I saw the following calculation in the training video on the official website, but I can't understand why in the calculation of the output error caused by the bias current, the errors seen from the non-inverting terminal and the inverting terminal need to be added instead of subtracted. Woolen cloth?
For example, when calculating Vib, why not Vib=ib(Rin-Req) but Vib=ib(Rin+Req)?