is this correct to use dummy poly for matching?

Status
Not open for further replies.
Z

zhangljz

Member level 5
Joined
Oct 19, 2013
Messages
81
Helped
1
Reputation
2
Reaction score
1
Trophy points
8
Activity points
648
Hello,

I want to use dummy poly to improve layout matching. I use "poly_dg" layer to draw dummy poly. Is this a correct way to do it?

when I do LVS, there are warning "Missing connections stamping poly nodev by layer poly_cont".

While if I use "poly_dmy" layer, the DRC rule of "poly_dmy" is different . In PDK, it says "dummy poly rule is drawn dummy poly guideline purpose only, this rule is ignorable at designer's own risk".

Which should I follow?

Thanks
 

Dummy poly addresses gate litho loading but it does not
do anything for other effects like S/D (active) litho, STI
strain effects and so on. Best matching is always like-for-
like and that goes for dummies also (a dummy FET covers
all FET litho aspects, while selected FET-chunks do part
of the job).

Do you know that poly litho is the way-dominant factor?

Seems like one of the errors is whining about poly not
connected (floating poly)?
 

Hello erikl,

Thank you. I searched some threads and they said poly_dg and poly_dmy are the same material. So I can use either of them to draw dummy poly. Just poly_dg is mainly for transistor, but if I don't give it an active region, it is also a dummy poly.

The DRC difference between them I think may because poly_dmy is mainly used for automatic dummy filling. So a larger spacing of poly_dmy can reduce parasitic, and has less impact to the circuit
 

I think, for DRC the important difference is that the poly_dmy layer is a recognition layer for poly dummies, and so will be subject to relaxed connection rules. You might find out this by checking the DRC rules' file - if it isn't encrypted, or simply by trying both possibilities.
 

Status
Not open for further replies.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…