Could use some help from the brains on this forum.  I've noticed that most of the time (in my limited experience) when a radiant slab is poured over earth that they put 2" of the pink down (Formula 250 ?).  I believe this is rated at 10 r's to the inch.  You can pic different #'s but let's just say the slab (heated) is 100 F and the earth is 50 F.  For 50 degrees of difference R-10 seems to be the standard.  When we move to the structure above we want a target temp of 70 F and maybe it gets to 0 F (outside) but a fair average might be 20 F.  With the same temp difference (50F) the walls and ceiling get a recommended R-19 and R-38 (or more).  Why the disparity?  With EPS at $1.00/sq./2" is a 2nd 2"sheet a bad investment? I know it's a diminishing returns and we have some other factors such as air infiltration but why 4X for a ceiling over a heated floor.  Looking for some learning.  Not my field so feel free to talk down.

Best,

Flathead

dirt does in MOST cases add some R-value.

also slabs may often not be 100 degrees, especially not as an average temp... more like 75-80 degrees.

also you don't have thermal bridging or air leakage.

but you raise a good point: I personally do not think most of our load calculators are particularly accurate for slab insulation. Passivehouse people put in a foot or more in some cases: that is overkill, undubitably, but their calculators show influences we don't typically consider, like the depth and speed of the water table.

Personally, I would do 3 or 4 inches standard.

For slab-on-grade, you use the design outdoor temp for sizing the insulation (not the earth temp). This is because the majority of heat loss is side-ways thru the “exposed” slab perimeter (not downward into the earth). “Exposed” means any perimeter exposed to any surface (stem wall, earth above frost line, etc) that is exposed to design outdoor temp. This is why you use slab-on-grade exposed floor PERIMETER (and not the floor AREA) when doing the building heat loss analysis. So for unheated slab-on-grade Ti=70 and To=0 you get 0.125(Ti-To) or 8.75 MINIMUM R-value. For 100 deg F heated slab-on-grade you get 0.125 (100-0) or 12.5 MINIMUM R-value.

where is your 0.125 factor coming from? If the number is not determined by code, seems like that is a variable that should account for fuel costs etc.

The equation and 0.125 is Siegenthaler's MINIMUM R-value recommendation for slab-on-grade floors. Code minimums are often less than Sieg's recommendation. When limited by insulation economics, we size the insulation so that the downward/side-ways heat loss for the heated slab is just less than 10% of the design upward heat gain provided by the hydronic heating.

Actually, my radiant specific heat load software considers both the edge and the field when calculating heat loads as both present a load. If I had to pick one, I too would emphasize the perimeter.

as long as he's slab on grade, perimeter is right, though I strongly dislike the insinuation inherent in that which says underslab insulation is not important... of course, it is. our calculators, including siggy's, are limited and not really indicative of the reality of heat loss to dirt.

as one limited example, my building site is solid ledge. rock conducts heat away a lot better than packed dirt or sand. I've see wet ground projects that failed to heat in the spring because moisture wicked heat away so effectively under uninsulated slabs. no calculator will capture that.

That's why in most cases I recommend exceeding code minimum (R10 typical) slab insulation and doing 3 or 4 inches underslab standard. If you are a slab on grade, then you have the ADDITIONAL consideration of the perimeter insulation. which should likewise be increased in most cases above a 2" minimum, in my humble opinion. you get a lot of pushback from builders though when you say you want 4" of edge insulation.

The practical reality is that 2" in the field will suffice where deep ground average temperatures are well documented and the load is easily calculated. Beyond that, the 2" XPS will stand up to foot and tool and labor for 2" is the same as one inch. Dropping more than 2" in the field of the slab (inside the frost line) is over-kill south of the Canadian border.

We are in agreement on the perimeter. We get complaints about the high cost of heating radiant slabs every week and the vast majority either forgot or did not do a good job of insulating the edge of the slab. They do melt snow around the doors though.

there is very little practical reality that meets your definition. unless you know the ground composition you are building on, and the water table behaviour (which can change), you cannot "easily calculate" a load. our standard calculators still use factors that include some assumption for ground conductivity that is most decidedly not universal.

Perimeter edge insulation is very important for both unheated and heated slab-on-grade because the side-ways perimeter heat loss can get very large. Underslab insulation is important for heated slab-on-grade because the downward underslab heat loss can get large. Underslab insulation becomes less important for unheated slab-on-grade in mild climates and builders will often only insulate the perimeter and first couple of feet underneath the perimeter area of the slab…and I suspect there are still some builders who won’t insulate underneath the slab at all...and that may even be appropriate in some locations/climate.

