How to validate the repetitive peak pulse current rating for a given MOSFET?

[PChit]

I was planning to use STP15N95K5, for one of the power converter application. I have RMS current for the switch around 4 A with repetitive peak currents around 30A (0.2 duty cycle with a frequency of operation of 140kHz). The switch operates at a DC voltage of 650V.

I am very confused about calculating the peak current. Whether the switch is capable of operating at the specified conditions. Can you please let me know if it is possible to use this particular switch with the calculations?

[ATENE]

​Hello,

With such a op freq, 140kHz -> pulse duration is about 7us.

Given 30 Amp and 650V as working parameter we are clearly out of SOA, as easily to be seen in the graph

Doing some calculation: 

For the correction factor K calculation using the formula K(t)= K(t1) [(t/t1)exp1/2] starting from the last readable value on thermal imp curve K(t1) = 0.04 at 10exp-5 sec

So, K(7us) = 0.04 (7*10exp-6/10exp-5)exp1/2 = 0.04*0.836 = 0.028

Thermal dissipation equation, 

V*I = (Tj-Tc)/Zth with Zth = Rth*0.028 = 0.0335

(Tj - Tc)= 650*30*0.0335= 652°C, as forecasted NO GOOD.

We suggest to use a bigger MOSFET in a package with better thermal dissipation behavior such as TO247, e.g. STW40N95K5 (or even STW40N90K5)

e.g. with STW40N90K5

K(7us) = 0.058 *(7*10exp-6/10exp-4)exp1/2 = 0.01534

Zth = 0.01534*0.28 = 0.004296

Tj - Tc = 650*30*0.004296 = 83.8 °C -> OK

Best Regards

Alberto Tenerello

High Voltage MOSFET BU, Technical Marketing

[PChit]

Hello Alberto Tenerello,

Thanks a lot for your reply. Can you please let me know how did you consider the K parameters 0.04(for STP15N95K5) or 0.058(for STW40N90K5).

Just for clarification, did you consider it for duty cycle of 0.02? (The actual duty cycle was 0.2)

Also are we suppose to consider just the (V*I) product of the MOSFET or the actual Power dissipation at MOSFET?

Best Regards,

Prajwal CP

[PChit]

Hello Alberto Tenerello,

Thanks a lot for your reply. Can you please let me know how did you consider the K parameters 0.04(for STP15N95K5) or 0.058(for STW40N90K5).

Just for clarification, did you consider it for duty cycle of 0.02? (The actual duty cycle was 0.2)

Also are we suppose to consider just the (V*I) product of the MOSFET or the actual Power dissipation at MOSFET?

Best Regards,

Prajwal CP

[ATENE]

Hello Prajwal,

sorry I misread the ​data. Actually the condition is even worse than in my calculation.

I considered the worst condition in which the peak current is constant during all the ON pulse duration but if you know exactly the duration of the Ipeak you can do a more refined calculation of the Tj the MOSFET reaches in the worst case. We cannot consider the only RMS value as it is not indicative of the effective working conditions and real thermal dissipation capability of the MOSFET.

Probably it is the case to put a current snubber in your design that will allow you to use a smaller MOSFET.

Best regards

Alberto Tenerello

[PChit]

Hello Alberto Tenerello,

Thanks for the suggestion.

Best Regards,

Prajwal CP