its well known that backup batteries barely work when they arent charged, you basically have to replace them every few turnoments because they drain so fast (or atleast the batteries we have)
what if instead, the vexnet key stores a charge and back powers the cortex when it brown-outs, keeping the cortex and the key online for a short period of time giving it enough time to recover
Rechargeable 9 volt batteries might solve your problem. 
IIRC these dont work unless you charge them before every match, let me search for the thread
Main robot batteries don’t work either if you don’t charge them. 
You just need to remember to charge/replace them, and if you don’t the robot will still work, it might just disconnect for a short time.
While I see your idea and it seems like a good one, I don’t know if it would be viable and hold enough power to keep a connection. And VEX just released the 2.0 keys, so I don’t think they’d want to come out with new ones for a while. Maybe in a new microcontroller it could automatically charge an internal or external backup battery.
i dont beleive a battery is a good solution, because surges are sudden and batteries discharge slower
i need to do some tests to see how much current the key and microcontrollers draw to see how much capacity it would take to keep it running
i know the CPUs use a verry small amount of power, if a low-power mode was added to the keys that would help a lot
Capacitors would drain too fast for this use. If you unplug an computer/electric device with LEDS, You will see that they only stay on for a second at most. 9V batteries is a good solution for this problem, because it can power a whole robot for a good amount of time, from my experience.
Ive never had motors run off of a backup battery, you sure it doesnt just power the processor and vexnet key?
Do the motors on your robot still run as well as when you are using the normal 7.2V battery?
And since you mentioned “for a good amount of time”, can I assume that it can power the robot for at least 2 mins (a single match)?
When you turn off the power switch on the robot but it is still kept on by the backup battery, you can apply power to the motors and they will twitch or move slowly, but not like on the normal battery.
That seems like the PWM signal leaking through the motor controller and not unregulated power from the battery
I believe the backup battery is incase the main battery gets disconnected and reconnected quickly - enough to keep the robot’s signal still connected - until the main battery reconnects, and you won’t have to wait a few seconds for the cortex to reconnect. I might be wrong.