Yowza! The emissivity of that galvanized steel is significant- you really ARE pumping as much or more heat into the basement than the first floor, since the surface area presented (due to the corrugation channels) is something like 30-50% more than a flat-surface first floor, and the emissivity of the tile is only somewhat higher (probably not more than 2x the emissivity of galvanized steel more than a year or two old, and the emissivity of the galvanized will go up over time.) A cheap application of radiant barrier would cut that significantly, but not nearly as much as "real" insulation.

Applying fiberglass insulation to the bottom of the steel would be very difficult to do well. An application of 3 to 5" of half-pound foam (Icynene, et al) would probably be the best approach (not cheap, by any means- as a ball park guesstimate multiply SquareFeetFloorArea x InchesOfThickness x $0.40 ). At only 3" you'd have ~R11, but it would have zero radiant & convective losses, performing better than R13 fiberglass. Then something needs to be applied to make a thermal barrier between the spray foam and the basement (even in an unfinished basement space) to meet code. The cheapest is probably 1/2" gypsum board applied to the underside of the steel trusses.

Alternatively, price out what 1.5-2" of foil-faced iso board (Thermax, etc.) applied to the under side of the steel trusses would cost. It'll only be R13 or so (for the 2" stuff), but with the foil facing, would behave more like R17-19 as long as the cavity is relatively well air-sealed (tape the seams with FSK tape, spray-foam caulk the other penetrations, etc.) It'll be way more expensive than rolled out aluminized mylar bubble-pack, but it'll be way better at doing the job too. In most places foil faced iso doesn't need a separate thermal barrier to meet code (but check.)

That said, it's still not the primary problem, if the Polaris won't turn on (or stay on) full-fire. If it's anything like 90% steady-state thermal efficiency, at ~200kbtu in it should have margin enough to keep reasonably tight well insulated 3.3k' house over a well insulated 3.3k' basement iup to temp inside when it's -15F outside. If it's the smallest 130kbtu/hr in model and the house is leakier than average (even if well insulated) you might be at the margin at -15F, but it should still do just fine during normal winter weather(average daily temps well above zero.) I'd be somewhat surprised if a SIP structure would have that much air leakage, but a well-insulated wind-tunnel is still a wind tunnel. Pressure door testing could tell a lot (and air-sealing is usually comparatively cheap-well worth doing if it's a leaky sieve.)

The Polaris water heater uses a negative pressure gas valve. You cannot set one of these up properly without an incline manometer or similar tool. If the filter gets plugged the unit will lock out. The burner assemblies on the newer units have a very fine mesh and cannot see any particulate. The installer did a good job installing the filter. IAt may be time to change or clean it. How long is the vent pipe system? Could it have exceeded manufacturers recomendations? Does the vent pitch properly? Could be icing in the vent line causing restriction. Along with agreeing you need to insulate the 1st floor if you have popping or thumping coming from the polaris you have a serious issue with the wrong combustion settings and could cause serious damage to the unit or your home. Make sure the tech that comes out knows how to work on a negative pressure valve

I wonder how is all turned out? Close that "open" system yet. Get some control of the design water temperature. Do a heat load?

I'd generally take a thinner aluminum plate that covers more area than a thicker one.

I would let you have it Jon :-). All the evidence leads to the extruded plate for better tube-to-PEX contact and higher output as a result. I use the thin plates for walls and ceiling where supplemental alumimum foiled rigid insulation will be used.

If there is a really big dT on the tubing, it is actually better for the polaris but I will wager that the problem could be an issue with the hot surface ignition (which they used to use, and they often messed up, but I'm not sure if they still use it) or as stated above, condensation in the vent or plugged filter.

Yes on delta T, doubtful on ignition. We have dozens of them in the field, yes with hot surface, and few problems when properly set up on the right system. When they say "open" heating system, all bets are off. The water heater will take the abuse but nothing good can come from it. Short cycling can be caused by many things, mostly user or installer error.

My experience with Polaris is many years old but I am re-thinking using them in low energy installs. When I did use them and did service work, the igniter was a one of the main issues, along with the gas valve on some of the early ones.

We no longer see ignitor problems on the Polaris and since it has no equal in the field...