Super simple and about as low tech as you can get...
Measure the circumference of the balloon. A possible way to do this is:
Find the burst radius of the balloon. Subtract a safety margin to get the launch circumference. Attach a number of eyes around the circumference of the balloon then tie fishing line to the first eye. Thread the fishing line through the rest of the eyes, once you get back to the first eye take the line down to the launch platform. Add extra line for the difference between the launch circumference and the current circumference. Put a big knot in the line at the launch point. Run the line through a switch arm with another eye on it. As the balloon gets bigger the line gets pulled through the eye on the switch, when the knot snags in the eye the pull on the the switch arm will activate the switch. Boom! (or is that whoosh?).
An even simpler version would be to attach the line to the top of the balloon and run it straight down to the launch platform via an eye on the launch pad attachment rod/rope(s). No eyes on the balloon at all.
The excess line could be put on a reel with a light spring to keep tension on the system to prevent tangling - fishing line needs to be controlled.
This should work pretty well, the force of the expanding balloon will be quite high so you wont need to make the line system too sensitive or friction free. It also doesn't require any calibration like the condom/second balloon version would require. Also easily tested in a commercial freezer by over inflating the balloon.
Issues I see:
o Attaching the eyes to the balloon in a way that allows for the expansion.
o Making sure the line will slide freely when it get super cold.
o Making sure the storage & tensioning reel still turns freely in the cold.
o Tangling, especially in any wind, also as the line gets stiffer due to the cold.
o Keeping the line away from the launch path
o The launch platform must not rotate independently of the balloon or the fishing line may tangle with the main line.
o I have assumed that the burst circumference is known and doesn't have much deviation.