What size pex tubing do you recommend?

Tube diameter has little to no effect on BTU output per SQ FT if properly installed. 1/2" is cheapest and easiest, and with only ~1800SF, 5 loops per bay at 250ft each w/ 6" OC is going to be more than adequate. There is absolutely no reason to run larger diameter, and anything wider than 6" OC would be a mistake.

You need to move a lot of BTUs of cold glycol in a SIM. Uponor suggest 5/8 or larger tube. 1/2 maybe for some stair treads where loop lengths are very short, you need a tight spacing.

Assuming the bays are covered you are designing more of an ice removal or anti-icing protection system, no real snow load to speak of.

So the tube and spacing as well as loop length dictates and SWT/ gpm dictates how many BTUs you deliver.

All the info about SIM design here.

https://www.uponor.com/getmedia/50a684b6-091d-4044-b8f7-c06f7705668c/snow-and-ice-melting-design-and-installation-manual.pdf?sitename=USA

What delta do you design at?

With a cold start SIM system, a 20* delta is largely ridiculous to get any amount of BTU output regardless of tube size. Space heating, sure- but a cold start snowmelt will almost never hold a 20* delta while also doing its job properly. 50+* is not uncommon and there is absolutely no downfall to doing that, quite the opposite in fact. Maximized BTU output with minimized upfront cost and the highest possible efficiency, all great things.

With that said, car wash systems typically don't need more than about 40-50 BTU/SF to keep the pads clear unless they never close the doors. Any respectable car wash closes down at about 0*, so my suggestion of 1/2" at 250ft is only a 15* delta at .8 GPM per loop, yielding 50 BTU/SF. This is not rocket science, guys.

As I'm sure you are aware, tube diameter is not really relative to BTU output. A 1/2" tube at 100* will put out an almost identical amount of BTU to 3/4" at that same 100* in the same slab. 1/2" tubing at 6" spacing will put out considerably more BTU than 3/4" at 9" spacing, assuming the same temperatures which obviously vary by flow rate. Pressure drop in 250ft of 1/2" at .8 GPM is about the same as 375ft of 3/4" at 2 GPM so while there is potential to carry more BTU with the same pressure drop and fewer loops, larger diameter tubing does nothing for actual BTU output after the dust is settled if spacing and delta are equal. I would personally rather fight 1/2" at 6" spacing than 3/4" at 9" spacing, but maybe that's just me.

Now if it's a large space like a 5000 SQ FT driveway with a 150 BTU/SF load, I'm all aboard the 3/4" train. In the case of a car wash however, there is not a chance I'm wrestling 3/4" when 1/2" is better from all aspects.

Yes, we would obviously need more details if it's not a regular car wash bay that gets shut down at a given temp, but that does not negate the fact that larger diameter tubing does not equal more BTU output per square foot. 1/2" at 6" spacing offers more BTU output capacity than 3/4" at 9" spacing, no matter how you slice it.

You're comparing apples to oranges, Bob. 6" spacing WILL allow more BTU output per SQ FT than 9" spacing. Period. Tubing diameter is irrelevant for the sake of this discussion. On that note, I'm not sure which Bluefin tubing you're looking at, but 1200LF of 1/2" at Supplyhouse is $280 not $360. 6 loops of 1/2" would be considerably less labor than 3 loops of 3/4". I will wholeheartedly agree that 1/2" is not the answer for every application and have mentioned that several times now, but having done this very task many times in one of the coldest areas of North America on top of personally owning a car wash which this post is specifically geared toward, I will say one more time that 1/2" at 6" spacing is the most feasible option here. It's cheaper, easier, and will outperform 3/4" at 9" spacing all the way around. The bays (2) in my personal self-serve wash are 14x42 and utilize 5 loops of 1/2" at 6" OC, 250ft each, with a 4ft heated entry/exit pad. Mine is automated (as most are, at least in my area) to shut down when outdoor temp falls below 0*, and the pump will not start until the door is closed when it's below 32*. We've had a few automation failures over the years where the doors don't close, or those fun days when it goes from 40* to 0* in 2 hours during a blizzard, and not once in 6 winters now has there been any ice accumulation. I run this on a basic Tekmar 519 stat with the floor sensor set to 50*, with an 80k Lochinvar Noble (so roughly 60 BTU/SF max capacity) set to 90* using P/S piping, and very seldom does it see high fire.