It's a definite possiblity. The second most common source of error, is not building to the calculated cross pattern. (the first is wrong data, usually the rim's ERD).

However the OP is convinced that it's not him, but the calculator that erred. It could be neither, with the spoke vendor having sent the wrong spokes. But the OP will never know unless he checks everything.

In answer to his original question as to which calculator is most consistent, they are all equal at 100% consistency. However the databases vary in reliability and there I'd say MrRabbit's is the one I'd trust most since he goes to great lengths to find and correct errors. However I don't trust any database and measure everything myself. That way if anything goes wrong I know exactly who to blame.