Partial Radiant Heat – Slab on Grade
![](https://secure.gravatar.com/avatar/d75e52df98dbd2613fb7019096fee156?s=52&d=https%3A%2F%2Fimages.greenbuildingadvisor.com%2Fapp%2Favatars%2F200x200%2Ft.png&r=g)
So I know radiant is beat to death here. Looking for some advice on it anyway! We have our house design pretty well complete. Double stud wall, around 2500sqft, climate zone 5. HVAC is going to be an A2W heat pump that feeds some circuits in both the garage and house. The A2W heat pump is going to feed a ducted fan coil that will feed most of the bedrooms, etc. It will feed a wall mount unit in the living/dining/kitchen room for A/C. The last thing I have it feeding is some heated floor in the open room. I have that sized such that at the design temp, the sqft of heated floor is less than the heating load of that room. The fan coil picks up the remainder of the capacity in this case.
All that said, what I’m looking for is how you would feed that portion of heated floor. It’s around 300 sqft of flooring in a room that totals around 800 sqft. The house is slab on grade. My intention is to pipe the PEX in the slab for that portion. I don’t know how well this will work at all. The other option is to put something such as warmboard or similar on top of the slab. But then I need to top the whole slab with subfloor to keep it even in height, which adds a decent bit of cost. I was intending to use engineered hardwood directly over the slab. Any advice is appreciated! Thanks.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb1.png)
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb2.png)
Replies
Tim_O
Sorry, I don't have any useful answers to your question. Among GBA posters there seems to be consensus that the approach you are taking is the right one. Spots of radiant, rather than all-radiant.
I hope you post your build as it progresses.
I totally agree. The forum has a number of people here that have a lot of expertise in it.
PEX in concrete is probably the cheapest way to do a heated floor. What you have to balance it against is the concrete has a fair bit of heat capacity, and the more heat capacity your emitter has the less responsive it is. Warmboard in particular is designed to be low heat capacity and highly responsive.
To get an idea of how big a factor that is, take the area of the floor and the thickness of the slap to calculate the volume of the floor in cubic feet. Concrete is about 130 pounds per cubit foot, and it takes about half a BTU to raise one pound of concrete by one degree F. Then compare the output of your heat to the amount of heat needed to heat the slab to a temperature where it's putting out heat.
For example, for a house built to modern standards I wouldn't expect a heating load beyond about 10 BTU/square foot. So let's say your 800 square foot room has a load of 8000 BTU/hr. The heating area is 300 square feet, at 4" thick that's 100 cubic feet of concrete or 13,000 pounds. To get 8000 BTU/hr out of the heating area that's 27 BTU/square foot, the rule of thumb is you get 2 BTU/square foot per degree of difference, so that slab needs to be 13.5 F above room temperature. Heating 13,000 pounds of concrete by 13.5 F takes 88,000 BTU. So if your piping is capable of delivering 8,000 BTU/hr, it would take 11 hours to warm the slab up.
And once it's warmed up, there's no turning it off. It's going to continue putting out heat until the slab cools off or the room heats up to slab temperature.
Whether this is a problem depends on the house and climate. In some places you're going to have a constant need for heat from fall to spring, the air handler is enough to cover the variations from hour to hour and having a slab cranking out a baseline is OK. My house has a lot of solar gain, on sunny days in the winter the heat turns off altogether and a big slab would have us opening windows.
If you want something with less heat capacity, Warmboard is kind of the gold standard, but it's expensive. You could also go with a thinner slab. Concrete gets weak if it's too thin, but what I've done is embed tubing in 3/4" of mortar, then cover it with backerboard and tile over the backerboard. There are also various aluminum plate systems that allow you more flexibility in flooring. The key to floors is you want something that is highly conductive of heat, you don't want cold spots or hot spots in your floor. The reason concrete is popular is it does have good conductivity.
If you're going a slab you shouldn't have to worry about differences in floor height if you plan it all out beforehand. Or you could do a concrete-free slab, I seem to recall the GBA detail library has that although I couldn't put my fingers on it. I'm not a big fan of concrete floors in houses and I am a fan of reducing the use of concrete in general.
Thanks DC. The calculations around concrete make sense. I was unsure how to model the connection between the unheated and heated portion of the concrete. If 300sqft has PEX embedded, does 400 sqft of concrete actually heat up, or maybe even 500.
When it comes to uneven floor, my issue is that I want to run the same wood floor down the hall, on a good amount of the first floor. So I'd be unnecessarily laying subfloor over the whole slab pretty much. We did look at a concrete free slab. Right now, we are looking at a monolithic poor using WarmForm. We have a load bearing wall down the middle and an on grade tornado shelter that complicate doing a slabless slab, but I'm not opposed to the idea if we could make it work.
