LTSpice Power Estimation of LTC6363 FDA

[davegravy]

I'm trying to model power dissipation for this IC and getting questionable results. I tried to reproduce the scenario given in the LTC6363 datasheet (page 29) for worst-case power dissipation.

The datasheet predicts 97mW total dissipation. The simulation matches the power dissipation through the load, feedback, and input resistors (12.5mW each resistor) but the term for quiescent current is way off. The datasheet predicts 22mW but LTSpice predicts 66mW.

Is LT's SPICE model wrong or am I using/interpreting the software incorrectly?

My LTSpice circuit is attached.

[davegravy]

Interestingly if I ground both inputs to eliminate the resistor dissipation terms, the predicted dissipation is ~20mW which lines up pretty close with the first term of the datasheet. Not sure then why the +/- 2.5V input case is so far off.