It seems perfectly clear that the root problem with AO-92 is that the batteries are no longer holding a charge as well as they did when new. I am not a battery expert, and unfortunately, we had no battery expert volunteers when the Foxes were being built (although we do now). Fox was designed for AMSAT engineering and ops to learn about cubesats. This is one of the things we learned. All the Foxes are complete--there is nothing we can do about that except work for better results on Golf.
Another issue is that AO-92 sometimes stays in safe mode after the voltage rises again as it exits eclipse and requires a command station to start the repeater again. That is not the design. I'll take responsibility for that behavior since I wrote that software. And believe me I *am* analysing it to ensure future birds that I work on don't have that bug (which is subtle to the extent that I have not yet found it, or even been able to reproduce it on the ground).
I'd add that these problems are most likely based on volunteer engineers with limited time butting up against fixed launch schedules set by large for-profit corporations with big payrolls. It's hard for me to imagine blaming the board of a volunteer, non-profit corporation for this situation.
73,
Burns Fisher, WB1FJ AMSAT Flight Software Volunteer