PC: pikes building department is wrong, flat out. the method neither corresponds to any accepted heat load methodology I am aware of, nor the algorithm siggy has in modern hydronic heating that I would recommend using (while I agree it too may not be perfect). It underestimates heat load, potentially quite significantly with both the vertical slab insulation question (the loss from which dwarfs all the rest) and the perimeter losses which are significantly higher than to ambient dirt, unless you have successfully insulated the edge/perimeter to minimum that affect, which is what a proper heat load calculation should show either has, or has not be achieved.

Also, slab temperatures are higher if the slab itself is heated, and therefor slab losses are higher than if the slab were not heated as well, so position of your emitter does matter (though it's not the only point of interest).

dmaceld: insulated crawlspaces are different than heated slab on grades. You've got extra dirt helping you out.

Agreed. That's why I emphasized I was referring to slab only. However, after you get about 4' to 8' in from the perimeter both crawl space/basement, and SOG would be pretty much the same, wouldn't they? jk wants to justify, in his mind, going with more than 2" of under slab XPS throughout the whole slab. My contention is that once you're well inside the perimeter, and heat the air only above the slab, you have quickly diminishing returns for XPS thickness greater than 3/4".

For cost/benefit analysis I would be inclined to use siggy's formula for the perimeter, and a plain heat transfer calculation for the slab area that is greater than, say, 8' from the perimeter.

Dmaceld: yes, it arrived, but none of the Options work for me. Chitina. I replied to your regular email, though. Thanks. john

It will be interesting to see what Sag uses for the perimeter. As I mentioned somewhere above, some of the engineer types here are going w/ 4" XPS either 4' or 8' from the perimeter, then 2" elsewhere, as mentioned above. That appeals to me viscerally, but 3/4"?? Man, the numbers may prove me wrong, but that is just plain skinny. Too, I need to find out what our earth temp is; that will make a difference, too. As for the perimeter vertically, sure, plenty there as that is what faces the -50 carp. Of note, I've been clearing trees the last few days, and again found that the ground (under undisturbed snow. leaves, roots) is only frozen down a few inches. That always amazes me, though I don't know the moisture content of the silt, either, and maybe it is so damn dry it just can't freeze. MC of silt and earth temp: gotta dig that info up to help w/ the picture here. Thanks again to all for the input. ON ANOTHER NOTE: I read on an engineering forum that an R for a slab within 0.5 meter (vertically) and 1 meter (horizontally) of the surface is about 2. I am not sure I completely assimilated the scattered info to draw that conclusion, but that is what I THINK I read. Comments? j

Hey guys, I just joined the green building talk and I am very impressed with the knowledge and experience you guys are posting. I have a couple questions about rigid foam insul. under conc. slabs. I work at an engineering firm where most of the ppl arent really understanding the importance of insulation, so I am working on making cad standards that the office can all use. I am wondering how to specify the slab construction. What do I specify for the type of rigid foam insul. to use under, where does the vapour barrier go (under the insul.?) and if I place the rigid foam extruding up along the conc. slab along the perimeter (thermal break) to make it flush with the slab itself, could you still build an interior wall along the inside of the foundation wall right up against it? do you leave a 2 inch gap around the slab to place that insulation vertical? Any advice or expertise would be much appreciated. O and I live in Kamloops B.C. Canada ;)

Cheers

It is all about temperature differential (ODT vs. IDT) with the added factor of slab vs. deep earth ground temperature. Here in MN (46-48F DET) we use 6 mil. poly vapor barrier on the grade, followed by 2" of XPS (see manufacturers of compressed strength). The frost line is the deciding factor for the width of the perimeter insulation and the depth by the difference of outdoor vs. indoor. 2" will stand up to foot traffic and multiple depths make everyone the job site unhappy. The ends of the slab can transfer a lot of heat and should be insulated.

I still design large commercial buildings using perimeter insulation only dependent on facility use, and deep ground temperature. The large the building the lower the design water temperature and thus the temperature difference between ground and slab. This temperature difference is significant, as in my early days of designing, it was brought to my attention by a hill-country resident of NW Texas. His ground temperature is 68F. For a commercial building in Souther Texas with and indoor design temperature of 55F you could have a net gain depending on infiltration and fresh air requirements. Perimeter insulation was nearly overkill for ROI.

The only place you would need more than 2" XPS (the terrible and ubiquitous 'high ground water' notwithstanding) would be in certain mountain areas of the lower 48, in areas of Canada and those where permafrost exists. Fairbanks would be one such location if I'm not mistaken.


BTW most heat load software designed for radiant floor heating take into account all these slab factors and "degree days" have virtually nothing to do with heat load analysis. The heat load is determined (following long established ASHRAE standards) by calculating the the heat loss of a structure under design conditions. Again, these have been established using weather data for the area in question and typically represent the averaged coldest 4 days of the year for a given location.

Badger: Interesting info again. Thanks. I can see where two layers of XPS might be awkward to work over. Maybe if a person glued layers together it would not be so bad, IF it holds. I'll keep that in mind, and I think SBS here carries 3" sheets for the perimeter. Yes, we do have permafrost in many areas, but thankfully not right near me. However, the ground is still cold; only one source on that, and he says "32 degrees, all year, down to 10 feet". It is interesting, too, that degree days are not directly used in calcs for heat loss, as I am only familiar w/ doing annual calcs using Q = UA.... I'll have to explore that some more, just for curiosity. In the meantime, I've adjusted my design to be much less "gross overkill" based on the info I am gleaning from reading here and elsewhere. Thanks to all; really helps. john

Jklingel, You steered me to hearth.com on healthyheating.com six weeks back. I have become an avid reader there, and I am close to solving my wood heat solar assist problem.
The only heat loss model I've found that allows some optimizing of slabs on grade is this one: http://builditsolar.com/References/Calculators/HeatLoss/HeatLoss.htm It is pretty rudimentary. Maybe the engineers here can gauge its reliability
At the risk of boasting, I think I have an elegant slab on grade solution. I'm doing a three-foot perimeter wall in ICF with an eight-inch cavity. I am floating the slab inside it, and buiilding AAC walls on top of it. (AAC is a low-density concrete block with a R value of about 1 per inch.) A 12-inch AAC block caps the ICF perimeter wall in both the literal and figurative sense, covering the foam on both sides, and completely insulating foundation concrete from the house's interior except at two exterior doors.
Isolating perimeter walls is key. Heat loss underground happens by conduction, or at right angles. Thus the edges of the footings lose twice as much heat as the bottom of the slab, and the corners of the footings lose three times as much. That's why builditsolar.com's model looks only at perimeter walls. Much of the heat loss, even in a slab on grade, happens above ground, so you can't ignore sill plates.
You probably still need insulation under the slab if you have radiant heat in it, depending on your solar design and prevailing weather. Back when, I planned to use the dirt under my slab as extra tons of heatsink. Turns out btus aren't as easy to come by in a gray Pa. winter as I figured, and what I need is a more manageable response time.

Todd: Glad you like those forums, too. There is a huge community of people willing to share/help in the world, contrary to what the newspaper may lead us to believe. The builditsolar heat loss program does use perimeter, but it still uses Q = UA(delta T) for the slab; I dug into the programming. If you use Siegenthaler's formula, apparently more accurate as it was modeled from ACCA (I think he said) data, you get a smaller heat loss. That was one of the big things that caused me to rethink huge XPS under the whole slab, and move some of it to the perimeter. Obviously, Q = U.... is not appropriate for a slab like it is for a wall, but I never knew what else to use. Now I do. Good luck w/ your project. j