The gravel is still in contact with the ground so damp and are still present.
Don't skip it, it ain't something like a countertop that you can unscrew and change.
Down the line you'd like to add radiant what will you do?

USE the XPS as planned.  In the long run it is worth it. concrete is about R= 0.08/inch  It is the difference between R 0.4 and R 10 in the whole floor. 

If you keep the house at 70 deg and the average groud temp is 50 deg the delta is 20.

Lets say the basement is 1000 ft^2  180 Days of heating per year

Option 1 no insulation  R=0.4 U= 2.5 BTU/(ft^2*deg* hour)

2.5 btu/ft^2*deg *Hour  X (1000 ft^2 area)* (20 degree temp difference)* (180Days*24Hours) =  216X10^6 BTU year

Option 2 R 10.4 insulation U = 0.096 BTU/(ft^2*deg* hour)

0.096 btu/ft^2*deg *Hour X (1000 ft^2 area)* (20 degree temp difference)* (180Days*24Hours) = 8.3X10^6 BTU year

Now this does not take into account a number of things (earth acts as an insulator to some extent) so the numbers are not  as bad as this calcuation depicts, but the Botom line is that  Insulating the slab can easily save you  50 million BTU's per year

50*10^6 BTU's /135,000 btu/gallon of fuel oil*90% efficiency =400 gallons of oil/year. You will pay of the insulation quite quickly. 

Eric

Eric, your calculation wildly overestimates underground heat loss, and xps under the slab will not pay for itself quickly. Put your numbers in this Canadian basement energy model and see for yourself: http://canmetenergy-canmetenergie.nrcan-rncan.gc.ca/eng/software_tools/basecalc.html

Jerkylips, you definitely want xps between the slab and the basement wall and the slab and the footers. Consider extending the xps under the slab for three feet or so around the perimeter.

If you add radiant later,you can add full coverage then as well.

Direct quote from building sceinces corp
“Floors Basement floor slabs are best insulated underneath with rigid insulation: both extruded or expanded polystyrene have been widely used with success. Although the energy savings of sub-slab insulation are not as significant as basement wall insulation, such insulations do offer a significant improvement in comfort and moisture damage resistance (including against summertime condensation).”

Todd
So you are suggesting he insulate vertically down 32 inches to the top of the footer and then around the footer + 3 ft of perimeter insulation? Seems to me that works out to around 7 square feet per linear feet of building perimeter. Why not just insulate under the whole slab and be done with it? For a 25X40 House, which was my example, you would use 130 linear ft X7= 910 ft^2 of insulation and I would use 1000^2. If you were convinced that the subsoil was always dry and low conductivity I might buy your argument if he wasn’t using radiant heat or living in it. This is not an unheated basement, this is living space and bedrooms for his kids if I remember correctly. It also contains a walkout. In the summer you are going to have a colder condensing surface, in the winter it is a large heat sink.
 Most studies I have looked at have an uninsulated slab being 6-30% of the heat loss of the structure There are a number of technical papers that model subslab heat loss. They get very different results based on soil composition and moisture levels. Today, probably the most stringent modeling program is the passive house one. From everything I have read most of the cold climate passive houses have between 5 and 16” of eps under the slab. On the high end of the spectrum for slab on grade, on the low end of the spectrum for basement slab
In my own house, I did both. I used ICF walls and 2” xps under the slab. I can walk comfortably on the floor in bare feet in the winter, in the summer I get to open the doors and windows in the basement without getting condensation on the floor and a musty smell.

I would agree that I used a very simple, 1 dimentional model to model the heat flow. I would agree that reality is quite different from my first aproximation.  In part that is why I derated my initial calc by a factor of 4.    I  can't seem the run the basecalc program, it keeps crashing.  I tried installing it a few times.  I  think the methodology is probably reasonable that they use for it.  

Why don't you run the numbers both ways as follows for my own house:

Basement 14X10 m  walls normal height, the insulation is ICF so 62 mm eps outside and 62 mm inside.  Basement walls are  on average 4 ft depth underground, one wall hase complete exposure, the slab is nomal thickness 100 mm, moist high conductivity soil, water table ~5.5 ft during heating season 3 ft durring spring.   6500 HDD so I guess I would pick detroit as my example city.
compare  50 mm XPS to no XPS under the slab.

