The UPS does not help. Note that similar coupling of the electromagnetic pulse from nearby lightning blew out my home computer via the Ethernet cable, even though my computer was on a UPS.
A: Lightning protection is needed whenever there is no guarantee that both ends of the wire will have about the same potential if lightning strikes. If you can establish a ground plane under everything with only one connection to the earth, you might be OK. This usually isn't possible if electric power or other wires leave the area of the ground plane.
Lightning currents can be huge, so even very low resistances can develop damaging voltages from one end to the other. The magnetic field from a lightning bolt changes in sub-microseconds, as does the electric field, so you have more than enough electromagnetic energy to get through cable screens. Conduit is better if it doesn't carry lightning current.
People who are serious about lightning protection put up tall towers with excellent earth grounds that are checked regularly. Lightning will prefer to strike the tower, which provides a cone of protection at about a 40° to 60° angle from the top, enough to cover buildings.
The trick is to isolate the tower lights from any other electrical circuit. Once the lightning is in the ground, it will dissipate over an area. That is why four-legged animals will die, but humans with their feet together may live when lightning strikes a nearby tree. Being inside a building is not enough, especially if the building has exterior lights.
This is a problem in risk management. Balance the chance of a lightning strike and its cost against the cost of lightning protection and the probability that it will be effective. Fiber-optic and wireless connections remove a lot of the risk.
The PolyPhaser division of the Protection Technology Group (www.transtector.com) has specialized in lightning protection for many years, especially for antennas. Its web site has good information about lightning and grounding.
Transmitter companies such as Rosemount offer external protection devices. Talk to PolyPhaser or MTL about what to do at the marshalling panel end.
I don't know of any device that will survive a direct hit, but a grounded metal building can protect even against direct hits.
A: Internal transient protection is relatively simple to implement in a 2-wire transmitter because there are only the two wires and the case or shield connection to worry about. In my opinion, all transmitters should have internal transient protection. I include it in all transmitters that I design. In my opinion, it should not just be an option. It should be included in the transmitters of all three cases that you list. I use transient protection zener diodes with at least a 600-W peak rating, such as the Littlefuse model SMBJ33CA.
Transient protection is not protection against direct or nearby lightning strikes. Nearby lightning strikes can cause a peak current flow of thousands of amps, and protection against that is not practical to include within the transmitter (even a very beefy circuit board copper trace can be burned-off). A direct lightning strike may cause the entire transmitter to vaporize or leave a melted mass of remains, and I don't know of any device that will protect against such an event. Some optional equipment is available, as you mention, for additional protection against damage from nearby lightning strikes.
In cases where cable runs will go outside a building and not underground, then you may want to consider the additional external protection.
David s. Nyce
A: Years ago, I normally used a metal oxide varistor in combination with a zener diode (in parallel, but with a small impedance in between). That was because the MOV could take a lot of peak power, which a small, inexpensive zener could not, but the MOV by itself was too slow to protect some MOS ICs. One problem with MOVs is that they can become damaged with repeated activation within their specified rating.
Now I use a higher power zener and no MOVs, because the higher power zeners are now available with a relatively small size (smaller than an equivalent MOV) and at low cost.
When building a lightning protection circuit (as opposed to just a transient protector), the first element in the circuit can be a high-voltage fuse, then a small impedance, followed by a spark gap of around 60 V (The spark gap is for the really high currents of a nearby lightning strike). The spark gap is followed by a small impedance and a power zener diode (around 33 V). The spark gap is available as a packaged component, with an arc-over voltage rating (usually around 60 V to 90 V). Extra-heavy traces are needed on the PCB that are able to support the current that will flow through the small impedances, into the spark gap and zener, for the amount of time until the fuse will blow. The spark gap and zener can handle repeated activations without degradation or failure.
A: You ask if transient protection is required for the transmitters if your listed three conditions are fulfilled simultaneously?
If the transmitters are "smart" (i.e., microprocessor-based), I would strongly recommend the use of internal transient protection unless the following additional conditions were all met:
- Your items #1, #2, and #3 and building/equipment/raceway grounds are industrial-grade level of compliance.
- The building is steel-framed (or grounding-conductor-framed) in a way that provides all equipment with the proper cone of protection.
- Further, the power distribution system in the building has a coordinated surge protection system in place down to the branch circuit level.
- The process is not hazardous so that safety is not an issue, because a run-away reaction will not result if transmitter components are damaged.
- Loss of process uptime is not an issue.
- Damage to downstream equipment (e.g., analog inputs) is not an issue.
- Transmitters are located in the same ground plane as the devices they are wired to.
- The plant site is in a very low lightning strike frequency area.
- PRIOR USE indicates surge protection is not needed.