Question on R-MosFET-C filter

[jimway]

Hi everyone,
I am supposed to design an R-MosFET-C LPF.
What has puzzled me is that the corner frequency of R-MosFET-C filter is strongly affacted by the deviation of process.
Here is the schematic of the filter.

Here is the Bode plot of its corner analysis.

My question #1 is how to improve this circuit.
#2 what cause the zero-pole pair (enclosed by the circle)? And how to eliminate it?[/img]

Added after 4 hours 9 minutes:

Is it possible to compensate the deviation of the process by the voltage control block?

 

[szekit]

yes, mos-resistor is susceptible to process variation due to Vth.

u need to derive the transfer function to see the contribution of zero.