In the office we joke about the kind of Stackoverflow answers that don’t actually answer the question but instead just ask “why would you do that?”
 
I don’t want to be “one of those”, and I do understand the convenience of being able to double up on IPIDs for testing, but I think the real answer here is - don’t do that.  Even though it may work (not always though!), it’s always come with some side effects, and I think you’re simply seeing that - how could the processor keep a single panel definition’s activity feedback in sync with two or more connections?
 
Perhaps a different approach would be to create a testing panel with a complete copy of the physical panel’s logic to keep them isolated or - if you really want to track the physical panel - create a separate IPID/panel and duplicate all of the signals except for the problematic ones (e.g. activity).  Panel joins are not “exclusive” - they can be jammed/driven by two sources.
 
Might that resolve the sync issue?
 
But - personally - we never double up on IPIDs - we just have a “test” panel definition.
 
All the best,
Oliver