I am not a fan of increasing the frequency as a cure because the definition of PWM includes the idea that you eventually get to 99% "on" time. You have to get the design stable at steady state DC operation, regardless of the switching frequency.

The higher the switching frequency, the more times per second the mosfet will have to endure being partially on while current flows through it. Increasing the frequency seems counter-intuitive to me.

 

The LM324 is far too slow and anemic to be used as a gate driver. Your Vgs will look like a triangle wave at 1kHz.

The gate charge for that MOSFET is 200nC. That is a VERY large gate charge, and the Rds(on) at ~0.15 Ohms isn't all that great. You probably got those as surplus.

 

PS, the datasheets are for us. The fact that you have them doesn't help us at all if you don't post them. Try that in the future, It make our end a lot easier.

 

I forgot to say a very important thing , I guess my head is really in bad state.

I tried this circuit first on a small 15v dc motor (when PWM was 5V peak from microcontroller , and when it was 15V peak from op-amp ) and it worked perfectly with no problem at all , would that change something.

sorry I forgot to mention that.

 

PS, the datasheets are for us. The fact that you have them doesn't help us at all if you don't post them. Try that in the future, It make our end a lot easier.

I'll make sure I post them in the future :].

 

The bus capacitor heat up a bit because of the ripple current on them, it's not good in the long run if they heat up too much, but just getting warm won't hurt them..

Not sure about the 50uF...

When the freewheel diode conducts after turning off the FET the capacitor is being discharged, right, then when you turn on the FET you will have a current spike through it, since it was discharged. Correct me if I'm wrong.

Where did you get you the info to put 50uF in series with the FET??

 

Where did you get you the info to put 50uF in series with the FET??

that's another mistake , the cap is across the input voltage , It's a drawing fault.

 

Glad to see the cap move!
Doing the math:
If you could instantly switch the mosfet on, the cap would deliver 160 volts/1.1 ohms, for a start surge of 145 amps plus the motor current. Very smoke if you do that!

 

Doing the math:
If you could instantly switch the mosfet on, the cap would deliver 160 volts/1.1 ohms, for a start surge of 145 amps plus the motor current. Very smoke if you do that!

I couldn't understand this.

but as I said , it worked on a small motor perfectly, so the reason must be related to using the 160v input voltage instead of 15v

 

So the cap is on the 160V power supply voltage. How far is it away? Can you post a picture of your layout?

Do you have an oscilloscope?

 

So the cap is on the 160V power supply voltage. How far is it away? Can you post a picture of your layout?

Do you have an oscilloscope?

no , I don't have an oscilloscope.




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

 

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?

 

please reply , even if someone can't find a good reason for it to work on low voltage only , so that I know how abnormal is this.

 

The voltage part that I can think of is that you said the mosfet was rated at 400 volts. The datasheet says it is rated at 200 volts. Still, the 200 volt mosfet might survive if you get the switching time correct.

Fix the timing and see what happens.

 

The voltage part that I can think of is that you said the mosfet was rated at 400 volts. The datasheet says it is rated at 200 volts.

The one I have says 450V for the 2SK1519.

Where are your electrolytic capacitors? The big ones?

If the switching current passes through long wires, the inductance of these wires also create a voltage spike. The electrolytic caps have to be as near as possible to the cathode of the motor freewheel diode and the source of the MOSFET.

Have a look at the attached picture, it's for IGBTs but also applicable to MOSFETS. The busbar parasitic inductances have to be minimized.

 

Where are your electrolytic capacitors? The big ones?
\

should I instal other caps , where should they be installed in the schematic , the only cap I have is the 50uF , 400v one .

 

should I instal other caps , where should they be installed in the schematic , the only cap I have is the 50uF , 400v one .

Where does the voltage come from? How long are the cables to your power supply?

It's getting difficult without an oscilloscope...

 

It comes from a transformer , then a rectifier , the cable are about two meters long.

the rectifier is just like this one
http://t3.gstatic.com/images?q=tbn:ANd9GcSuGf4eSMwxv5sd2il_9IXuYWPTZjzeiKKRgv195hSuPVFXw3cw


if it is very important to have an oscilloscope , I can try to buy one.

 

It comes from a transformer , then a rectifier , the cable are about two meters long.

So essentially you don't really have DC-bus capacitors.
IMO you need to add low-ESR DC-caps the way I described before.

Please have a look at this paper.

if it is very important to have an oscilloscope , I can try to buy one.

If you have other projects to do, this will help you very very much.

 

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

and is not a dc signal , that also would explain the capacitor heating and the mosfet breaking , is that what you mean?

I thought that this rectifier contains the capacitor in it , that's really a stupid mistake.