Hello all! I am trying to do NMOS characterization. When I change W/L I get a change in Kn'. Kn' should be a constant but I am getting different values for different W/L.

Unfortunately this is a theoretical model. In practice you will have short-width and short-length effects and so both Vt and Kn' will change with W and L.

Okay but then what formula should I follow?

To get accurate results for varying L/W you'll need to simulate it in ngspice.

I have tried simulating the design in xschem with W/L - 0.42/0.15, 7/8, 0.42/1.