Hi. A client of ours is asking for a label to be put on equipment not requiring an arc flash label. The label is meant to be an indication for untrained management that conduct safety audits to understand that certain pieces of equipment don't require arc flash labels. After a careful reading of NFPA 70E and the NEC, I can see only two cases where this label would be useful (by which I mean, we wouldn't have included the equipment in a study):

1) in cases where the equipment is less than 50 volts.
2) in cases where the equipment is never examined, serviced, etc, while energized- but even in this case, the commentary in the 70E handbook says that for equipment over 50V, inside the limited approach boundary a hazard is presumed to exist.

Recent tests have shown that it's pretty near impossible for single phase loads under 208V to sustain an arc, so I'd imagine that most plug-in equipment could have such labels, but I'm curious to know if anyone else has gotten this request and if so, what you did.

Thoughts?

This actually touches on one of my pet peeves for IEEE 1584 and NFPA 70E. There is a gap between what NFPA 70E requires for equipment that requires labels and what IEEE 1584 covers. The end result is it leaves us hanging without clear guidance.

NFPA 70E, 130.5(D) requires labels for ... panelboards, industrial control panels... It also requires the label to have the arc flash boundary.

So that is what my clients ask for. This clearly would include single phase panelboards, and industrial control panels with small loads.

However, IEEE 1584 states in section 1.2 single phase ac systems and dc systems are not included. So now what do we do to fulfill the labeling requirements?

It might be thought you could use 70E table 130.7(C)(15)(A)(b), but it has large gaps in what it covers. For example for a panel rated 240V, you have to have a fault clearing time of 0.03 s. If the upstream breaker(s) doesn't clear in that time, what goes on the label?

Lately I've been using ANSI/IEEE C2 Table 410-1, Note 2 (which sites EPRI Testing) stating that arcs will not be sustained for more than 2 cycles and applying that to the IEEE 1584 equations to get an arc flash boundary. Not technically right, but what else is there? At least it is a round-about way of getting back to same answer of the NFPA 70E table answer.

Is your client looking for a IE level or just a label? A generic "Arc Flash hazard Exits" may be enough - like the example in the NEC Handbook.

70E/IEEE joint testing has shown that arcs can be sustained for more than 2 cycles, and the arcs are a little different. IEEE 1584 has precisely ONE successful test at 208 V that is used to estimate the equations for the entire range of arc gaps, currents, and voltages down to that point. So this is definitely not right.

So in my mind going by standards, there are 6 choices when it comes to the under 250 V case:
1. Use a generic label as per NEC as suggested. This is side stepping the issue because the very next question will be...ok, what's the boundary and PPE. But if its going to be ignored anyway and its just to meet a Code requirement, that's good enough.
2. Use the tables in 70E. Works for some but not all scenarios. Overestimates.
3. Use the "125 kVA or less, one transformer, under 240 V" exception from IEEE 1584. This is being re-examined so at some point it will decrease (down to around 40-60 kVA) but until then, it's there. This is limited to 3 phase. It is technically incorrect (but over-estimates) to use 3 phase results for single phase.
4. Use actual test data. Works OK I guess if you have unlimited funding.
5. Use the TABLES from NESC (IEEE C2) outright. There are 2 clauses but the net result is 4 cal/cm^2 across the board at the normal working distance that can be re-normalized to calculate an arc flash boundary. There is more to this by the way than just "2 cycles". The 2 cycle claim is from test work from PG&E that has been invalidated (thus the reason that a subcommittee for IEEE 1584 is relooking at the 125 kVA rule). Instead use the table values because they actually lab tested these and you can look at what went into the EPRI reports for free. They only got to around 3.2 cal/cm^2 with a 250 V arc with a 20 kA available fault current. You could dig into the source further but you are in a defendable position to simply use the NESC (consensus safety standard) as your basis rather than trying to play games with the math behind it.

Personally, I like a combination of #3 and #5. I can use #3 for "low level" conditions where incident energy is not going to be a concern anyways and then switch to #5 for everything else. This approach relies 100% on consensus safety standards and not on vague interpretations of things.