Hi Pete,
though to anlyze without seeing. If you can send some cross-sections, stack 
up and adhesive materials between flex and rigid could be helpful.
This board parameters are very "large". I can testify I don't see CAF or this 
type of latent phemona in flex/rigids of 10 mils drills and 3 mils lines/spaces.
I believe there's no CAF there: too short time/no voltage/too low temps.
A well manufactured board will survive also at much higher temps. You could 
cycle it at least to 130 C.
Shorts could be formed by misregistration between layers, poor drilling(as 
already stated in the forum), and also poor manufacturing of inner layers. 
Other problems caused by bad manufacturing are shorts around the window of 
interface between the rigid and the flex. The routing of the rigid could "smear" 
some of the copper lines beneath this line.
hope I could help

Roberto Tulman
VP Technologies & CTO
Eltek Ltd 
[log in to unmask] | www.eltekglobal.com
Tel: +972 3 939 5004 | Fax: +972 3 939 3704 |Cell: +972 54 8074204