On a 37.5 KVA transformer, with a 5% impedance and 240V secondary, the worst case fault current is going to be just over 3000A. With cable it will be reduced, so on an arcing fault it is unlikely that you would ever see the full 3000A. Maybe something like 38%.
As far as breaker tripping, the FLA of xfmr is 156A, so the breaker would probably be a 200A. Assuming non-adjustable trip, for a anything over 1500A, the breaker should trip within 7 cycles, or around 0.12 secs. for the 38% of the 3000A, your looking at about 2 secs.
I think part of the issue that is possibly being overlooked, is that at lower voltages, and certainly with smaller transformers, the energy behind the fault, available to sustain an arc is extremely limited, and therefore the amount of energy available to cause a non-recoverable injury is unlikely. Hence the need for only Level 1 PPE.
Case in point:
Many years ago, I had the unfortunate experience of grounding a single phase of a three phase 208 volt lighting panel in a grocery store, with my 3/8" wide, 12" long Heavy-Duty Round-Shank Klein screwdriver. I was changing out a bolt on breaker, the srewdriver slipped, and Wham!! Other then feeling quite shaken (no bodily fluids expelled)and stupid, some ringing in my ears, and looking like a racoon from my safety glasses, (my face was coal black from the flash), the leather gloves and insulated handle on the srewdriver seemed to work. I still have the srewdriver (minus a rather large chunk of steel on the shaft that became flash welding material) to this day, as a reminder to think safety first. Needless to say, it was a short day of work, and much, much beer was consumed (prior to lunch)
This is certainly not scientific proof, but IMHO, I would find it hard to recommend, or enforce, anything over PPE level 1 for residential panelboards, or small 208Y/120V lighting and power panelboards.
