What Radiant Floor Heating Actually Feels Like — And When I Tell People Not to Buy It

Almost every radiant floor conversation I have starts the same way. Someone stayed at a hotel in the mountains, stepped onto a warm bathroom floor at six in the morning, and has been thinking about it ever since.

I understand it completely. But I want to set expectations honestly before anybody spends money, because a correctly designed radiant floor in a well-built house does not feel like that hotel bathroom.

It feels like nothing. That's the point.

The math behind the feeling

Floor surface temperature isn't a setting. It's an output, and it comes from a very short equation:

Floor surface temperature = (upward heat load in BTU/h per sq ft ÷ 2) + your room setpoint

The floor transfers about 2 BTU/h per square foot for every degree it sits above room temperature. So it only gets as warm as it needs to be to replace the heat the room is losing.

Run real numbers on that. A well-insulated, code-built house today has a design heat load around 10 to 15 BTU/h per square foot — roughly a third of the 30 to 35 that mid-20th-century housing stock needed. At 12 BTU/h per square foot upward with a 70°F setpoint, that floor sits at 76°F.

Seventy-six degrees. Your hand is warmer than that. The floor reads as neutral to a bare foot, not warm.

And that's at design conditions — the coldest day of the year. On a mild October afternoon the load is a fraction of that, and the floor sits only a couple of degrees above room temperature. You won't feel it at all.

A worked example out of the Uponor manual I design from: a bedroom at a 65°F setpoint with a 19.8 BTU/h per sq ft upward load lands at a 74.9°F floor surface. That's a well-designed radiant floor at full output.

What you do feel is different. Radiant heats objects and people directly instead of heating air and blowing it around — no drafts, no cold corners, no blast of hot air followed by twenty minutes of slow cooling. Most people say "the house just feels right" long before they say "my floors are warm."

If someone is selling you a floor that's warm to the touch all winter, they're selling you either an oversized load or an under-insulated house.

The ceiling nobody mentions: 30 BTU/h per square foot

Radiant floors have a hard output limit, and it's the most important number in the conversation.

I design to a maximum of 30 BTU/h per square foot of radiant floor — a cap that exists for comfort and for floor protection.

The comfort side comes right out of that surface-temperature equation. Standard practice limits floor surface temperature to 87.5°F for most floors and 80°F over hardwood. Work it backward: at a 70°F setpoint, an 87.5°F ceiling means the most a room can carry is 35 BTU/h per square foot — and over hardwood, where the cap is 80°F, it's only 20.

The honest consequence: if a room's verified heat loss divided by its available floor area exceeds 30 BTU/h per square foot, radiant alone will not heat that room. No amount of hotter water fixes it. The floor cannot deliver more than the surface will give up.

That room needs a better envelope or supplemental heat — usually a panel radiator on the cold wall. I'd rather tell you during design than have you find out in January.

In an older Denver house with single-pane windows and no wall insulation, a room with three exposed walls and a wall of glass can blow right through 30. That isn't a failure of radiant. That's radiant telling you the truth about the envelope.

Floor coverings: what fights you

The covering sits between the tubing and the room, and it's insulation whether you want it to be or not. Every bit of R-value it adds gets paid for with hotter supply water — roughly 12°F of supply temperature for every 0.5 of covering R above R-1.0.

• Tile, stone, and bare concrete — best case by a wide margin. Quarter-inch ceramic tile is about R-0.23; marble and quarried stone are in the same neighborhood.

• Hardwood — moderate. Half-inch oak is about R-0.56. Workable, but the 80°F surface cap over wood tightens your output ceiling further.

• Carpet — the one that ends projects. It's not just the carpet, it's carpet plus pad, and they stack. A quarter-inch nylon saxony at R-0.88 over a quarter-inch bonded urethane pad at R-1.04 totals R-1.92 — that assembly alone adds roughly 22°F to your required supply water. Move to a half-inch wool plush at R-2.20 over a thick pad and you're past the point where the system makes sense.

Deep carpet doesn't reduce radiant output. It effectively eliminates it. If plush carpet is going in the main living areas, I'll tell you radiant is the wrong system for that space — and I'd rather say it early.

When I tell people not to buy it

Four situations, and I say this out loud on site:

1. The envelope can't support it. If the load exceeds 30 BTU/h per square foot, radiant alone won't get there. Sometimes the right answer is insulation and air sealing first — and then radiant works beautifully at a lower water temperature.

2. The floors are already down. Tubing in a slab goes in before the pour, and that window doesn't reopen. With finished floors the alternative is staple-up between the joists — which starts around 150°F supply water instead of the 98 to 120°F an insulated slab needs, and delivers less. It's a real option, but those supply temperatures climb high enough to push return water out of a condensing boiler's efficiency window. You lose the low-temperature efficiency that made radiant attractive in the first place.

3. Deep carpet is going down. See above. This one is physics, not preference.

4. You want fast response and deep setbacks. A slab is a thermal battery. That's a feature — it buffers supply-temperature swings and rides out interruptions gracefully — but the floor doesn't turn on and off quickly. If your habit is dropping the thermostat well back at night and snapping it up at 6 a.m., a slab will frustrate you.

One exception I'll always argue for: an interior bathroom with tile has almost no calculated heat loss and technically doesn't need heat at all. I'll still loop it. That's the barefoot-comfort floor everybody is actually picturing — a deliberate decision rather than an accident of load.

What good design buys you

When it's right, the floor sits at a modest temperature, the boiler runs at the lowest water temperature the emitters allow, and return water stays cold enough that a condensing boiler actually condenses. Low water temperature is where the efficiency lives — that's the whole reason radiant and mod-con boilers belong together.

Getting there means running the procedure room by room: heat loss, then feasibility against the 30 BTU/h cap, then tube spacing, then supply temperature, then flow, then head, then equipment. Every step feeds the next. Skip the first one and everything after it is a guess.

More on how I design and build these: https://www.bydesignhydronicsandair.com/radiant-floor-heating-denver

If radiant is on the table for a new build or a remodel, book a site review at https://www.bydesignhydronicsandair.com/contact-us-1 — ideally before the slab is scheduled. I'll walk the house and tell you honestly whether the floors can carry the load.

If they can't, that's a useful answer too.

— Ryan Van Inwegen, ByDesign Hydronics & Air, Littleton, CO