Tags: architectures, classic, four-transistor

————————

     This is very interesting version of current mirror. This time current is mirrored from M2 to M1! Yes, really! It sounds a little bit weird, but let’s see how it works.

     For sure, Iout arises as a result of Iref current. However, Iout current is mirrored from M2 to M1. If IDS1 is lower than Iref (IDS1 < Iref) voltage VC goes up and as a result VA and VB voltages are adjusted to the situation when IDS1 = Iref, what leads to

Iout = Iref

     If IDS1 is higher than Iref (IDS1 > Iref) voltage VC goes down and again VA and VB voltages are adjusted to the situation when the condition IDS1 = Iref is valid.

     Comparing to the classical mirror, four-transistor mirror requires much higher output voltage Vout, at least VGS2 +VDSsat4, to operate properly.

     It’s biggest advantage is the fact that it adjusts VA, VB and VC voltages in respond to changing Vout voltage to maintain condition:

Iout = Iref

     Hence, the output resistance of the four-transistor mirror is very high.