A-B, like several other mfrs, take the IEC style Motor Protectors and submit them to UL for 489 listing, so that they can be used for branch circuit protection of things OTHER than motors. UL-489 does NOT REQUIRE the Phase Unbalance trip biasing, but neither does it PRECLUDE it. So when these devices are UL-489 listed, they still have the "differential trip bar*" that biases the thermal trip setting, meaning if you do NOT pass current through all 3 poles, the trip level will be LOWER than the setting, meaning you will get NUISANCE tripping, not lack of tripping.
* IEC Overloads have a separate spring loaded balancing bar on the trip mechanism that is held at abeyance by equal pressure from the three separate thermal elements acting on it equally. If you lose one phase, the pressure from that thermal element is lost and the balance bar causes the trip mechanism to be shifted to already be closer to tripping, so it takes LESS current on the other two remaining elements to cause a trip. By looping one phase back through, you are ensuring that all three of the thermal elements are exerting the same pressure.
The Short Circuit Current Ratings (formerly "withstand" ratings) of most of these IEC style devices are based on them being designed for use on 400V nominal systems that are ALWAYS 4 wire Wye where the Line to ground is never more than 230V. So the asymmetrical fault capability of the mechanical design is based on the amount of energy that could flow to ground, and at 230V, it is less total energy than it would be at the full phase-to-phase voltage. When these devices were retested for 480V, the asymmetrical fault rating was done at 277V and it passed, so they were listed for 480Y277V only. But with 480V delta, the L-G would be 480V, so a lot more fault energy and they will not pass , they will likely explode catastrophically.
I don't use them much because as best I can tell they are not suitable for 480 V unless it is a wye system, and mostly I do not have any way of knowing that.
