Why treat the racks as one? I would use a dual probe tstat. Racks vary in temp up and down typically anyway I would not add complication of rack a and b it is too difficult to make sure everything is exactly the same. If you have a single poor connection the temps on the rest of the string will be off. I personally would probe the middle of each rack and set each independently. A single run of 11" flexwatt should not be longer than 25 feet, or 24 panels. I would also not use the 20w 11 if anything goes wrong it generates a huge amount of heat. most racks I have seen and my personal one all use 4 inch it also allows a max run of 60 feet.
Personally I would highly recommend the herpstat 2 (195) it has dual probes and a relay (500w per probe) . The VE 300x2 is 265$ (canadian supplier the manufacturer does not ship to Canada) it has the same power specs.
Many use melamine, I think it is ok, I am not a huge fan myself. I would suggest sealing any exposed edges as formaldehyde is used in the composition of the core and it off gasses. What the long term health effects are unknown. I would still lean to PVCx as it is easier to clean and in my experience racks are difficult to clean anyway. There are a lot of surfaces, 10 per snake. PVCx also is a fire retardant... Most softwoods are fine but need to be sealed and same for most plywood again it needs to be sealed. I would not use MDF. Depending of the design I would look at 3/8 baltic birch ply it is no more expensive and massively better quality (robert bury if there is one by you...)