MOSFET power measurement

[xzrjk1]

Hello everyone!

I experience some troubles with a stepup converter for my power electronics course. Here are the general informations: P=75W; Vin=13.5V; Vout=48V; Iout=1.5A; f=500kHz. The simulation was successful and calculated the mean losses (off+switching+on) per periode of the MOSFET with about 2W. I got this result by simulating the drain source voltage across the MOSFET as well as the drain current and multiplied them to get the losses. Then I simply calculated the mean over 1 periode.
Now, I attempted to do the same with the experimental setup, but I recieved a mean loss per periode of about -3W across the MOSFET. The current is partly negative, I assume due to the body diode and the MOSFET is of course getting pretty warm.  So, I am not sure how to objectively compare the measurement results with the simulation results. Please find my measurement results in the attachment.
I am grateful for your help and I hope someone can explain this to me.

Thank you!

[nuno]

Don't forget to simulate layout and passive component parasitics.

[owiecc]

How do you do simulations? What models do you use? Do you simulate nonidealities in the circuit? Do you include component temperature rise in your model?

How do you do measurements? What do you use to measure current and voltage? How fast is you mosfet (switch time)?

[xzrjk1]

Thanks for your replies!

I didn't simulate the layout influences on the circuit.

I did my simulations with the models provided by the suppliers. It included the driver IC, the mosfet, the diode and the output capacitor.
I measured the voltage with a differential probe of an oscilloscope directly across the mosfet. For the current I used a 0.01R shunt resistor between mosfet source and power ground. The plot attached was done in matlab with the csv-files aquired from the measurements. The mosfet (FDMS86255) has a typical rise-time of 4.5ns (10ns maximum) and a fall-time of 6.2ns (12ns maximum).
Maybe I should really consider to include temperature changes into the simulation?

Thank you! 

[owiecc]

What is the bandwidth of your voltage probe and scope? If you are using a galvanic connection to the scope why don't you use a passive probe? Ideally probe would be connected to PCB with an adapter (figure 4.11 of Tektronix Probe Fundamentals).

For current probe can you give some details of the resistor used? I know guys doing GaN characterization use plenty of 0201 resistors in parallel to reduce added inductance. We use T&M SDN series for SiC loss measurement.

Do you have some pictures of you setup?

You can try including parasitics and junction temperature effects in your simulation but it is not very easy to do.

[nuno]

The input capacitor plays a relevant role in the behaviour and should also be simulated with parasitics. I don't know how to do the complete analysis, but noone will do it with only the information you provided; MOSFET driver circuit, circuit topology, layout, simulation results (are we suppose to compare the measurements you provided with what?), ... And things such as "typical rise-time of 4.5ns (10ns maximum)" are completely useless, as rise/fall times depend on how you drive your FETs.