Mahmood,

Erik and Andy hit the problems with respect to voltage transients and noise on signal lines. I second Andy's comment about using plenty of ferrite beads.

With regard to radiated vulnerability testing, I recall a field strength of 10 V/m from 30 MHz to 2GHz being used to validate some equipment at a previous employer.

At the same employer a locomotive control system was designed and installed on over 300 locomotives. Locomotives have got to be the worst situation to deal with. US locomotives use a nominal 74 VDC battery system to power control and starters. (European is 110 VDC, I think.) The battery voltage is about 68 VDC without the engine operating. On start, battery voltage dips to 18 VDC on fresh batteries - and about 13 VDC on older batteries. (Railroads usually don't keep fresh batteries in their locomotives.) As the starter motors crank the engine to speed (takes about 10 to 15 seconds on average), voltage to the control circuitry rises to about 35 volts until the engine starts and brings the alternator and battery charging circuit into operation.

Bruce