That should give the comparison for where I live.  Makes for some good food for thought.

Cheers,
Eric

 PS.  I am quite happy I went for a fully insulated slab.

hey guys, thanks for the responses.

a couple points of clarification -

basement is unfinished now, not actively heated. Down the road we plan to finish it & zone the forced air heat, but no radiant. I was considering tubing for radiant to use solar hot water to heat, but we had to draw the line somewhere on budget & that got nixed.

The basement walls are going to be insulated w/2" of rigid foam.

Once excavation was done, we found that with the slope of the lot they poured less full height basement wall, & we're getting more exposed on the sides. That's a good thing, since we can put in windows now & count those rooms as bedrooms. I'd say that more than 1/2 of the basement is going to be framed wall, with the same insuation as upstairs - 1" spray foam, r19 batts, & 1" of foam sheathing.

All this discussion is probably academic at this point. We bought our lot about 3 years ago, & it appraised out lower than we were expecting. Because of that, we have to come up with more money at closing, & are needing to cut a few other things. I don't think I'm going to be able to swing the extra $1800 to insulate the slab. We do have a lot of windows on the south side of the basement, so I'm thinking/hoping that the solar will heat the slab during the day & temper at least some of the heat loss through the slab.

Be a little bit carefull about large basement windows. Finshed basements and bedrooms may add quite a bit to your tax assessment. Worth finding out before you do it.

personaly I would cut other things before I cut the insulation, like leaving a bathroom unfinished. You only have one chance to insulate the slab.
Every one that builds ends up making hard choices about what to do and what not to do!
Good Luck,

Eric

Posted By eric anderson on 24 Mar 2010 01:28 PM
Be a little bit carefull about large basement windows. Finshed basements and bedrooms may add quite a bit to your tax assessment. Worth finding out before you do it.

personaly I would cut other things before I cut the insulation, like leaving a bathroom unfinished. You only have one chance to insulate the slab.
Every one that builds ends up making hard choices about what to do and what not to do!
Good Luck,

Eric

I hear ya..  The problem is, we've already had to cut a fair amount of other "non-essentials".  The mechanicals & insulation were the last resort, but we've cut enough things that there isn't much left.  There are a few things that we just refuse to sacrafice, like our hardwood floors. 

And as far as basement goes, it won't be finished for at least a few years.  All we're doing now is putting in the windows & roughing in for the bathroom so that shouldn't affect our tax bill.  Very good point, though...

For a non-structural slab over 32" of gravel methinks you could get a lot out of either going EPS instead of XPS, or 1" XPS instead of 2" at a substantial cost-savings. Well drained gravel itself has an insulation value of ~R4-R5 in that depth in a cold-side-down configuration (which it is.) A couple inches of EPS would bring it up to ~R13, reducing the XPS to 1" would still yield ~R9-R10, which is about what you'd get with 2" of XPS on top of hard-packed clay instead of gravel. In the quad-cities groundwater temps run ~ 50-54F, so that's probably about the temp of your sub-slab soil. You don't need a lot, but you need SOMETHING more than gravel to minimize that heat loss.

A separate vapor retarder (6mil or thicker poly sheeting) between the foam and slab is a good idea if you're going with EPS, or XPS less than 2" thick, for ground-gas control (radon, water vapor, etc), since 1" XPS is semi permeable, and EPS at 2" is highly permeable. XPS is semi-IMpermeable at 2"+ and is usually adequate on it's own as a ground vapor retarder as long as you seal the seams. (You may have been planning on a vapor retarder sheet anyway, since it's much better than 2" XPS on it's own.)

Our basement is also "extra" at this point but you can really tell where the footings are temperature wise. We have about a foot of gravel and it has a "French drain" out to the end of the lot so it stays pretty dry. We used 2 inch under the slab, but since your high and dry you could probably go 1 inch, but I would put something under there. For us the 2 inch was a big help in putting the pex down via staples.

