Replies

"It's the difference between a $50 coffee pot or a $500 coffee pot made to special MIL SPEC requirements. The $50 one puts enough broadband and narrow band noise into a missile guidance circuit that the missile couldn't keep on signal."

That sounds like the emissions from contacts closing/opening. A solid state relay would eliminate that and still not bring the price to $500.

"It will also leak line to line and line to ground potential direct to the steel table it's sitting on."

Only the line's potential could be coupled to the table and that would be capacitively coupled. Any magnetic coupling would be small and just result in tiny eddy currents.

MIL-461. You are not considering cord length and radiated/conductive emmissions. E-fields, H-fields, etc...There is lots that is just too long for a post. Try dc/ac inverters mounted on printed circuit boards. Triacs switching(the contacts you refered to) thermistors, certain chemical compositions of carbon brushes, and so on.

There are no commercial machines or equipment which emmit less than 20db from 4HZ all the way up to 2GHZ without signifigant and costly redesign or modifications.

If you really want to have fun, try to TEMPEST comply your desktop PC. You'll find out very quickly that a $600 computer becomes a $7000 computer in order to get it there.

I don't want to overstate the obvious, but obtaining a good ground in sand is almost impossible. Most of Iraq is sand. I doubt the military attracts too many engineers to support deployed units. Cases like this were why we incorporated mil-specs into defense procurement in the first place! If the pump which electrocuted this soldier was being powered by the same generator that powered the radar units on site, what do you think the reliability of those units must be? Not good for much!