Well Paul,

What I mean is as following.

when i only simulate the SC CMFB, the voltage produced is as expected. however, when i combined the full diff opamp and the cmfb, the cmfb cannot work properly. in the end, the opamp cannot work.

i use spectre RF (the pss and pac analysis).

when i use trans analysis from Mentor graphic, the correct dc point of the opamp can be obtained.

is there something wrong with the Spectre RF? or any design rule for SC CMFB?
thx.

I was able to run a simulation with Spectre from the netlist you sent and did not find anything that would suggest a problem. However, as you did not send a simulatable netlist (there were no analyses, options, etc.), I cannot really know what simulations you ran, or which outputs you observed. Nor have you really described what problems occur (you only say that the simulations cannot work).

If you would like more help, I recommend that you construct a netlist that demonstrates the problem (be sure to test it to assure it both runs and exhibits the problem) and send that to me. Be sure to include a description of the circuit (inputs and outputs, function, etc) and the problem.

-Ken