Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
I believe this will be very foundry, technology and cell-library-
specific. You should at least say what you're looking at in hope
somebody here has walked down that road already.
Or, you might read into the library usage docs where such cells
probably would have their own category and a couple of
paragraphs about what they're for and how to invoke correct
ones for the application.
ENDCAP cells are silly cells that extend the layout a little bit to make sure that the wells are long enough. There are some variants of this, depending on which side you are trying to extend. an ENDCAP cell for left/right side might be different from and ENDCAP for top/bottom sides. Corners sometimes can be special cases too. This changes with technology. In old techs, there were no ENDCAP cells at all. Then we started to see the need for side caps on the left/right of a standard cell row, perhaps around the 90nm era. Then we started to see top and bottom as well in the 32nm era more or less. In newer FinFET nodes you can have top/bottom/left/right/corners.
ENDCAP cells are silly cells that extend the layout a little bit to make sure that the wells are long enough. There are some variants of this, depending on which side you are trying to extend. an ENDCAP cell for left/right side might be different from and ENDCAP for top/bottom sides. Corners sometimes can be special cases too. This changes with technology. In old techs, there were no ENDCAP cells at all. Then we started to see the need for side caps on the left/right of a standard cell row, perhaps around the 90nm era. Then we started to see top and bottom as well in the 32nm era more or less. In newer FinFET nodes you can have top/bottom/left/right/corners.
Dumb question, why do we need to extend the well at all? If each std cell has the well inside it wont we get continuous well by abutment? I see these cells too and would like to finally know what endcap cells are doing. Seeing them on core boundary, surrounding macros..
there are two reasons to extend the well:
- DRC compliance
- timing environment
For the first one, it is possible to draw an individual cell that never passes DRC unless it abuts with another cell. just because the well has to be X nm away from edge or the minimum P/N zone has to be Y nm wide.
For the second one, standard cells are characterized assuming they are in the middle of a row of cells and do have neighbors. Without neighbors, their timing changes only so slightly. If you want to learn more about this, try reading about the well proximity effect.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