Funny story on the French drain, when the inspector came over he ran a hose and dropped it in the sump pit and turned it on, waited a bit, nothing, walked around the house and came back to it, still no pump activation. He finally asked me what was up because he couldn't pass the house without the sump pump running and verifying it worked and drained correctly. I told him we had a French drain and he said "Ahhhhhhh". Then I manually turned the pump on to drain the lower part of the sump pit and he was happy. I am not sure how much water he ran down there before he came and found me

At least go with 1" of foam to provide a thermal break.  Not knowing where you live, etc, here, with a daylight basement as you describe, the foam is required by code. 

BTW - poly plastic under the slab is essential.

Sorry for the delay, Eric. Gotta build at full speed between Nor'easters dumping 4 inches of rain ...

First, you measure thermal breaks in inches rather than feet. A 4 inch strip of foam would separate a four-inch slab from the foundation walls, and you'd need another 6-10 inch strip to lift the slab up off the footers. The breaks are important because heat loss by conduction happens at right angles; footer edges lose heat twice as fast as a slab bottom, and footer corners lose heat three times faster. Heat loss is also much greater at the perimeter of the slab than the middle.

I ran your numbers a couple ways: With Detroit climate and soil, your basement would lose 15.1 gigajoules of energy in the course of a year. Skipping xps under the slab raises the loss to 17.5 gigajoules. Insulating the first meter of the slab's perimeter (leaving the middle uninsulated) drops the loss to 16.5 GJ. A GJ is roughly one MCF of natural gas. At $12/MCF, full coverage is not saving you big money.

If you poured the slab inside the ICF foam, you have one of the two breaks and your heat loss is then 14.3 GJ. If you butted the XPS directly to the ICF foam, you have both breaks and the loss is 13.7 GJ.

Radiant heat changes the picture, as do walkouts, soil conditions and building codes.

Thanks for running the numbers. In my case the insulation under the slab runs from icf wall to wall so there is a complete thermal break. So I assume I work out to around 13.7 GJ or 12.9 M BTU (which makes more sense to me in terms of energy units).
 
I guess we look at this differently. My first thought based on your numbers was that I should have used 6-10 inches of EPS under the slab. I would consider this much heat loss through the slab quite unacceptable. But based on that programs calculated heat loss, the difference between insulated (2 inches XPS and non insulated is only around 1/3 or 5MBTU. Since neither radiation nor convection are likely to be a big factor, this implies that the ground is effectively R20 from a conduction point of view. I have a hard time buying that number, but that does not mean it is not correct . In my calcs I was estimating that the slab heat loss was about 8Million BTU/ year and you are calculating it as~12 million BTU. My assumption was that since the ICF walls continued at least 48” below grade the temp under the slab would be a uniform 50°, but this might not be accurate.
 
I realize in an average American house, that runs ~10 btu/ft^2/ heating degree day, the basement slab loss can effectively be ignored.
 
My heat loss for the winter for the whole house is around 18 Million BTU this year, 155 gallons of propane(90% efficiency) and ~3/4 cord of wood (65% efficiency). I don’t know how much of the propane usage is DHW or cooking so actual heating BTU’s are a bit less. I had initially calculated that the house would use around 10 Million BTU/year so I am off a good bit in calculations. One area I know is problematic is the use of a woodstove. When I run the stove, the temp in the house ends up much higher then I like, even though the stove is quite small. The higher delta T accounts for some of the heat loss. The other thing I never really realized was how much the contribution all the various exhaust vents contributed. All that being said the slab contributes a significant heat load in a well built house and should be dealt with accordingly.
 http://www.passivehouse.us/passiveH...ehouse.pdf  These are the houses we should be building.
 
http://www.treehugger.com/vinyl-mcmansion.jpg This is what we are building.
 

Cheers,
Eric

> under the slab would be a uniform 50°,

The soil right under an uninsulated slab may start out at 50F, but it quickly rises to something in between the deeper soil temperature and the slab temperature. Best to use something like "BaseCalc" for calculating losses.

I didn't give you comparables on thermal breaks, Eric. The 17.5 GJ for no XPS coverage assumed that there were no breaks. So you need to compare the 13.7 GJ figure with 16.4 GJ. In other words, with equivalent thermal breaks and no other xps under the slab, your heat loss would be about 20 percent higher. Putting a meter of xps around the perimeter would reduce it to 15.8 GJ, or 15 percent higher. The difference there of 2.1 GJ, or 2million btus, is a truer picture of how little heat is lost through the middle of slab, as opposed to foundation walls and slab perimeter. Most NG users would look at the $25/yr savings and ask why bother.

