Can anyone please provide me the information on what kind of effects are appearing externally before mosfet damage..

It would be really kind if anyone helps me..I need the external effects causing before mosfet failing.

thanks in advance.

 

Did you read ANY of the links provided? This is all addressed, including a way to autopsy a dead MOSFET to analyze its mode of failure.

 

MOSFETS do not give warning signs except maybe getting hot. They work until they quit, then they just quit working in a tiny fraction of a second.

It is like asking, when I put a firecracker in an oven and heat it up, what are the warning signs that it is going to pop? There aren't any. It just sits there until it pops.

 

Can anyone please explain me the behaviour of mosfet and provide me few youtube links to see practiaclly how the mosfet behaves?

what are the limiting factors?

and why does it break?

http://www.4qd.co.uk/serv/MOSFETfail.html

Limiting factors are listed in data sheet, it'll break when you do what exceeds the data sheet.

I think you just sprained my brain.
It's impossible to answer that except to say, "Stay within the specifications provided by the manufacturer".
The only possible way to test a transistor and be sure not to blow it up is to test for whether it meets the manufacturers guaranteed parameters.

I don't think that "stay within manufacturer's specs" is the end of the answer to the question of "how not to blow up a MOSFET. Datasheets are good at promoting "datasheet coolaid" - fantastical ratings based on theoretical things and hairbrained calculations involving infinite heatsinks and such. The ratings they show pertain only to the silicon, and totally ignore the limitations and parasitic properties of the package. The datasheet specs are out to lunch when you try to apply them to real world. Some people recommend not exceeding 50%-70% of manufacturers specs "just to be safe", but even that is a recipe for failure in some instances. For example, IFR1405PBF: datasheet says 169A max @25C continuous, 118A max @ 100C continuous, and max pulsed current 680A. It's in a TO-220AB package. It says right there on the datasheet in the fine print that package limitation is 75A, but that's more than some datasheets tell you; some datasheets omit this crucial information. So, any number about 75 is totally useless on this datasheet. Really? 75A?. My welder operates around 75A and it needs 4AWG thick copper cable to carry that. Am I really to believe that the cute little leg of a TO-220AB fet can carry the same amperage? I'm not comfortable with passing more than maybe 10A through it, and it would probably be pretty darn hot even at that. So I say that datasheets can be off by at least an order of magnitude, and the thumbrule of 50-75% of max is not even a safe bet in some instances. And that's just the current; how much do you think the gate capacitance changes when you take the package into account? Or any parameter really? I think this is the whole point of the setup he's trying to make, to determine some real world parameters for these fets.

 

I'm gonna have to put a "thanks" on you for the most intelligent post in this thread.
The completely nebulous questions by the OP are impossible to answer, but you found a way to cover at least a little bit of it.

This drawing is a method to test the "Zero gate voltage drain current".

 

I would like to know few answers for external effect of mosfet failure:
Hot Carrier Injection:
- at which power dissipation level does HCI start to be an issue for our MOS-Fets -> literature, maybe testing
- what are the short time effects of a “hot electron” -can it be detected?
- what kind of physical damage processes are caused (over a long time)
- does HCI reduce the power capability of the device, any effect on SOA SafeOperatingArea ratings?
- what device characteristics are changed over long time – how can they be detected from the outside? -> literature and testing

 

This sounds suspiciously like coursework doesn't it?

 

This sounds suspiciously like coursework doesn't it?

I don't know, but it sure doesn't sound like a circuit project.
I had to google to learn about HCI. The links there and the wiki entry had several useful citations that address the questions listed.