Just checked, calc fails at 11^0.41703 (curiously not at 41703, but 41704, maybe a floating point thing). Maybe it changes base to e and uses the Taylor series for e^x, but fails if x is too precise for what it expects?