PMOS diode-biased cascoded current mirror

From ICclopedia

Schematic Diagram

Circuit Netlist

* dev <nets>        <values>                                                                         
* -----------------------------------                                                                                         
I1    n5 0        50uA                                                                                                        
I2    n3 0        50uA                                                                                                        
V1    vp 0        1.3V                                                                               
V2    vp n2       0.45334V                                                                                                    
M1    n1 n5 vp vp pmos W=220u L=3u                                                                                            
M2    n4 n5 vp vp pmos W=220u L=3u                                                                                            
M3    n2 n3 n4 vp pmos W=110u L=0.35u                                                                                         
M4    n5 n3 n1 vp pmos W=110u L=0.35u                                                                                         
M5    n3 n3 vp vp pmos W=41u L=3u

SPICE Simulations

Operating Point Analysis

Operating point DC measurement results (re-formatted for display)

n1 = 1.045
n2 = 0.847
n3 = 0.646
n4 = 1.045
n5 = 0.847
v1#branch = I1 + I2 = (-50) + (-50)  
v1#branch = -100uA
v2#branch = -50uA
vp = 1.3
(v2#branch/v1#branch) = 1.0

And our relevant transistors' device parameters at the DC OP (re-formatted for display):

device  m5        m4        m3        m2         m1                                                 
model   pmos      pmos      pmos      pmos       pmos                                                  
gm      275.6 uS  959.3 uS  959.2 uS  620.8 uS   620.8 uS                                              
rds     1.095 MR  92.41 kR  92.41 kR  201.36 kR  201.35 kR                                             
id      50 uA     50 uA     50 uA     50 uA      50 uA                                              
vgs     0.654 V   0.399 V   0.399 V   0.453 V    0.453 V                                            
vds     0.654 V   0.198 V   0.198 V   0.255 V    0.255 V                                               
vth     0.337 V   0.390 V   0.390 V   0.337 V    0.337 V                                               
vdsat   0.302 V   0.080 V   0.080 V   0.141 V    0.141 V

DC Analysis (Sweep)

For measuring the variation of the mirrored output current under different applied loads. We apply a DC sweep to V2 (our load voltage) from 1.3V to 0V in 0.05V increments.

We are plotting the output current magnitude vs drain voltage.

Monte Carlo Analysis

In our Montecarlo analysis, we are measuring the effect of transistors’ mismatch on the performance of our current mirror i.e. how small variations in individual transistor parameters when added together can affect the performance of the overall circuit (please refer to the NMOS counterpart article for additional setup details for this simulation).

Results

For the cascoded current mirror, we can make the following observations

  • The absolute minimum voltage needed for the mirror output branch to be in the sat region is the vdsat of both mirror and cascode devices: vmin_abs = 0.110 + 0.198 >> 0.308
  • However as shown with the current sizing, the mirror output can go up to maximum 0.8V (at least ~0.5 vload voltage relative to vp) to yield its maximum Rout. (seen from the plot above where the slope is linear).
  • Error Measurement: We see a variation of 49.924uA to 50.002uA over a linear operating vload range of 0.8 to 0V. This is equivalent to an error of 78nA or 0.156% relative to our reference.

Figures of Merit

Output Resistance Rout: 10.256MΩ (measured from 0.8 to 0V linear range)

Compliance Voltage Vmin: 0.8V (or ~0.5V from VP)

Please keep in mind the figures above are likely to be optimistic given we have not considered mirror performance over PVT (where we will expect a large variation from the diode based reference).

References

Toolchain