@Matt Venn @Uri Shaked I am traveling today. Can you rerun fill in the next 24 hours? I can pick it up tomorrow morning my time. Also, regarding the missing mcon, you can add that too. Or I can add it. Let me know.

Hi Mitch,

I think what happened is you finished tapeout well in advance. We pushed out an update a few weeks before submission. I didn't catch that you didn't rerun. I have launched a new tapeout of your die. I'll transfer the resulting data over and XOR it to what I currently have. Sorry for any confusion.

Thanks for looking into this!

My pleasure. There's also another DRC violation I am working to fix. It's more complicated to describe. I'll send you a screenshot of it once done to make sure it doesn't cause any circuit problems.

Thanks! tt08_integration comes from https://tinytapeout.com/runs/tt08/tt_um_tiny_pll

No, this is one of those special rules I can't talk about right now. We're working with Skywater to change that.

I see - so nothing we can do ATM to avoid this happening again on TT09, right?

At this point in time, no. Just make sure nwell is >= 0.7um away from URPM.

Is there a way to ensure that using KLayout DRC?

We are not going to roll out any more rule changes for 2411. I hope to have an addition set of rules by end of February. Let me look at it a moment. I can come up with some rules for that.

Excellent - then we can add it to our own "precheck" action, in addition to the existing rules from the MR rule deck

Here's a simplistic rule to check uprm to nwell spacing. Run it by

klayout -b -rd input=<layout name> -rd report=klayout_nw_urpm_check.xml -rd thr=8 \
    -rd top_cell=<cellname> -r <path>/nwell_urpm.drc >& klayout_nw_urpm_check.log

This will flag more than just the geometries causing the error in tt08_integration, but at least you'll see where it can occur. It runs pretty quick. I ran it on TT08 and it flagged eleven violations, all in tt08_integration, four of which were the real ones ID'ed by Skywater. Run this on the new TT. We can get on a Zoom meeting if you see anything that will be difficult to fix and I can tell you if it's a real error or not. And I can refine the rule if it flags too many shapes.