OMG , does that mean that the signal I have is still like this shape

http://upload.wikimedia.org/wikiped...ifier.en.svg/600px-Gratz.rectifier.en.svg.png

Not exactly, since you HAVE a capacitor, but I think it's too small to supply the current you need for the motor, so the rest of the current has to come through the 2x2m wire (inductance) which will cause a voltage spike when the FET is being turned off..

This is what I imagine could be happening, you can only be sure if you measure it with an oscilloscope.
Don't you have some electrolytic caps lying around, something like 220uF/250V or 470uF? Even if they are not suited for this application (they may heat up too much to last for a long time) you should see a difference. Although I'm not sure how you will find out what exactly happens if you have nothing to measure. :-/

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=495-4259-ND

 

so , should I connect , say , a 1000uF , 400 volt cap accross the two terminals , would that be sufficient.

 

If the cap is small , the charging slope would be high and the output wouldn't be a good dc signal , Isn't that what you mean?

http://upload.wikimedia.org/wikipedia/commons/thumb/a/a3/RC_Filter.png/400px-RC_Filter.png

 

and do I have to use a mosfet driver , I think at this low freq. (1KHZ) there would be enough time to charge the gate without the need of a mosfet driver.

Do you have access to an oscilloscope so you can see how the GS waveform looks like? Sounds like it stays half saturated a lot of the time. Does the mosfet get very hot?
Does it get less hot if you lower the PWM frequency?

-- K

 

can't any one think of a reason for it to work if the input voltage is 15V , and burn the mosfet when the input voltage is 160V?

I don't have much details about your motor but as per your circuit and your post ,it may be like this.....

Many new designers miss this factor while designing circuits power dissipation of the device,even in simple passive components like resistors.
You burn it,you learn it....

In the datasheet of K1519 the channel dissipation is about 200 watts at 25°C and less at higher temperatures.Now you are using a DC motor of 160V and about 15Amp ,(160*15) = 2400 watts ,it's power is quite large even if you are using PWM which cause the motor to use less power then also I think the power will be quite large.

Good Luck

 

If the cap is small , the charging slope would be high and the output wouldn't be a good dc signal , Isn't that what you mean?

http://upload.wikimedia.org/wikipedia/commons/thumb/a/a3/RC_Filter.png/400px-RC_Filter.png

Since we don't know why it dies, wether because of overcurrent or overvoltage we have to rule out every possible cause.

A 50uF in my opinion is not sufficient to provide the current necessary, especially not when you are near the valley of the rectified voltage. So current has to pass through the bus bar inductance which is a 2x2m(!) wire. When you turn off the PWM this inductance will try to continue the current flow in the same direction therefore it will create a voltage spike which has it's "+" at the freewheel diode cathode and it's "-" at the MOSFET's source.

If you want to put a 1000uF/400V you can do that, make sure you don't exceed the fullbridge rectifier rating when turning on mains. (maybe do a simple softstart with a NTC on the transformer's primary).
You may also put a MKP capacitor 470nF with high pulse strength, for example the B32686/B32656/B32654 series from epcos in parallel with the electrolytic cap.

What type is the 50uF capacitor you installed?

If this still doesn't help you can also install an RC snubber/ a small (470pF max) capacitor from source to drain or a RCD voltage clamp.

Problem is still, without oscilloscope it's all just guessing.

As others stated you also have to consider that the MOSFET current handling capability is exceeded, even though their peak current capability is quite high.

So far you have a few things to change:

-Charge the gate faster with a mosfet driver and a smaller Rgate (will decrease switching losses
-install electrolytic caps as near as possible to your board from cathode of the freewheel diode to source of the MOSFET

 

In the datasheet of K1519 the channel dissipation is about 200 watts at 25°C and less at higher temperatures.Now you are using a DC motor of 160V and about 15Amp ,(160*15) = 2400 watts ,it's power is quite large even if you are using PWM which cause the motor to use less power then also I think the power will be quite large.
Good Luck

But is the power dissipation , the max power the mosfet can endure , the absolute voltage rating is 450v , absolute current rating is 30A , that means in this absolute conditions , the power of the current passing through would be 450 * 30 = 13500W .

