miso a écrit :
No the DC effect will be much less. In Tim's circuit the transistors have way different dissipations and the output transistor at the same time defines the output current and sees the power change which is bad. In a Wilson or enhanced Wilson mirrors, not only the two transistors that define the current both have low VCE and low dissipation, but the one that sees the varying power totally shields the current defining transistors from the power variation, as does the cascode one.
As you note some base current error still remains but that too can easily be dealt with.