How does everyone handle lighting transformer disconnects? The manufacturer, such as Daykin, provides installation instructions showing that they are to be wired to the line side of a control panels main disconnect. In some cases I've seen these lighting transformer enclosures mounted within the control panel and other times they are mounted to the exterior and connected with a chase nipple.

I always model all feeds into a panel and label it with the worst case IE. In my opinion when a device such as this is connected to the line side terminals and is not de-energized by the main disconnect it should receive a label of its own. This is especially true if it is mounted to the exterior of the control panel.

A big issue I've run into is with panels that have large incoming feeds. I have multiple cases where the upstream OCPD is between 400 and 600A. The feeders from these devices come into a panel and terminate onto the line side of the main disconnect. Then, either 14 AWG or 12 AWG wires also connect to the main disconnects line side terminals and route to the lighting transformer disconnect. When I model the circuit the IE at the lighting transformer disconnect is incredibly high as it cannot trip the OCPD so the calculation runs for the full 2 seconds.

My customer believes we should make a duplicate arc flash label and label it the same as the control panel it is wired to. I have concerns with labeling a device with a relatively low IE that I can model and find to be much higher.

Aren't these devices single phase?

Even if they are 'tapped' on the line side of the panel disconnect, don't they have about 1ft of #12 conductor in front of them? If so, why not put them into your model and treat them like any other device?

This kind of sounds like the issue with PT drawers. Typically a small wire comes off the main and lands on the PT although the PT itself usually has a small high voltage fuse on top of it. Theoretically the arc flash potential in PT drawers is insanely high but in practice I've never seen significant damage to anything around it. Since a #14 wire is used as the "fuse" for most of the IEEE 1584 arc flash testing because it quickly vaporizes and provides the arcing fault path, my best guess is that even if an arcing fault were to happen, the wire would vaporize just as it does in the arc flash tests and either the arc extinguishes, or continues to arc but down on the main lugs in the panel below it. Other possible explanations is that the constriction effect prevents significant current flow, or the fact that the wire resistance increases so much with the increase in temperature that this alone puts out the arc (an unintentional current limiting effect). Either way I've just never heard or seen such a thing happen.

If you do not have a primary fuse no larger than 125%/250% (depending which NEC rule you are using) of the transformer FLA, you have a NEC code violation. If the transformer enclosure does include the OCD, you may have an issue with IE. This is single phase, however, it is complicated by the fact of the small size of the device you mentioned.

The solution is to use the NEC tap rules to feed an external fused disconnect disconnect, that can then feed the transformer. The fuse will mitigate the IE to less than 1.2 cal/sq cm inside the transformer enclosure.

Here is the only problem with the added disconnect, it is all interpretation, I say you are not leaving the enclosure, if you close nipple the switch:

This is for the NEC 10 Ft Tap Rule
"For field installations, if the tap conductors leave the enclosure
or vault in which the tap is made, the ampacity of the
tap conductors is not less than one-tenth of the rating of
the overcurrent device protecting the feeder conductors"