> If 300sqft has PEX embedded, does 400 sqft of concrete actually heat up, or maybe even 500.
With 300 sqft heated, we needed 26.7 BTU/sqft and a surface temperature 13.3F above room. We had 13,000 lbs of concrete so it took 86.6 kBTU to heat it.
With 400 sqft heated, we only need 20 BTU/sqft and a surface temperature of 10F above room. We have 17,300 lb of concrete so it takes 86.7K BTU to heat it.
With 500 sqft heated, we need 16 BTU/sqft and a surface temperature of 8F above room. We have 21,700 lbs of concrete so it takes 86.7 kBTU to heat it.
With the same slab thickness, the heat is the same regardless of the area.
For sure, all the heat makes it to the room no matter what. I guess what I mean, I'm trying to design it such that the floor temp would be in the 75-80 range in my heated area most of the winter. So my question is more along the lines of, how wide does the 86.7k btu spread before radiating through the floor. Probably not an easy thing to model.
I'd say 75 won't be noticeably warm, 80 you'll notice but not hot, that's a good temperature.
At 80 and air temperature of 70 you'll get 20 BTU/sf, that's 6000 BTU/hr with 300 square feet of heated floor. With air at 72F it's 4800 BTU/hr. At a floor of 75 you get 3000 BTU/hr at 70F and 1800 BTU/hr at 72F.
You said your design load is 15k BTU, I'd guess your design temperature is about 5F, so 60 degrees of heating (assuming a break-even temperature of 65F). So plotting your heating curve with those four numbers, I see:
1800 BTU -- 57F outside
3000 BTU -- 53F outside
4800 BTU -- 46F outside
6000 BTU -- 41F outside
When it's below 41F, the air handlers start kicking in.
I have a Chiltrix heat pump, it has automatic outdoor reset and allows you to program in a temperature curve. So you'd want to program it so that at 41F and below it's at 100%, then straight line to 0% at 72F.
A2W heat pumps tend to like a smaller temperature drop than boilers, 10F is normal. Half inch pex is good for about one gallon per minute, 1 GPM at 10F is 5000 BTU/hr.
You sure you don't want to do the upstairs bathroom floors too?
Yep, 5f is the design temp. Likely the downstairs primary bath and upstairs bath get heated floors too.
Closer to 75 is where I keep the small bit of heated floor I have. It's noticable, not super warm, but sorta warm. It might be more like 77-78, my infared thermometer is an Amazon special.
I need to review the heat pump specs. I had been planning for chiltrix for a while, their customer support seems great and knowledgeable. However, I had not realized it drops off that far at 5F. For some reason, I thought it was more capable, but I think I mixed it up. The Hydrosolar EVI pump maintains ~19k BTU up to 122f, so I can use it for DHW even.
With a wood floor I think you should consider not putting it directly on the slab. I don't have a specific recommendation, in the GBA detail library there are lots of variations of floors over slabs. If you put your insulation between the slab and the floor you can decouple them.
That would give you room to run stuff under the floor, not just the heating but also things like wiring, which can be handy. It also makes it much easier to adjust for different construction in different sections and get them all the same height.
I'll take a look! I wasn't intending for hardwood, but more likely engineered hardwood. Spending all my money on hydronics, gotta cut back somewhere! Lol.
Slabs with the proper temperature water running though them do not have problems with overshoot.
With all due respect to DC, radiant heat has been working for 75 years, some systems better than others, much like any heating system
One never has to heat up the slab from 'cold' but once a year.
Probably the worst problems come in houses with a lot of sun loading, where the system may be off for many hours during the day.
In a well insulated house, when the sun goes down, the temperature will take hours to drop, and even then, the slab is still pretty warm, and by the time you would have noticed it, the slab will have recovered.
But a well insulated house shouldn't have enough glass to cause that problem, right?
THe moral is that make sure you have a primary secondary loop so that you can lower the water temp to the slab, since you will be outputting as hot a temp as you can for the fan coils.
>Slabs with the proper temperature water running though them do not have problems with overshoot.
One realization completely changed the way I think about floor heat: the output of a floor is entirely determined by the difference between the surface temperature of the floor and the air in the room it's in.
I don't know if that is obvious or non-obvious, it wasn't obvious to me until suddenly it was. And it changed everything. There's a lot going on when you design a heated floor -- tube spacing, flow rate, water temperature, loop length, the materials in the floor and below it -- but it all comes into focus when you realize that what really matters in the end is surface temperature.
