End Device Auto Association pitfall?

I’m using auto-association so my wireless sensors can find a coordinator to upload their data to. It works great. Let’s call this project A.

Once I’ve deployed a bunch of sensors (end devices) and a bunch of coordinators to service them, I’d like to move on to another, unrelated application for XBees. Let’s call this project B.

I’m going to have a bunch of coordinators for these two entirely different projects in the same “room”. (Think humongous aircraft hangar.) The coordinators and end points from project A will use (my home grown) packet payload protocols that are incompatible with project B.

My question is this: When a sensor from project A attempts to auto associate, might it discover 5 PANS from project B, terminating active scan, causing it to associate with a coodinator from project B?

Data transfer will fail (because of the said incompatibility) and the sensor will sleep a bit and try to auto associate again.

My concern is that some sensors from project A might never discover a coordinator from project A. Is this possible?

If this can happen, what is the solution?
Should I just assign each project a unique PAN ID and force all coordinators and end points from that project to use it?

Cheers,
cesium

Hmmm,
That won’t work. On p.18 I read:
After applying these filters to the discovered Coordinators…
It seems that the PAN filter is applied after 5 PANs are discovered. In my case, that’s 5 PANs from project B.
cesium

OK,
My brain hurts.
If the coordinator performs some sort of random delay before replying to the Beacon Request, then by adjusting the Scan Duration, maybe we can create a high but non-unity probability that each coordinator will be found. Lather, rinse, repeat, and we might find our “quiet” coordinator?
A little help?
cesium