I also checked another mosfet with 1000V,5A absolute rating , I would expect its max power enduration to be 5000W , but the data sheet says power dissipation = 150W.
http://www.irf.com/product-info/datasheets/data/irfng50.pdf

doesn't that mean that this power dissipation value is the max allowable power lost from heating because of channel resistance.

 

What type is the 50uF capacitor you installed?

it's electrolytic .

I will try this and see.

 

You don't have any heat sinks on the MOSFET or the diode. They will be dissipating lots of power during switching/conducting. You need to have heat sinks on them, or they will burn up very quickly.

You've run traces at odd angles. I generally try to avoid doing that where possible. I keep angles divisible by 45°.

note : not all components in the layout is true , if i don't find the component in eagle , I use anything with its dimentions.

That is fine, but you should use the NAME tool and put in the actual part number instead of what is there.

 

But is the power dissipation , the max power the mosfet can endure , the absolute voltage rating is 450v , absolute current rating is 30A , that means in this absolute conditions , the power of the current passing through would be 450 * 30 = 13500W .

I also checked another mosfet with 1000V,5A absolute rating , I would expect its max power enduration to be 5000W , but the data sheet says power dissipation = 150W.
http://www.irf.com/product-info/datasheets/data/irfng50.pdf

doesn't that mean that this power dissipation value is the max allowable power lost from heating because of channel resistance.

Losses transformed into heat are mainly switching losses and losses due to the On-resistance. Multiplying Vmax with Imax doesn't make sense. You mentioned you mounted the Mosfet on a heatsink but it can't be seen on the layout.

 

That is fine, but you should use the NAME tool and put in the actual part number instead of what is there.

OK .



I also did install a heatsink .

there is also a 1Mohm resistance I installed between gate and source , to prevent the gate from being floating , based on the advice of "The Art of Electronics" , but I don't think that would affect this problem .

 

Losses transformed into heat are mainly switching losses and losses due to the On-resistance. Multiplying Vmax with Imax doesn't make sense. You mentioned you mounted the Mosfet on a heatsink but it can't be seen on the layout.

this method is the one debjit625 used to tell me that the mosfet doesn't satisfy my power requirements.

does that mean that the max power the mosfet can endure is actually 200W .

I did install heatsink , but I don't know how to put it in the layout.

 

Do you have access to an oscilloscope so you can see how the GS waveform looks like? Sounds like it stays half saturated a lot of the time. Does the mosfet get very hot?
Does it get less hot if you lower the PWM frequency?

unfortunately , I don't , the mosfet does get very hot when I put the 160 volts , but with the 15v , it doesn't at all

 

But is the power dissipation , the max power the mosfet can endure , the absolute voltage rating is 450v , absolute current rating is 30A , that means in this absolute conditions , the power of the current passing through would be 450 * 30 = 13500W .

I also checked another mosfet with 1000V,5A absolute rating , I would expect its max power enduration to be 5000W , but the data sheet says power dissipation = 150W.
http://www.irf.com/product-info/datasheets/data/irfng50.pdf

doesn't that mean that this power dissipation value is the max allowable power lost from heating because of channel resistance.

You got it all wrong man....
That datasheet says that
Absolute Maximum Drain to source voltage at 25°C of temperature can be 450V i.e.. you can apply up to 450 Volt DC from Drain to source. As its an enhancement mode MOSFET these are normally off ,so it says with zero Vgs ,if you apply 450V at Drain to source it will not conduct much current.It is also known as drain to source break down voltage after you cross this limit the normally off channel will starts conducting a huge current and destroy the MOSFET.

Now the other thing is the Id (Drain current) it says that max at 25°C the channel can pass 30A and if peak current i.e.. pulse like yours it can be 120 Amp, check for the pulse width it should be smaller or equal to 10 micro sec and duty cycle should be smaller or equal to 1%.

