We have drc rules which apply to a slot layer and to holes in metal. Checking the slot layer is not really a problem, but identifying which holes should be considered slots is a real problem. Anyone have a suggestion?
Do you have a fairly strict definition of what proper slots should look like? Such as a certain width and length?
I wonder if "good slots" could be identified using those criteria and then other metal holes that "almost" meet those criteria could be identified as bad slots.
Slot checking has been an interesting issue for so many years but I'm still not personally aware of the best techniques to check them. It would be nice to get some input from different people with different perspectives.
The tools are advancing so fast right now that I have to wonder if metal slot checking could become easy with the help of some new capabilities. Discussing the current limitations (no pun intended) might be a great place to start.
A valid slot is one created with a metal slot layer or a hole in metal that is > 2 microns wide, <= 10 microns wide, >20 microns long < 250 microns long.
They must be 10 microns from the edge of metal, must run in direction of current, must be 10 microns apart.
The problem is with determining how to identfy the which holes to check. No matter which width we use for the hole, if the hole is larger than that and it violates a rule, it does't get checked. Then the issue is which 10 or less micron holes are not actually slots (false errors).
You are right this is a very sticky issue.
Maybe we could chip away at this.
Would it be helpful to have DRC output holes in metal that are too wide to be slots?
If it's preferred they aren't output that way, then we should still be able to make sure that any non-slot holes (that cause other problems) aren't missed.
I guess I am not certain what you mean. Assume we have holes in metal1 and we "add them back" into metal1 (and check to see if that metal can be considered wide, requiring a slot). Then the slots may or may not be considered slots, based upon certain criteria. The criteria picked may ignore some holes that should be slots and include others that shouldn't.
So do we agree to allow all holes or some holes? The criteria is one thing, then the implementation is another.
I didn't realize you were adding the holes back in to the metal and then checking it for width.
Would it be feasable to simply check the holes to see if they qualify as bona fide slots, by choosing only holes that meet the specified criteria for hole dimension (regardless of presence in fat metal or narrow metal) ?
Then, if metal had bona fide slots in it, would the metal truly be non-wide metal (if the true width is actually reduced by the slot opening).
Maybe the derived metal layer could be the result of adding in just the holes that were too narrow to qualify as slots. I realize those three suggestions may not help at all but I'll leave them there just in case they inspire ideas from someone else.
I guess the whole process needs to account for wide metal spacing at the same time as metal slotting. This does seem to be a complex problem.
Retrieving data ...