|
You haven't missed too much. The cosmic ray bit-flip guys have a point, but that is very hard to shield against and probably won't kill your servers. See this paper for a deeper dive. https://digitalcommons.usu.edu/cgi/viewcontent.cgi?article=1... In regards to the Faraday cage (grounded mesh), assuming you have a very good path to ground and a sufficiently robust mesh material, you are shielding against any emissions with a wavelength larger than λ. You will have better results versus wavelengths that are much larger however. If you want a deeper dive on that, you can actually read Faraday, or Wikipedia or just ask your search engine about a basic dipole antennae and work up from there. Many computer components are somewhat shielded/tolerant of standard levels of RF interference in the environment so you might not even need to bother (depending on application). Normally (in my experience) you see problems arise with sensors, digital control lines and some human interface devices in very noisy environments (industrial facilities with a lot of high power variable speed electric motors). Regarding aperture size λ, figure out what source of RF you're most concerned about, determine the range of wavelengths emitted, concentrate on the most powerful bands and then make your aperture smaller than that. Examples are 'the sun', cosmic rays, north korean nukes, the welding robots on the other side of
the wall etc. Finally, if you're going to shield, don't half ass it or you waste money. Gaps around doors, or anywhere you poke cable through a wall can cause problems if not handled correctly. |