Hi guys!
I am not so familiar with mosfet's so I have some questions I hope you can helpe me with. I want to use a power mosfet as a switch in a buck regulator to be able to transform a lot of effect. So far I have experimented a little with my design and found that my mosfet gets really really hot, and eventually it will blow. This happens even though I am not even close to the ratings of the mosfet. Mosfet is supposed to be able to handle 200v and 50amps, and I am "only" using 100v, 2amps.
So, is it normal for a power mosfet to get this hot?
So far I have used a small heatsink on the mosfet.
I am grateful for all tips regarding how to prevent the mosfet from blowing again.

 

Hallo,

When you would post a schematic over here, we could have a look what can be improved.

Bertus

 

Edit: Obviously you were above its power ratings.

Chances are you are not properly switching the MOSFET. They must be rapidly switched from fully ON to fully OFF and back to minimize heating. But we need a schematic, as bertus stated, to make any suggestions.

 

Thanks for the super quick reply!
Here is a part of the schematic (if I did this right) showing the buck circuit, the pwm modulator (on the bottom) and the gate driver (on the right side). The power mosfet used is named IRFP260MPbF and is driven by a 50k Hz pwm signal.

 

Edit: Obviously you were above its power ratings.

I am sorry but I don't understand..

 

The exact part number you use may impact your success or failure (or the failure of the mosfet. Since voltages are not labeled in your drawing, i assume the chips ate running at 5 volts and the voltage supply is something higher.

If I assumed correctly, you will need a MOSfet that can be switched at "logic level" voltages (5 volts). Most standard MOSFETS need 10 volts to drive the gate. If you are using a standard mosfet and only supplying 5 volts, the mosfet will not be on completely and start getting hot, really hot depending on the current actually allowed to flow through the partially on mosfet times the. Voltage drop across the mosfet = watts of heat the poor chip must try to share with its environment.

 

What's the part number for the gate driver chip?
The 130 ohm series gate resistor value seems too high for driving at 50khz.

 

I am sorry but I don't understand..

All transistors has a power rating based upon their thermal resistance to the ambient environment. Since you transistor got too hot and failed it obviously was trying to dissipate more power (or more exactly exceeded the maximum junction temperature) then it was rated for. The power dissipated in a transistor is simply the current through the device times the voltage drop across the device when it is ON.

 

The mosfet is switched at 20v thanks to the gate driver (uc3807). The resistors between gate driver and mosfet are found from testing with different values, and the values used are those that gave the best gate signal waveform.
So if I understand you guys right, it is not normal for a power mosfet to get overheated at these conditions. And therefore a larger heatsink will not solve my problem?

 

There is no bootstrap cap in your gate driver. Cant find UC3807. UCC3807, which was found, is a current mode pwm controller 8Pin IC.
Your gate needs about 10V higher than the source to turn on.
Maybe, the gate driver detail?
Oh, just found UC3707. Is it what you are using?

 

UC3707 is not the right one here.
I am assuming that you are using around 100V input.
You need to use a gate driver with a bootstrap, like IR2117, which have a connection to the source of the MOSFET that you are driving, and uses a bootstrap cap to get a voltage higher than source (which is not at your circuit ground).

 

I am sorry the gate driver is UC3708.
And the Vcc is 20v

 

The answer is the same as that of UC3707.

 

Thanks for reply,
I do not really understand why I need a bootstrap gate driver.
Isn't the gate voltage higher than source when all parts of the circuit uses the same ground, and gate driver output voltage is 20v and source voltage is 14v ?

 

Thanks for reply,
I do not really understand why I need a bootstrap gate driver.
Isn't the gate voltage higher than source when all parts of the circuit uses the same ground, and gate driver output voltage is 20v and source voltage is 14v ?

The bootstrap is needed. If your source voltage is 14V and your gate voltage is 20V; 20v - 14V = 6V. Your only putting 6V on the gate, barely turning the mosfet on. The boot capacitor allows the gate voltage to be 20v higher than the source.

Here's a good link on boot strapping a mosfet driver; http://www.fairchildsemi.com/an/AN/AN-6076.pdf

 

Thanks for the great tips RamaD and shortbus,
I will read the document you posted later.
I have searched around a bit and found that it is possible to have the mosfet connected to the "low side". Doing this my circuit will end up with different ground points for input and output and this makes things more complicated. Especially considering the voltage feedback to the pwm controller.
What do you think about this idea?
Is this more trouble than a bootstrap gate driver?
And if I use this configuration will I still need a bootstrap gate driver?

 

Thanks for reply,
I do not really understand why I need a bootstrap gate driver.
Isn't the gate voltage higher than source when all parts of the circuit uses the same ground, and gate driver output voltage is 20v and source voltage is 14v ?

The source voltage is not 14V. When the transistor turns on, the source voltage should rise to equal the drain voltage for proper switching of the voltage to drive the inductor (in that state the input inductor voltage is higher than the 14V output voltage). To do this with an N-MOSFET, the gate must be at a higher voltage than the source (and drain), which requires a bootstrap circuit.

Alternately you can use a P-MOSFET switch with the source connected to the 20V. In that case grounding the MOSFET gate will fully turn on the MOSFET. (But remember to observe the maximum gate-source voltage rating for the transistor you use).

 

Thanks for good explanation crutschow!
But before I start looking for a bootstrap gate driver that can fit in my design I would appreciated some opinions and tips on how to connect the mosfet on the low-side. This will creates problems with my voltage feedback to the pwm controller.
As far as I understand it would then not be necessary with a bootstrap gate driver?

 

Thanks for good explanation crutschow!
But before I start looking for a bootstrap gate driver that can fit in my design I would appreciated some opinions and tips on how to connect the mosfet on the low-side. This will creates problems with my voltage feedback to the pwm controller.
As far as I understand it would then not be necessary with a bootstrap gate driver?

The problems you mention and the complicated circuit required to solve those problems are why a low-side switch is seldom (if ever) used for a buck converter.

 

Thanks again crutschow.
So than I guess I am left with the choice between a p-channel mosfet or a bootstrap gate drive? I have googled for different bootstrap ICs that can fit for my specifications (50 to 300v input) but so far I have not found any better than the one RamaD mentions (IR2117). Do you know of any others that may be more suitable?
(Am I terrible at searching or are there not many bootstrap gate drivers for high voltage applications?)

Or is it possible to use the gate driver I already have and just add a capacitor, resistance and a diode?