Apparent inconsistency of MOSFET power rating in datasheet: SOA v. max power

I am trying to use an IPP114N03L G (datasheet).

I calculated the maximum power it could handle as follows:

From the first two pages of the datasheet: Rthjc = 3.9K/W Tj(max) = 175C

Assuming I can keep the case under 50C, max. power dissipation should be: (175 - 50) / 3.9 = 32W

So 1A at 30V should be about the maximum.

Now looking at the Safe Operating Area graph on page 4, at 30V it says ~150mA max. and that is with Tc at only 25C.

Why do these figures disagree?

In practice, I have killed two of these devices now, the last one was at around 20V 200mA, so it seems the first calculation is wrong. The case was cool.