Sensor circuit is mounted in a cavity inside a thick metal cylinder with a single 3-wire cable running out between it and the central controller of the telescope. We added a pair of Zener diodes to prevent overvoltages from entering the circuit while massive metallic housing makes it highly resistant not only to mechanical but also to electrical disturbances from the environment. We are actually pretty sure that the tilt sensor we built is a good candidate for surviving the EMP warfare. The only problem that might jam this device that comes to mind is the future development of gravitational waves astronomy and possible breakthroughs in physics that might turn it into an experimental science - a playful postgraduate mingling with miniature black holes in the vicinity of the telescope has a fair chance to trip this sensor.

If anyone wonders whether all this effort was worth it... well, yes. In more than one occasion since its seeing the first light, our humble sensor did stop the telescope from experiencing a nightmare. This is a fine confirmation of the rule that states that no automatic safety mechanism is effective if there is a switch to turn it off - such as IS the case with software limits in telescope motion control applications.

Carl Sagan once lamented about which side of the telescope intelligence resides at... But having all this in mind... It should be clear why there is no item such as "thisRover.power_off" in Spirit and Opportunity command lists :)