Now, I'll agree that the best way to control the surface temperature is by modulating the temperature of the water being fed to the floor. All of the methods of control do that one way or another, whether it's turning the flow on and off with a zone valve or adjusting the speed with a variable speed pump, modulating the temperature directly is just the most effective way of doing. Is modulating the water temperature always going to give you the surface temperature you want, when you want it? Of course not.
Now, is it possible to have a floor that mostly or almost always gives the surface temperature you want? Sure, I said so in my response, it can happen. The less heat capacity your floor has, the more sensitive the surface is to water temperature, and the more responsive it is to the controls.
It's absolutely true that people have been building heated floors for a long time -- longer than 75 years, the Romans had them. But that history is littered with many floors that just never worked -- they didn't produce enough heat, or produced too much, or in order to produce enough they got too hot to be comfortable. What those floors all had in common was they weren't designed. There's an awful lot of "build it and see what happens" in this area.
The key to designing something is understanding it and being able to predict its behavior. Heated floors are particularly tough because every one is hand made and slight variations in materials or construction can have an impact on performance. But if you're not starting with an understanding of the behavior you're just guessing.
The house I grew up with, in a development of 100s of similar houses had a heated uninsulated concrete slab in zone 5 . It worked as well as any baseboard system, in that is was not perfect, but was not a noticeable issue. Most of those systems were replaced not because they did not heat well, but because they were copper coils in the slab and rotted away.
When you talk about large temperature swings or inability to recover, you are talking about very badly designed systems.
In most badly or standard insulated houses the struggle is to get enough feet of tube in the floor to output enough BTUs with low enough water temperature to have a comfortable floor temperature. Poor[or no] design leads to too high a water temperature to get the needed BTU output, which results in [taadaa!] overheating the slab and temperature swings
In a well or super insulated house I just cannot see this being an issue.
Most of the conversations here center around the floor temperature not being high enough to notice, which in my opinion can be dealt with.
I love radiant heat, but the real issue is once you insulate the floor and stop all the heat loss and exfiltration, the floors aren't cold and who needs it? I went to the time and trouble to run tube[retro above the slab] in the entire downstairs of my current house and have never hooked it up. An insulated floor is not a cold floor
Oh, and my last radiant concrete slab was in my garage, and because the boiler was in the garage, I realized the heat never ran there. So to keep the slab warm, I bought a cheap setback thermostat and had the slab come on for a short period of time once a day. If I was working there I could set it to whatever I wanted.
In the narrow circumstance that water cool enough to prevent overshooting of the room temp caused too long of a warm up time one could use this strategy
The other complication - we do have a good amount of southern glazing in that room. Decent overhang, but it's mostly exposed by the equinox. The glass will all be lower SHGC, in the 0.3 range I think.
But makes sense, seems feasible, all in all. Might take a little controls work between the floor and fan coil to dial it in.
I found LoopCAD to be useful in calculating the floor temperatures of the individual zones. You start by doing your Manual J for the heat loss and then adding the radiant. It will calculate the btu per hour to heat the zone. I found that bathrooms, porches and basements is where radiant makes sense.
Attached is a sample pdf the download/evaluation version is free.
Awesome, I'll take a look! Always like to learn a good tool. I used Borst calculators for a lot of it. They have some radiant design tools, solar gain calculators, heat/cooling load, etc.
If you're going to be using air handlers for a fair bit of heating and cooling I'd suggest looking at one air-to-air and one air-to-water rather than doing it all with air-to-water. It's not that air-to-air is better, it's that it's a lot cheaper.
I suspect that you're going to end up with more than one heat pump. In zone 5, I'd imagine your design temperature is in the single digits, with a 2500 square foot house in zone 5 I'm going to guess your heating load is around 25k BTU/hr. Something like a Chiltrix CX 34 (see https://www.chiltrix.com/CX34-air-to-water-heat-pump/ ) which has a nominal capacity of 34k BTU/hr, is really going to fall off at your temperatures. At 5F it's rated for 17K at 95F output, which is fine for floors probably but for air handlers you probably want warmer water, and at 113F it's rated for only 14K at 5F. It's also MSRP $4800.
You could get something like a 18k BTU Mitsubishi Hyperheat ducted unit with both compressor and air handler for about $4400, it's going to keep its capacity a lot better at cold temperatures. And it gives you good redundancy.
i'd consider thermally isolating the portion of the slab that will get the heat. insulation under at least that portion of the slab and maybe a thermally broken keyway between the adjacent portion or SS reinforcing across the joint at the thickened portion. that way you'd avoid the warmboard; otherwise i'd get them to do a slab-fold. pouring concrete at different heights is relativlely straighforward for a contractor. worse case is that they pour the other area as a come-back while doing a pour in another area [like patios or porches.]