One really needs to run the numbers with the proposed insulation to accurately quantify the slab-on-grade heat loss and then decide how much insulation should be used. The problem with slab-on-grade soil heat transfer analysis has less to do with the analysis then with accurately knowing the actual conductivity of the soil where you are building, which can vary significantly depending on the actual soil composition and actual amount of water contained in the soil. One can either test the soil conductivity at various times of the year or one can assume a conservative soil thermal conductivity and then run the numbers. Bottom line, you need to know the insulation MINIMUMS and you need to run the numbers to determine how much additional insulation and how far you will be deviating from these MINIMUMS in order to accomplish your design objective. As a MINIMUM when insulation economics are a design constraint (and economics usually are a constraint for most customers), we design the floor heat loss to be just under 10% of the design upward heat gain.

Might want to look at CreteHeat, 3 inches will give you R 14 no thermal breaks/moisture and no off gassing so R value will remain stable.

OK thanx fellas. Mission accomplished. Here's what I think I know. I now realized how important the perimeter insulation is. Probably explains why Morton buildings 15 years ago (in the Northeast) only insulated the perimeter 2' and the rest was on bare ground. If economics allow, I would consider more than 2" of xps but could consider beefing up the outer perimeter 2 or 4 feet as a compromise. Thanx for John's #'s. I'll file that one away. Just finished some seminars in Dec. with him and have a high regard for his knowledge and teaching ability. I see how important knowledge of the soils is and will consider that it's condition now may not be what it will be in April. I seem to remember the comment of one notable on the forum that, "Insulation of soils is a messy subject". Thanx for getting me into the traveling lane.

Br,

Flathead

FWIW: Since XPS is blown with HFC134a with a greenhouse gas potential ~1400x that of CO2, and EPS is blown with pentane at a paltry 7x CO2, when going higher-R on slab insulation it's far greener to bulk it out with EPS rather than XPS, since the lifetime greenhouse gas potential of the XPS would far exceed the CO2 offset by the energy use reductions after the first inch, in a sub-slab app.

Radiant builders/designer may prefer XPS due to it's better staple-retention when tacking the tubing directly to the foam, but if you're going more than one layer (recommended in all cases, to limit thermally bridging gaps from developing over time), making the top layer an inch of XPS, and the rest Type-II or or Type-IX (not really necessary for a residential slab) would at least beat break-even on lifecycle greenhouse emissions in most cases. If local codes allow it, even low density Type-I EPS has adequate compressive strength under a 4" non load-bearing slab, and can often be had for VERY cheap on an $/R basis from roofing-foam reclaiming/recycling outfits (eg. insulationdepot.com) . When using reclaimed roofing foam under a slab it has to be EPS or XPS , and NOT polyiso, since the latter can wick & retain moisture. Iso is OK for insulating the interior faces of basement walls though. Type-II EPS is also used in roofing applications and sometimes (rarely) so ist Type-IX, but you have to be specific if local code (or the mechanical load) demands something denser than Type-I.

About CreteTherm or CreteHeat.....
It's more expensive than XPS and one of them says you still need plastic underneath anyway. So on a 1500 sf project crete heat would be about $1000 extra which translates to a good amount of labor. Nice system though.

I don't have the foam stapler. How do you recommend attaching the pex to the insulation? I would probably put down 2 layers of 1" xps if I don't use crete heat or crete therm.

I think I usually expect it to be a 0.50/sq ft upcharge typically but your mileage may vary based on your location. If you drop a vapor barrier and wire mesh, it isn't that much of an upcharge. if you don't, it's still MUCH faster. Figure how much time there is in vapor barrier install, plus foam and taping, plus mesh, plus tying. that vs snapping together panels and walking in the pipe as fast as you can unroll it.... it's significantly different.

it's definitely not for everyone... if you have time but not money, it wouldn't make sense. but it has a place.

I think crete therm suggests 6 mil poly under their system. Crete heat claims it is a vapor barrier so no poly needed. Would you use poly?

I think I'll be contacting you tomorrow about a design consultation.

I'm not a moisture expert. if I had a really wet site I probably would for actual water permeation issues rather than vapor. in typical conditions, no, I wouldn't.