Here is a question that comes up from time to time so let's see where everyone stands. Since IEEE 1584 does not address single phase arc flash calculations, yet the NFPA 70E requirements for labeling (and most everything else) applies for voltages 50 V and above where equipment may be serviced etc. while energized, how do you address the arc rated PPE requirements? Category 0 goes away with the 2015 Edition.

For arc flash labels on 120 V single phase systems/circuits, how should the arc flash protection be determined?

Base it on 208V 3 Phase Calculations

NFPA 70E PPE Categories Tables

Something else

Yeah, but now we would be able to say these circuits pose little arc flash risk therefore "No PPE required".

1. Category 0 does not "go away". It becomes the default minimum PPE, same as most folks treated it. So it is required for all electrical tasks where an arc flash may occur, no matter how small.
2. All calculation methods published fail to address arcing conditions where the arc is not self sustaining and stable. They grossly overpredict because they are designed for the stable case.
3. DC 130 V arcs have been tested. At 20 kA they last for 0.8 secs at 0.5" and arc flash at 18 inches is 1.3 cal. For 120 V AC obviously time and incident energy will be less with zero crossings.
4. NESC sets a threshold of 4 cal for under 300 V but this is partly because in order to alleviate issues with other hazards like molten copper they start at "Category 1" in that standard. The 4 cal limit is based on testing up to around 40 kA or so.
5. Currently 1584 sets a threshold of 125 kVA and 208 V as "need not be calculated" which has been widely interpreted as meaning less than 1.2 cal. Other published data suggest this should be lowered to 4.5 kA (45-75 kVA).

So I am using three results...use the 1584 "rule" where it applies, and then if its not too much trouble model the single phase case as 3 phase, usually when I have a large transformer and distribution breakers that clearly reduce it significantly, but applying an upper threshold of 4 cal. This last step just captures a few extra "1.2" cases that exist due to small upstream breakers. SKM defaults to Lee since it is outside of the valid 1584 range. Lee is most likely off by a factor of 4 or more. Trip times should also be limited to 0.5 secs based on published test results rather than 2 seconds.

Ultimately I think the best approach for "margin" cases like this is going to be a table driven approach since empirical methods so far are no good this low.

I think there is too much shooting in the dark concerning this issue. In order to put this issue to rest, we need empirical evidence of what level of available short-circuit will be the threshold of an unsustainable arc at typical gap distances in 120/208V and 120/240V panelboards or 120V industrial control panels. Once we know that level of short-circuit, for equipment below that level of short-circuit, then all we have to do is determine the clearing time which one would guess at 1/2 cycle. Using that we most likely would be able to conclude those panels would be below the 1.2 cal/cm^2 level. For short-circuits above that sustainable level, then I think again empirical evidence will help. I would guess that the standard IEEE 1584 equations based on 3-phase short-circuits don't quite fit for single phase short-circuits.

As for what I currently do: Apply the IEEE 1584 equations to the equipment as if it were 3 phase equipment and find the clearing time from the equipment TCC's. Obviously, 1584 says single-phase panels are outside of bounds for these equations, but what else is there? Declaring them to be of too low incident energy without empirical evidence is premature. Using this method, I quite often find 120/240V & 120/208V panels to have fairly high incident energies, quite often into Category 3. So if there is a general belief that this equipment would have a lower incident energy to warrant the PPE=0 category, the current IEEE 1584 equations don't support that contention. Clearly some additional guidelines are needed. Again, empirical evidence would be helpful.

IEEE 1584 testing has a single value at 208 V and I think 4 more at 250 V. All other test cases were so unstable that they didn't get any data. In DC testing tests have been conducted by Kinetrics at 130 V and they were able to get stable arcs at 5 kA and 20 kA at 0.5" spacing but any further and they could not get much of an arc. Even at 20 kA it self extinguished in 0.8 seconds. Most cases that have been published for even 208 or 240 VAC extinguish in about 1-2 cycles but there are some exceptions.

The real challenge though is that even while it is arcing the arc is so weak that it doesn't match the calculated arcing current to short circuit current ratio as calculated in IEEE 1584. And the arc can even extinguish and then reignite if it is 3 phase. Even with all this to consider though it would seem like we can simply do a lot of testing to measure the threshold below which arcs don't sustain and then use that as a lower cutoff for incident energy calculations. Not so! There are documented cases where the incident energy exceeds 1.2 cal/cm^2 and even a few lab cases where it got up over 4 cal/cm^2. This precludes any kind of "cutoff" approach. Like it or not, we're going to have to address this area. Hence the reason that I'm thinking that in the end it might have to be table driven.



NESC quotes empirical evidence and sets a threshold at 4 cal for 250 V or less regardless of the equipment configuraiton. Test work was performed by EPRI.