA drain source leakage current when gate source is at zero voltage. Injecting high negative pulses will kill the gate. Most of the FETs otherwise require negative voltage. This FET is like a NPN transistor and does not operate at zero voltage. Perhaps it was designed as an IGBT and called a MOSFET.

Avoid inductance and capacitance in gate source area. Charging the gate will make it in ON state for much longer period.

Resistance between gate and source used by you can be 10V zener. This will also remove that negative voltage effect. Only input resistance to gate is required and diode is not required. Zener will allow 600mV negative which is enough to shut off the transistor.

MOSFETs are all special devices and you need to look at the catalog carefully.

I didn't connect to load. but I replaced 4 MOSFET by 4 IRFP 460. Before I used these IRFP, it's operated normal.

IRFP460 requires much lower Gate source voltage (perhaps about 6-8V range) and also works much like NPN transistor with about 4V threshold voltage. No negative voltage required.

but I used IRFP for many power supply like that. all of them are in order now. I think has problem with my power now but I didn't determine the cause.

Well, if you design properly then it gives lots of margin for error else may fail will small change in control signal. IRFP has maximum gate voltage of +/-20V and Other one +/-30V. Yes, IRFP are more serious designs and can take some extra banging but will lose their life for sure.

I design for 30 years life span and not for just for a day or two. I also take care for adverse conditions and my instruments work 365 days 24- hour on condition for over 10 years. Only things fail are capacitors as they are not for longer life. I prefer to use hirel capacitors also.

As an engineer, you need to plan all possibilities that are realistic to keep things safe.

I feel first of all you should check your transformer to see if it gets satisfaction with your circuit frequency! If the transformer is designed in way of F1, but you work at F2, you will fail to work at F2 if F2<F1. Of cause the device will be broken by current or say heat. If your transformer bias for some thing, say, aymmetry active pulse. it obviously fail. The 460 is rated stand 600voultage and 20A I remember. The 380v will a bit higher for its use. so you should control the supply not greater over 300v for safe working. On other side, you would put the diode parallel with resistor between grid and source in otder not to broken the mos transistor reversely.if you hope to accelerate on/off, you can use a capacitor parallel with the ressistor insteading of a diode. then you can also apply capacitors on the drain and source of mosfet as a buffer fo r spikes

feel first of all you should check your transformer to see if it gets satisfaction with your circuit frequency! If the transformer is designed in way of F1, but you work at F2, you will fail to work at F2 if F2<F1. Of cause the device will be broken by current or say heat. If your transformer bias for some thing, say, aymmetry active pulse. it obviously fail. The 460 is rated stand 600voultage and 20A I remember. The 380v will a bit higher for its use. so you should control the supply not greater over 300v for safe working. On other side, you would put the diode parallel with resistor between grid and source in otder not to broken the mos transistor reversely.if you hope to accelerate on/off, you can use a capacitor parallel with the ressistor insteading of a diode. then you can also apply capacitors on the drain and source of mosfet as a buffer fo r spikes

To find an answer you must investigate a circuit characteristics with oscilloscope help under low input voltage (50V-300V, grounded PS). Take into consideration an overvoltage on gates and overcurrent in power switches during ON-OFF and OFF-ON transitions.