i want to know how to implement tanh function using cmos and how to implement exponential and logarithmic amplifier using cmos except the transistors are in weak inversion region
The LPNP should be suported by the fab because you can always make a lateral poly base PNP in plane vanilly CMOS but typical the NWELL diffusion profile determine the DC beta and the maximum current density. Furtheron the extraction is complex and typical done only for fixed unit cells. And as I understand there are some limitation in Tanner for a SpecificCell->Subcircuit extraction procedure.
So you have 2 issues
1. Complex subcircuit model (typ 3 PNP's)
2. Extraction procedure for multi unit cell structure