I'll take a look at the slab fold, I had not heard this term before. We were intending for a monolithic pour, seems cheapest/easiest.
The heat conduction to the non heated area is not a problem as all the concrete is still part of conditioned space. Getting exact numbers for this is not easy math, but pretty easy to set up in Therm. The problem is the steep learning curve of Therm. Roughly speaking, you won't get much heat 2' away from the heated area.
I would add in floor heat for your bathrooms and entrance mudroom.
DC suggestion earlier of a mix of air to air and air to water is definitely worth a look. All hydronic, you are probably looking at around a $15k BOM. I would go further and say anything air to water is probably not worth the ROI, simple resistance heat for these limited areas will be much simpler and won't cost all that much extra.
If you really must have air to water in there, I would look at something like a Sanco2 unit feeding a large dual coil buffer tank. The domestic hot water would run through the coils which would let you use glycol for the Sanco2 and your space heat. This setup is limited in the BTU of space heat you can get but will easily do a couple hundred square feet.
Thanks all! I need to double check my sheet, but our heat load is closer to 15k BTU if I remember right. It's double stud walls, triple pane glass, and targeting under 1.0ach50. At my design temp, the ArcticHeat and Hydrosolar A2W pumps put out enough capacity for the house. Garage would have to drop off or use electric backup.
The reason I want to go A2W is because I can do some of the world myself on the install to help even out cost. I want a ducted fan coil for the rooms of the house, wall mounted fan coil for the great room, and wall mounted fan coil for the garage/shop (for A/C mainly).
The Sanco2 I like, but it only does heat, so I would need two additional heat pumps between house and garage at a minimum.
My suggestion was Sanco2 plus a cold climate heat pump. You'll have to run some ducting already for fresh air, with your low load, the ducting that can carry both fresh air and heat/cool is maybe 1" larger.
With slab on grade, I would say even going for 2x6 or 2x8 sleepers and running the ducts through that space is worth the extra bit of framing and lumber. This would mean standard nail down floor install, no need to glue down, you can now use warmboard for your floor heat and you can even run you plumbing in there.
Your 15k heat loss puts you right about something like a Midea DLCSRBH18AAK/DLFSABH18XBK. Internet price on this is $2500. Nothing hydronic you can come up with will come even close to the cost and efficiency of a modern air to air unit.
That's more or less what I'm intending. Same duct system for ERV and north end of the house heating/cooling. Due to the southern glazing, I think the great room needs it's own mini split of sorts. And the garage as well. And I really want that heated floor... both house and garage. So when you pencil all that out, it becomes 1 A2W vs 2 or 3 A2A plus something for DHW. The numbers start to get much closer, but A2W leaves more I can do myself. And I picked up some of the pricier hydronic bits already, NOS, etc. I did a lot of side by side spread sheets, haha. Including resistance heat for a couple thousand Btus. Resistance in the garage becomes a decent bit pricier, so going that route, I'd have to forfeit the heated shop floor I think. I am intending to well insulate the garage and build carriage doors, but it won't be as good as the house I'm sure.
Are you saying 2x8 on it's side? Or laying flat? I'll look through the detail library on wood floors. I guess, my thought was engineered wood flooring over bare concrete was the best option. Carpet planned in 1st floor bedroom, but not intending to heat that floor. Bathroom and mudroom are on my list.
For ducting, we do have a good bit of space using floor trusses. My cfm and duct calculations say we don't need that much ducting. Worst case room was in the 50cfm range. The floor trusses have a lot of room.
YOu can self install air to air heat pumps. The work is all the same except for the refrigerant, and it should be possible to get a tech to do that. For what HVAC guys want to install a system, you buy a whole set of replacement units instead of having warranty coverage.
High volume name brand units tend to be reasonably priced compared to lower volume air to water.
I self installed 3 mitsubishi units 15 years ago. Tech came in reflared the ends of the stock lines[his choice] soldered on right angle fitting on for convenience, nitrogen pressure tested, vacuum tested and opened the valves. Some years later, one unit ran low and it was a line fitting on the outside unit that needed to be retightened.
Otherwise flawless
I have actually done it, one heat pump and two vehicle AC systems I had to custom build, even machined my own fittings.
With that said, I'm not certified. And for my house, would need that or an HVAC tech, which for 2-3 units is going to add up quick.
This article has a good description of minimizing concrete:
https://www.greenbuildingadvisor.com/article/minimizing-concrete-in-a-slab-on-grade-home
Yeah, we are looking at that. Without a basement, I don't feel comfortable not having a storm room of some sorts, so we have a concrete floor there and then load bearing down the middle of the house. Creates some complications. Not insurmountable, but WarmForm makes a nice and easy solution, and it has some speed advantages too.