<@U03B029DKU5> Points to Observe - The COMP fill...
# ieee-sscs-dc-23
g
@aquiles viza Points to Observe • The COMP fill is too close to active COMP areas. • Maintain a distance of 3.2 units between dummy_poly2 and COMP. • The METAL fill is too close to active metal areas. • Ensure there are no dummy metal pattern fills in the following areas: MIM CAP, Poly fuse, Metal fuse, so avoid areas above MIM caps. For Poly fill: Space between two(or more) not connected _dummy_poly2_ = 2.4 Space of _dummy_poly2_ to COMP = 3.2 Space of _dummy_poly2_ to *poly2*(ACTIVE) = 5 Space of _dummy_poly2_ to Nwell boundary = 1 Space of _dummy_poly2_ to DNWELL boundary = 2 Space of _dummy_poly2_ to LVPWELL boundary = 1 Space of _dummy_poly2_ to Dualgate boundary = 1 Space from _dummy_poly2_ to Resistor marking layer (RES_MK) = 19.7 Space of _dummy_poly2_ to Pad = 6.7 Space of _dummy_poly2_ fill to active circuit metal1 = 2 Space of _dummy_poly2_ fill to active circuit metal2 = 2 My idea is to create a function to avoid all this above mentioned layers, and make the distance +10um (except for RES-MK). I did not read all the rules yet, but we need more people on this if we are going to create the scripts from scratch, there is a lot of rules. CC @Atif Khan @Juan Sebastian Moya The relationship between Dummy_poly2 and Dummy_COMP needs to be clarified, as they should interact in a specific manner. I am currently investigating this further to understand the details.
a
Yesterday I've read the rules, they gave a lot of indications on different areas: • Placement with shifting between dummy cells on the same layer. • Different separations according to the layer (On the message). • Offsets between consecutive metal dummy layers. • More and more criteria. The script is unable to execute all of this in the state it is now. We have to understand: • How the TilingProcessor understand "Expressions" to make more complex petitions, and • If it's smart enough to apply shifting rules on the same dummy layer. For instance, klayout can place arrays with vectors which basis is on both X and Y, but I don't know if TilingProcessor can. • It seems that Dummy Poly should be over Dummy Comp (???), How can we indicate this is a mistery, maybe with another cell composed by poly/comp dummy. This requires to manually specify which areas should be filled with one or another