<@U017X0NM2E7>: It would be an error, essentially, to draw a poly or diffusion resistor that way, s...
t
@Mitch Bailey: It would be an error, essentially, to draw a poly or diffusion resistor that way, so I'm not too worried about trying to make the L and W calculation work out correctly for those cases.
m
I agree that it is incorrect to draw resistors like this. Is there anything that can be done to alert the user that the resistor shape is invalid? Can magic drc flag non-rectilinear shapes?
t
Usually the extraction flags the non-rectilinear shapes. But I recently had to fix an issue with the calculation of a device with three sides in the same layer. I expect the calculation of resistors has the same error. The L and W calculations are reversed between resistors and FETs, so that code is duplicated. If that is the case, then fixing it should be simple and then it would both flag a warning as well as extract a reasonably accurate L and W---probably good enough for simulation. Then it would be the job of the designer to match the device parameters in schematic (at which time they might realize that wasn't what they intended at all).