# MOS current mirror region of operation

The mosfet M1 is clearly in saturation as gate and drain terminals are shorted. Why we take M2 also in saturation? I know Vgs1 = Vgs2 and they are identical MOS but i cant justify why M2 should be in saturation, as drain voltage of M2 can be any value..

• Can you share the exact text you're asking about, that says we can assume M2 is in saturation? May 27, 2020 at 14:50
• The basic current mirror can also be implemented using MOSFET transistors, as shown in Figure 2. Transistor M1 is operating in the saturation or active mode, and so is M2. en.wikipedia.org/wiki/Current_mirror May 27, 2020 at 14:58

According to Sedra/Smith text (because textbooks do a better job at explaining things than I do), Microelectronic Circuits, on the subject of basic MOSFET current mirroring:

Now consider transistor $$\Q_2\$$ (or $$\M_2\$$ from your picture): It has the same $$\V_{GS}\$$ as $$\Q_1\$$... (from Figure 8.1) For proper operation, the output terminal, that is, the drain of $$\Q_2\$$, must be connected to a circuit that ensures that $$\Q_2\$$ operates in saturation... To ensure that $$\Q_2\$$ is saturated, the circuit to which the drain of $$\Q_2\$$ is to be connected must establish a drain voltage $$\V_o\$$ that satisfies the relationship $$\V_o \geq V_{GS}-V_{tn}\$$.

Microelectronic Circuits ed. 7 by Sedra and Smith

So in order to have a proper current mirror, we have to assume that the M2 is in saturation. Drain voltage for M2 can be arbitrary (to an extent) with that condition of $$\V_o \geq V_{GS}-V_{tn}\$$.

The drain voltage of M2 "can" be of any value but it's only useful as a controlled current sink when Vgs is high enough that the pair functions as a good current mirror.

The compliance voltage for saturation (where the current is mostly constant) is:

Vout > Vgs - Vt so if you are operating M1 at a current that creates a Vgs only slightly above Vt (in other words, a very low bias current) you can have Vout almost zero and still get constant current.

Here you can see a current mirror compared at 100uA/1mA/10mA and the behavior as Vds increases from 0V to a few hundred mV. At higher currents, the compliance range of voltage is less (you need more Vds to maintain a constant current). Vgs is about 1.41V to 1.73V in this simulation, depending on current.

If M2's drain voltage drops too low then of course it won't be in the saturation region but, importantly, neither will it be a good current mirror. M2 should have a drain voltage no lower than M1's to ensure it remains in the saturation region but some devices may be able to go lower without entering the triode region too deeply.

The red markers on the curves below show where the diode connected MOSFET M1 sits: -

So there is some scope in some places to drop the drain voltage of M2 below that of the diode connected MOSFET and remain reasonably in the saturation region. It all depends on the particular MOSFET characteristic.

The 2N7000 for instance has a fair bit of leeway to drop below M1's drain voltage: -

No real MOSFETs were harmed in the creation of the above pictures.