While soil is not R20, Basecalc estimates that there is 10 feet of it in Detroit before reaching the water table. Conduction means heat loss from the middle of the slab goes straight down until it isn't going much of anywhere any more.

Posted By toddm on 06 Apr 2010 05:12 PM
I didn't give you comparables on thermal breaks, Eric. The 17.5 GJ for no XPS coverage assumed that there were no breaks. So you need to compare the 13.7 GJ figure with 16.4 GJ. In other words, with equivalent thermal breaks and no other xps under the slab, your heat loss would be about 20 percent higher. Putting a meter of xps around the perimeter would reduce it to 15.8 GJ, or 15 percent higher. The difference there of 2.1 GJ, or 2million btus, is a truer picture of how little heat is lost through the middle of slab, as opposed to foundation walls and slab perimeter. Most NG users would look at the $25/yr savings and ask why bother.

While soil is not R20, Basecalc estimates that there is 10 feet of it in Detroit before reaching the water table. Conduction means heat loss from the middle of the slab goes straight down until it isn't going much of anywhere any more.

A highly dubious datapoint to generalize about, eh? 

Is that average water table ~10' below the average surface soil?

The site-specifics could be (and probably will be) off by quite a bit from that estimated by the tool, in either direction, and I'd be loathe to make a decision based on a such a generalization. (It might not be hard to discover what the TRUE water table level is at that site.)

You can (and should) enter whatever water table depth is appropriate into BaseCalc.

Your county soil conservation district can help you customize basecalc. Local well diggers know what's down there, too, as do septic and soil engineers. My point in citing basecalc's Detroit default was simply to suggest that per-inch measurement, like much of conventional insulation wisdom, doesn't necessarily translate.

Interesting discussion all around.
 I have not been to successful trying to get Bascalc to run on my computer and asked Todd to run the numbers for my situation.
 I picked Detroit for similar heating degree days as Connecticut not soil conditions.
Where I live no one builds on land with a water table 10 ft down. In fact water runs continuously out of my footing drains ~ 6 months a year. The first thing you do here before building is dig a batch of deep test pits and look for the mottling level and a place you can legally put a septic system, ie ground water at least 2 ft below the surface, then figure out where the house goes.
 I was not suggesting that the soil was R20 per inch, just R20 total in the calculation for the math to work. My point was not that the calculation from basecalc was incorrect, just that it does not make sense intuitively.
 Look at it this way, convection is not an issue, radiation should not be a factor so you are left with conduction as the mechanism for heat loss. If we ignore the condition with no insulation under the slab and simply focus on the condition with a fully isolated, insulated slab.  If we take a basement slab, isolated from the ground by 2” XPS Underslab temp a uniform 50°F, Temp in the basement 70 deg.  Assume a 6 month continuous heating season this should account for the worst case scenario. IF we assume an R value of the XPS of 10, you get U= 0.1 btu/ft^2*degF*hour.
 0.1 btu/ft^2*deg *Hour X (1000 ft^2 area)* (20 degree temp difference)* (180Days*24Hours) = 8.6X10^6 BTU year.
IF Bascecalc is correct in determining a number that is 13 X10^6 btu/year, one of my assumptions must be incorrect. Either XPS has a higher conductivity then predicted, the temp under part of the slab is less then 50 deg, or there are heat losses due to radiation or convection. Including the sides of the slab increases surface area by 4% so if you include that you are at 8.9X 10^6 btu/year. Where does the other 4 Million btu’s go?

That is my question.
cheers,
Eric

The other 4 MBTU goes nowhere, soil being a relatively poor conductor. Delta T shrinks by the day until the soil in contact with the slab is so close to 70 degrees that it barely matters if you used xps or not. Your formula assumes that lost heat zips into the ether; it doesn't work for what it is effectively a closed system. (more or less, depending on site specifics.)
This is not a lonely view held north of the border. Rescheck ignores underslab insulation as well, except in walkout situations. Even for slabs on grade, you reduce UA by sinking perimeter walls deeper and insulating them better.