The R value of dry sand is about R0.6 per inch of depth. A mere 3" layer will add up to about a whopping R2 (or almost, R2). To add up to the same R-value as 2" of XPS (R10), the simple arithmetic says you need about 16-17" of DRY sand. (Wet sand is much more conductive <==> lousy insulator.)

It's more complicated than that though- lots of other factors enter in:

Sand is highly permeable and not at all air-tight. If you're using the sand as an insulator you need a very good vapor barrier under your sand layer, and good soil drainage below the sand. You also have to be sure that the floor coverings above the slab don't create a vapor trap to allow any spillage or moisture seepage finding it's way into the sand to dry toward the interior of the building. (Don't hose off the slab to clean it every day, or you can kiss it's insulative value goodbye.)

Sand does have a significant amount of thermal mass (I'm gonna guesstimate it has a specific heat about 1/2-2/3 that of concrete per unit volume, or about 40-50BTU/ft^3/degree-F. At 16" depth that's something like 60BTU/ft^2/degree-F, similar to that of a 8" thick slab (but not all entirely inside the insulation, since it IS the insulation.)

Both of these affect the actual R value of the sand layer, and it may vary considerably with temperature differences between the sub-soil and the slab, and the R-value is not linear with depth. (You may only need a foot, not 16" to achive R-parity with 2" XPS, I don't know for sure what the function looks like.)

But if you're running radiant in the slab, bear in mind that you need to heat UP a substantial portion of that sand- the lag time on the system may extend beyond 12 hours. (I'm pretty sure some passive-solar designers have modeled the lag time- can't point to any data/studies off hand though.) Using some sand over XPS to add inexpensive mass to the system may not be a bad idea in a very well insulated (or superinsulated) building.

In VERY sandy soil with excellent drainage on all sides you can probably get away with insulating the slab perimeter 4' in and calling it a day. I've seen/read-of installlations where it was R10 on the perimeter, R5 in the center- that's probably still cost effective most of the time, and the response time of the radiant slab will be much more reasonable (yet still quite slow relative to low-mass radiation.) If the shop building is intermittent-use only, insulating the floor and going with radiant ceiling, , low-temp baseboard, or radiant panels might be a cheaper/faster reacting, more suitable for using temperature setbacks during idle times to conserve fuel.

The bigger question is why would you need a heat sink? An uninsulated slab will stay warm well into the fall, saving energy, but it will also stay cold well into the spring, offsetting those savings. Unless you expect a substantial cooling burden in the summer, or there is a seasonal shift in energy costs, or passive solar works really well in northern Idaho, you are no further ahead.
The Canadian govt has done lots of work on basements, (but alas not on slabs on grade.) The official verdict there was insulation under the slab is optional, except in the case of radiant heat. Gotta agree myself. Absent an argument for storage, I'd want the heat I put into the slab to go up rather than down.

My thought on why a big heat sink would be good is that it would let the building "coast" through if the heat source was off for a few days. I should mention it will be heated with an outdoor wood fired furnace, so if I wanted to take off for a weekend and not feed the fire it shouldn't cool down much. The building will be in use typically at least 40 hours a week.

It's not my intention to try and cheap out on the insulation and didn't really expect to be able to replace foam with sand but the idea of improving the insulation and heat storage with something that's nearly free definitely has my attention. The idea of a couple inches on top of the foam is what I'm considering most. It seemd to me like if the pex is up in the slab then the slab would heat up first, then very slowly the sand. So, if the sand improves the heat storage, and slows the downward movement of the heat before it even gets to the 2" foam then what, if anything would be the downside?

No downside really, except for a slow warmup after a period of no heat. You'll need heat loss calculations to guesstimate how far from home you can get, but concrete weighs about 145 pounds a cubic foot and has a specific heat (ability to store it) of about .2. Sand weighs about 100 pounds a cubic foot and has a specific heat of about .22. (Easier to warm than concrete.) So, with 4 inches of each under your shop, your floor will warm or cool at the rate of about 80,000 btu/degree/hr.

sounds like you are almost there.
no sand, great 2 inch rigid insulation! the real deal, now you can staple your pipe to the foam, easy/fast. lay your wire mat on top of pipe if required (7'x20' flat road mat watch for sharp edges) .
Sand is going to slow everything. A 4, 5 or 6 inch slab will hold a lot of heat mass.
thats what I would suggest.
Dan

The vapor barrier goes down first, sand for leveling, 4' of 2" XPS at the perimeter minimum, flat wire for tying in the middle. If you plan to set and forget you system this will work quite well. Response time should be addressed by outdoor reset, built-in to all ModCon boilers.

Professional design help, priceless.

