protection of electronic Box against EMI

Dear All
I am installing an electronic box, including analogue and digital boards, computer and modems, on a locomotive.
the box collects signals from several sensors installed around the engine. (from 3 meters to 15 meters of distance)

the box is from steel.

I would appreciate your advises on protecting the box from interfering signals from high power generator, relays, switches and etc. (the box is installed very close to the high power cables with above 600 volts and 1200 amp signals)

Is it necessary to use extra protection for each board? (over voltage and/or over current protection)
Do I need to change the box with an aluminum one?
considering the metallic structure of the locomotive, what do u suggest for earthing of the system?

your comments on different aspects of the system is highly appreciated.

Many thanks

Grounding the box is a given but interface of all the wires requires additional shielding, filtering and protection from overvoltage, short circuits, live disconnect/connect, stray magnetic field pulses and stray electric field pulses. Magnetic is current/inductive coupling and Electric is voltage/ capacitive coupling.

Standard tools are CM chokes, feedthru capacitors, ferrite beads for differential filtering and noise suppression caps to limit noise bandwidth. Wires should be orthogonal to high current wires and low impedance over the range of signals being transferred. ESD protection and isolation from high Voltage breakdown must be analyzed from risk analysis of the cable paths.

Differential inputs and outputs may be necessary and also insufficient without CM chokes.. Polyfuse may be useful in case of faults

The design needs lots of consideration DFT ( design for testability) with a suitable user interface and rugged design so that it can still work after being dropped and run while vibration exists with high temperatures. (HALT/HASS design principles)

I would not recommend perform that replacement, due aluminum have not so good shield performance if compared with other ferromagnetic materials such as cooper and iron. Basically, you must employ materials able to reflect electromagnetic irradiated fields.

However, the greater ability to shield noise stemmed from external sources, also increase the ability to confine the noise generated inside the enclosure, and this means that depending on the application, you will also need to employ at internal surface some absorbent material to prevent inner noise generated not reach critical levels.

Take a look on this guide which covers these concepts: Engineering Aspects of Electromagnetic Shielding




+++

I agree Steel box is better for magnetic shielding but biggest problems will be the cables not the box. Steel must be anti-rust coated and welded ground stud if painted with lugs between nuts..

Copper ground planes on board usually shield on one side well enough, but cable dressing on interface to the board is ignored and conducted noise on cables can then radiate inside the box. Laptops use 1mm steel wire gauze/mesh or copper foil & adhesive to wrap around sensitive internal cables or create a ground plane.

Avoiding common analog, power and digital current ground paths is also important from ground shift using a star or single point ground or a heavy distributed ground.

Traction motors radiate a significant amount of RF noise in the 50 ~150MHz band from what I recall when we measured TV ingress noise during mid 80's from nearby railroad tracks 100m away

Coaxial CATV ground connections and shield transfer imepdance not being perfect contributed to the problems.

Many thanks for your valuable comments and advises

on the locomotive, all the electronic equipment (including high power and low power) share a common ground which is the structure of the loco.

in order to provide a level of protection, is a complete isolation of my electronics boards from the box and locomotive body, useful?

and for limiting the inner radiations and noises and also the effect of external magnetic and electric fields, someone has advised me to use sort of aluminium foils or coating, inside the box.
do you think it is wise to do this?

Best regards

Although electric isolation is more commonly used for protective purposes, in cases of noises reaching high levels, it is also quite effective. However, DC-DC converters used for generate isolated bus are another noise source.

In short words, there is no general rule but each case requires a detailed investigation of impact owing to the adoption of certain solution.



+++

LIke I said, your cables will be much bigger problems for conducting and radiating noise (Ingress and egress) A test setup might be a noisy Boost regulator (PFM) SMPS or a lousy BUck regulator with fast transitions <10ns with a dummy load and wrap the long wires around your cables for ingress testing.

A Lingren Faraday Cage is designed to attenuation >120dB using 2 layers of copper gauze like mosquito netting with Berrylium Copper fingers on all contacts and door gaps. All gaps and slots will act like 1/4 wave antenna. THe surface does not need any more unless you are getting in high GHz radar.