I admit to using the floater approach on a couple of fixture changeouts for the few Residential service type jobs I have done [other than family and friends].
I tried as much as possible to put the J-Box's outlet for the change from 60 degree NM to 90 deg. NM, in a location that was not so visiable - trying to give it some degree of hiding, rather than sticking out like a sore thumb!
All the ones I did were fortunatly in an attic space, rather than something like a shallow soffit, or 1st floor ceiling, so there was some room to work around.
The end result of having the 1 gang blank cover on the J-Box was not too bad, plus these customers understood the entire reason for it.
After reading the ideas in this thread, I was also pondering some techniques for recessed Incandescent fixtures, which will limit their potentials for fires.
Asbestos, or glass braiding Insulations [both on the conductors and the fixture it's self] would be one step, but the conductor's heat sinking would be overcome with Aluminum wire.
I don't exactly like Al, but it won't heat sink like Cu or Ag [Silver].
This temperature situation has been a consern in Fluorescent lighting for quite some time [90 deg. Insulation within X inches of the ballast], so why did it take the companies so long to require the recessed fixtures this also??
I know very little of legalistic stuff - please pardon my ignorance
On a semi-related issue, I had the task of rewiring a high intensity tanning bed once, for a good customer that owned the tanning salon. I hooked up all kinds of tanning beds for him at this salon - we learned a lot from each other, too!!
The High Intensity unit had five [5] 2,000 watt HID lamps in the top [only the ballasts, capacitors, contactors, control equipment and unit subpanel was in the bottom].
The original wiring in the top used insulation similar to that used on Ballasts and Transformers [TEW, I think] - something like a max. temp. of 100 + degrees Celcius.
These conductors began to fail after 6 months of normal usage. Failure was a result of the insulation becoming "dried out" and brittle from the high levels of unfiltered UV light available, rather than just the ambient temperature it's self. When the top unit was raised / lowered as a person used the machine, it would move the conductors and create cracks in the insulation. When the lamps were first started, they were pulsed with a 2.5KV series of pulses, just as HPS lamps are. The ignitors are remote from the lamps, and in the case of this unit, they were in the lower section near the Ballasts.
Needless to say, the cracked insulation only allowed the ignition pulses to flow out through the cracked insulation.
Visible history showed that random starting was achieved over the time that the insulation's breakdown was the newest.
After repeated events, the insulation failure became too great, which kept the ignition pulses at the cracked points closest to the grounded frame. At these points, the ignition pulses would work on the conductor until there was a direct contact to the grounded frame. This would not appear as a short circuit since it was on the output of the Reactor type Ballasts.
One by one, the lamps failed to work, until they all quit. That's when I was called in.
The end results were using Glass Braided high temp silicone insulated conductors in the top part, relocating the ignitors to the top, sending the Ballasts [Reactors with intregal Autotransformers wound to the cores] out to be rewound a repotted - they had signs of overuse, resulting in charred terminals and windings.
This was an invaluable lesson to me and the client.
BTW: The company that made the unit had no knowledge of this occuring, as this was kind of a "Prototype" unit. They were so glad to be informed of this by us, that they sent reps and design engineers to the site to help with everything plus expand on the situation.
P.S. They paid for all expenses on our end!
Scott SET