Read this article about sand over vapor barriers before doing this. This is a very informative article about just this subject. Talks about how the ACI (American Concrete Institute) changed it's recommendations for slabs in 1996.

http://www.buildingscience.com/documents/insights/bsi-003-concrete-floor-problems?full_view=1

Tom Pittsley
[email protected]
www.eebt.org

Tom,
Great article, no sand contact with the slab. The concept of placing pipe in the slab for additional heat mass just got compounded. sand insulates and is a sponge as well.
Morgan, where is the insulation? not even barrier. I would think that using a mod con to heat the earth under this huge slab would cost a fortune. Every uninsulated slab short of the perimeter i have see( though few) seems to cost a lot to heat. I could see this if the fellow was using a free fuel boiler like wood/solar. I am just not convinced. I think that even a modest amount of slab insulation will make a large difference in op cost. Have you had good results with perimeter insulation only systems like this?
Dan

Sounds like

concrete
plastic
sand
foam

is what you want.   You will get more thermal mass but I agree that the insulation from sand isn't worth much.

The extra thermal mass might be useful if you have low nighttime electric rates, solar heat or wide swings in daily temperature.

Assuming fully insulated slab,
Standard in applications we have done would be
Concrete
2 inch Rigid foam / pex stapled to foam
barrier / 6 mill plastic
sand, pea gravel, compacted earth, crushed rock substrate

If you want more mass, add more concrete. I can tell you this , you can have to much mass. The more mass the more storage but at some point it becomes moot. In a house I just finished we used in order: crushed stone compacted, 10ml plastic, 3" of geofoam, 6" concrete. The slab is being supplied with 95F water from a geothermal heat pump. It takes almost 6 hours of run time to move the slab up one degree F. Any location uninsulated will be a heat sink and not what you want, you do not want to warm the earth below the slab, it is a lossing battle. I debated weather to put foam then plastic or plastic then foam but opted for platic then foam to minimize the potential for the plastic to be punctured. I think either way would work, just no sand on top of plastic.

These two videos show how I did it on my last project.

http://www.youtube.com/watch?v=J7ICP024ogw&feature=channel_page part#1

http://www.youtube.com/watch?v=JkXceHPU7zw&feature=channel_page part #2


Hey Dan from Blueridge company,
Remember me? Tom Pittsley, you donated some of the radiant equipment we used for the MIT solar decathlon project. Want to get involved with the next one? The Boston Architectural College has teamed up with Tufts university for this years 2009 solar decathlon competition and need sponsors. I am acting as advisor for the team and they need some help. Your equipment showed very nicely in 2007 and was one of the things people seemed to be most interested in. It seems people are starting to understand that radiant heat isn't really a luxury anymore but a means to save energy and be efficient, comfort being the bonus.
Thanks
Tom Pittsley
[email protected]
www.eebt.org

I agree that a high mass radiant surface can cause problems - it's an uncontrollable heater.   Say it's 40F at night so the slab heats up, then the sun comes out and it's 70F outside.  You can't turn the slab off and the house gets too warm.

There is also the issue of not being occupied.  In a low mass house or office, you can turn the heat off when you aren't there and save energy.

Outdoor reset will control any well-insulated building. The vapor barrier protects the insulation. Perimeter insulation does the bulk of energy saving as typical earth acts as an insulator. There is much debate on this issue but many commercial properties are still specified w/o sub-slab insulation. It does not pay for itself on the large scale.

As for heating up the earth. One could say the same about walls and ceilings. It is a matter of time, temperature and square feet. At some point the soil stops moving energy.

What is the vapor barrier protecting the foam insulation from?

Outdoor reset doesn't solve the slow reaction time problem caused by thermal mass - it only gives a somewhat quicker indication of the heat loss changing.

A really smart thermostat would be able to predict the future heat loss - by using previous day patterns, sunshine, weather reports, etc and would then stretch the normal comfort limits when there will be changing heat loads or costs. For example, say I normally prefer 70F +- 1 degree. But *if* there is a good reason, I can live with 65-75F. If it is a 10F cold and cloudy winter day, might as well just heat exactly to 70F. But if is 45F at night and expected to be 70F and sunny later in the day, then let the temperature drop to 65F at night and take advantage of that free heat that will be available later.

Vapor.

Reaction time has to do with heat loss and gain. The better the building the better the performance. Internal gains notwithstanding, I have designed many slab systems using T87s.

Your strategy may work, but dollars are small. Some modcon boilers have reset and setback in one. Tekmar can do practicall anything with a building but you have to pay for controls also.

I suggest you spend the money now and invest in P2000 insulation. not only is it a vapor barrier but it is R-4 per inch and is the best radient barrier on the market today. The products site is http://p2insulation.com/flashfiles/index.html Your money will comeback to you tenfold with energy savings. sand is cheaper on the short run but will cost more on the long term. You should always be thinking on the long-term when building a home.

Radiant barriers under slabs are a waste of time.

Agreed.  A radiant barrier without some airspace next to it is worthless.

EPS foam is only way to insulate under slab.