https://open-source-silicon.dev logo
Join Slack
Powered by
Hey <@U069XNVQM4H> , the layout rules PDF defines ...
# ihp-sg13g2
d
Hey @Krzysztof Herman , the layout rules PDF defines "All features are on a drawing grid of 5 nm (0.005um)". The Passiv layer in the bondpad_70x70 cell is from 2.137x2.137 to 67.863x67.863. This is most likely the reason for the DRC error.
k
Hi, I have just enter my office. Let me check and I will paste you here some DRC more verbose output
d
sure ... no rush
k
whwre can I find your most up to date gds ?
d
k
so what is the issue with the bondpad ?
d
You wrote on Friday: "Additionally there were many (184) Passiv_offgrid errors (maybe the bondpad position is offrgid)"
k
have you run DRC on IO cell separately ?
d
no
the DRC rules in the Open PDK don't check for Passiv offgrid right now
wait, the DRC rules Andreas sent last week do! Let me check
k
I have just run 20240515b sg13g2-maximal on my small desing and I do have offgrid errors
d
did you try this bondpad file?
bondpad_70x70.gds
I don't see when running the maximal DRC over sg13g2_IOPadIOVdd
k
so it should be ok
b
@Krzysztof Herman General question: Do off grid errors matter for the upcoming tapeout? Our layouts were drawn mostly on a 1nm grid.
k
@Boris Murmann I suppose that the off-grid check is part of the rejection test
I can confirm you tomorrow since today is holyday so nobody is here
👍 1
I am currently fixing my small design aligning polygons to 5 nm grid and it fixes those DRC's
b
It is unlikely that we can fix all errors (there will be too many violations in our design). The current PDK does not restrict the grid and does not check for it, so that naturally led to people not abiding by the 5nm number that is only there on paper as of now (I think).
k
you are right, we have this default 10, 5 and 1 nm grids available
👍 1
Hi @Boris Murmann I have the answer: "I guess it might not be an issue. In case that you have the GDS, just upload to our rejection test and check the result. I am not 100% sure, but I think the grid is not checked."
b
@Krzysztof Herman Thank you. I have a gds in a pull request from last week. Would you be able to run it? The gds passes the minimal rule deck, but there are many errors with the maximal deck. It would be good to know which ones could be waived, if any. We have been working on further cleanups already, but getting things 100% clean for the maximal deck is challenging.
k
@Boris Murmann great to hear god news from you, let me check the GDS.
👍 1
@Daniel Schultz @Boris Murmann we just found that the sealring PyCell has issues: (1) it should not have the fillers (this is fixed), (2) it should have a EdgeSeal.boundary box 30u away from the outer EdgeSeal.drawing
d
@Krzysztof Herman can we draw the EdgeSeal.boundary box easily?
k
just draw it using 
EdgeSeal.boundary
layer and keep 30 um distance
👍 1
b
Will do.
d
@Krzysztof Herman do I need to move the entire chip or can this box be in the negatives?
k
I guess it can be relative. Todays afternoon I will be able to run rejection test on our platform so then we will know all the exact issues.
4 Views
Open in Slack
PreviousNext
Powered by