[SOLVED] Output resistance of Cascode Current Mirror

[farhan89]

Hello,

I have a small confusion regarding the Output resisitance of Cacode current mirror and Wide wing cascode current mirror. As per my knowledge, both of these have same output resistance. But when I calculated thru AC simulation its giving me different values for resistance.
Kindly see the attached images. Both configuration have identical transistors. M4=M1 and M2=M3, M2 and M3 has twice width of M1 and M4 and M5 gate length is 4 times of other transistors. IREF is 50 uA and V_OUT is 1.2 V, Ibias is 200uA, in DC simulation I can measure the I_OUT (100uA) , for AC simulation I am inserting 1 V AC at output and measuring the current at output (R=V/A). In Cascode it is around 569K ohm and in Wide swing it is 600 Kohm.

Could anyone explain me why is this so ? If I decrease the Ibias (in wide swing) then resistance is decreasing.
Any hints

 

[D.A.(Tony)Stewart]

Does load capacitance and frequency explain it?

 

[farhan89]

Does load capacitance and frequency explain it?

In both cases I am sweeping from 1 Hz to 100 Ghz, Frequency will effect the parasitic capacitance , but how the resistance value changes ?

 

[dick_freebird]

I view that difference in resistance as trivial, and probably
due to the operating point differences between wide swing
and classical cascode operating points. The Vds is likely
partitioned differently and my gut feeling is that the classic
version has the guard (upper) running with lower Vds, so
closer to linear and less improvement to Rout. You might
find some clues, inspecting the OP results for the output
legs of both versions.

 

[AMS012]

Output resistance as all know is Gmcasc/(Gdscasc*Gdsmain). In the two cases that you are comparing, the gate voltages of both devices M2 and M3 could be different in both cases, which causes the change in resistances. Dump the DC analysis to a file and look at the Gm and Gds values.

In AC analysis, simulators take the small signal model and do the calculations, in which case the swing does not matter.

 

[farhan89]

Thanks everyone, the problem was the operating points, in regular cascodes, Trasistors were operating at different operating points than wide swing configuration. Now the output resistance is almost same.