Equivalent width of a transistor (Cadence extraction)

Hi, all:

I have a question on the effective width of a transistor layout.
the layout is at http://wangxf.nease.net/temp/transistor.bmp

(The drawing is not to scale, the shape is actually a square.)

I drawed the layout in cadence and did the extraction. Cadence gave an
effective transistor. The effective width is around 47um, while the effective
length is around 0.5um.

I know the effective width is due to the corner effect.
But I have some confusion about the effective length. Why it is half of the
physical length. I tried different length, and it seemed the effective length
is always half of the physical length. Can anyone tell me why? Is Cadence
extraction accurate?

Thanks.



Best Regards,

Xiaofeng

 

Xiaofeng,

The answer is entirely dependent on the technology and the rules being used
to do the extraction - any computation of effective widths and lengths would be
done there, presumably. It's not something that the extraction tool (whichever
you're using) would do on a general basis.

Regards,

Andrew.