All the above ratings are considered if the MOSFET is at 25°C,if the temperature increases then the rating decreases and its very common that your MOSFET will be not at 25°C it will be higher than that.

Now the channel dissipation is 200 watts any more than that it will be destroyed. So your Vdss and Id must satisfy that.

For example a 3 ohm resistor with 12 volt drop across it with 4 Amp of current through it, will dissipate 48 watt, so for the circuit you need a 3 ohms 48 watt resistor ,if you use a 3 ohms 1/4 watt resistor than it will burn up. Now here is the fun the same 3 ohms 48 watt resistor can have 4 volt drop across it with 12 Amp running through it.

Now I can say that the voltage across the resistor can be 12Volt and the current can be 12 Amp but not at the same time and it still dissipate 48 watts of power not (12 * 12 = 144 watts).

Good luck

 

You got it all wrong man...

that means that with my 160 volts , I have a max current of 1.25A.

so what should I do , Is there a mosfet that can take 160*15 = 2400W , or should I connect more than one of them in parallel .

 

what about this .



It says the max power the channel can dissipate , not endure , and it doesn't say that I have to make sure that the multiplication of the current and voltage doesn't exceed power dissipation value .

 

that means that with my 160 volts , I have a max current of 1.25A.

?? No.

Maximum power dissipation is given for and depends on the maximum allowable die temperature if the MOSFET's case is maintained at 25 degree C.
It depends on the thermal resistance from die to case.

http://www.aosmd.com/res/application_notes/mosfets/Power_MOSFET_Basics.pdf
http://www.irf.com/technical-info/appnotes/mosfet.pdf

Please have a look at THIS guide for power MOSFET loss calculation

 

and it doesn't say that I have to make sure that the multiplication of the current and voltage doesn't exceed power dissipation value .

Because that's not the way it's done.
I used IGBT's with a max breakdown voltage of 900V and max. current of 51A@25°C in a TO247 package. Would they have to be able to dissipate 46000W ? No. Forget this, multiplying these values doesn't mean anything.

You have to make sure that voltage spikes don't exceed the breakdown voltage Vds and that you don't exceed the maximum continuous collector current which translate into maximum die temperature. What size is your heatsink? It may not be big enough, since the MOSFET gets HOT. If I look at the datasheet the RdsON is 0.25Ohms @20A @120°C case temp, so this alone can't be the reason.

Switching losses are probably also increased because of your huge gate resistor and there is the question of the busbar inductance which may lead to voltage spikes.
Address one problem after another and you will get to the problems source.

 

Because that's not the way it's done.
I used IGBT's with a max breakdown voltage of 900V and max. current of 51A@25°C in a TO247 package. Would they have to be able to dissipate 46000W ? No. Forget this, multiplying these values doesn't mean anything.

You have to make sure that voltage spikes don't exceed the breakdown voltage Vds and that you don't exceed the maximum continuous collector current which translate into maximum die temperature. What size is your heatsink? It may not be big enough, since the MOSFET gets HOT. If I look at the datasheet the RdsON is 0.25Ohms @20A @120°C case temp, so this alone can't be the reason.

Switching losses are probably also increased because of your huge gate resistor and there is the question of the busbar inductance which may lead to voltage spikes.
Address one problem after another and you will get to the problems source.

So , I didn't understand , will this mosfet work under these conditions or not ?

according to debjit625 , It has no chance of survival , does your words mean it could work if I fix those other problems.

 

Hard to say if it can or can't without doing some measurements. For your project I would at least measure the voltage with an oscilloscope and the temperature with a temperature probe. I also would measure the peak current putting a 1R resistor at the source, but then I would also need an oscilloscope.
I think with a proper heatsink it has a chance of survival, yes.
If you want, put two MOSFET's in parallel (yes you can do this since they have a positive temp coefficient).
If your problem is current, it will be solved. If your problem is voltage then you'll have two dead MOSFETs.

You still didn't show/say what's the size of your heatsink, I suspect, it's too small.