One more time a bug due to Cadence:
impossible to post a message without "Normal 0 false etc..."
In reply to threepwood:
Assuming I've understood things correctly and you're using either the final IC6.1.2 ISR or latest IC6.1.3 ISR...
Option A: A simple option (w.r.t. setup) would be to de-embed the dcsweep. Set up the temperature and additional parameter sweeps (like Process and/or Voltage) at the ADE XL level (design variable sweep and/or corners) so you get the expanded set of sweep points to check specs against.
- Advantages to this approach include ability to perform more detailed analysis at each dc point (calculator expressions, device checks, etc.), use the output of each dc point as an input to other tests (using calcVal() and VAR() interactions), and of course you can take advantage of multi-job and/or distribution queues to run things in parallel (more useful when working with larger extracted rep's) assuming you have enough sim licenses. ;-)
Option B: A more complex option if you want/need to use Spectre's built-in dcsweep analysis would be to help the range checking along with a couple of calculator expressions. One would return the minimum dcsweep result, the other would return the maximum dcsweep result, then set the min and max spec check accordingly.
- It keeps the Spectre run compact, but you would loose visibility into which individual temperature(s) are failing spec.
Option C: Ping your Cadence AE for assistance when s/he returns from their manditory shutdown vacation. Also ask/tell them to file an SR on your behalf reporting the (broken) dcsweep integration issue in ADE XL.
Hope this helps!
In reply to aplumb:
Thanks a lot Andrew for this precise answer.