I am looking for data on what R values would be good for my new dream house. Opinions are great too but I would really like more real data based knowledge. How do you know how to balance the R-values in each area? How can you simulate the radiation/conduction/convection to get the optimal ratios? Are roofs usually higher R-value just due to the internal convection? Why don’t people insulate more under the slab or on the footing/foundation where we have conduction to the outside world? Ever grab a steel pole in the winter with your bare hand? It sucks more heat out of your hand than the outside air does. Based on reading stuff from this and other websites, my first rough approximation is R-10 below slab, with a thermal break of r-5 to the footers which are R-20, foundation R-30, walls R-40, roof R-50. How do I know if I have them balanced right? I am only guessing so far. So the house will be on a sloped site in Manistee Michigan, 3 sides with a 9 foot poured foundation, and one open with a walk out basement, Guestimate size is 20x35 I think I prefer to use SIPS above grade, for the walls and roof as I have some experience with them. I also prefer EPS foam for the footers/foundation.. Thinking radiant floor heating (so maybe I need R-20 or better under the slab) with a condensing tank-type hot water heater. I have read a lot of the forums here and I have learned a lot already. Any other good sources of current info? Good books? Any new and emerging thing to think about? I know for instance that Cree is making great strides with LED lights up to 254 lumen-per-watt. This means in 5 years or so LEDs may be the most efficient light for most areas. Only problem is a lot of heat is generated so if I design lighting I may need to think about higher heat fixtures needed for the LEDs of the future. Anyway, I am open for suggestions as I am in the beginning stages now.

The math works out for the highest heating loads to the ceiling, walls and foundation in most construction. It is due to the surface areas and temperature differentials and dictates how a Manual 'J' load is performed.

The highest temperatures are driven to the highest part of the house by convection so we insulate the attic best (also the cheapest per R-value. The walls and incidentally the windows are next due to their direct exposure to the outdoors and finally the foundation, which by definition is on or underground with steady deep earth temperatures in all but permafrost areas and with a certain resistance or R-value of its own presents the least load per square foot.

So for return on investment if follow that the ceiling is 1st, the walls and windows 2nd the foundation a distant 3rd. Radiating the slab will present a higher load but only marginally. So, as each area is important it pays to know where to invest and the climate in which you are building. Thus a proper heat load analysis before the ink dries on the CAD drawing.

No need to guess with all of the energy modeling software available free on the web. http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=10&sqi=2&ved=0CG8QFjAJ&url=http%3A%2F%2Fwww.grabows.us%2Fltu%2Fhpb%2FSolarThermalDocs%2FEnergy%2520Modeling%2520Software%2520Evaluation.doc&ei=hZuOUK7TM-yH0QHt8YGwDw&usg=AFQjCNHCKoYywYyUTMZdwvPGsySqAIsX6w&sig2=VoETiYK6LwaXQbHAtnAI3w

I used UCLA's user friendly HEED. You might start with UCLA's Climate Consultant just to get a handle on what strategies work best in MI. The Canadian Hot2000 has a very good basement energy modeling component, as does DOE2 and its derivatives. IIRC, Google Sketchup, a 3-D design tool, has EnergyPlus as a plug in, but that model is the opposite of intuitive IMHO.

Thanks Morgan, Todd,
So there is no actual data about heat loss from a building envelope? Is it all theoretical? I am happy to work through the math of a heat load analysis with the energy modeling software or by hand. I am sure all of the software was put together by some very bright people, but still software depends on the value of the input and how it calculates.

So far I haven’t found HEED or Climate consultant all that useful. I will try some of the other software that was evaluated in 2009 that you linked to Todd.

Thanks again, looks like my simple question of R-value balance has no easy answer. And as always, current "state of the art" will be laughable in 50 years. I will be gone but I hope my home is still standing.

That's right- there are no simple-answers that get you to the optimal balance- every house is different, as are the site & shading factors, and the design considerations aren't always perfectly aligned either:

It may not be cost-effective from an energy use point of view to go with U0.18 triple panes windows, but from a comfort point of view it might still be "worth it" in places where sub-zero F winter temps are common.

In other situations super efficiency heat-rejecting windows might be cost-effective in some cooling dominated location, but the cut in daylighting might be worth backing off from the absolute lowest solar gain factor.

In re LED lighting: Yes, you really can get 254lm/w light, but at a color temp and color rendering index that would be downright dismal to live with. But the amount of HEAT derived from high efficiency lighting is very LOW, since a greater fraction of the input power is being radiated away as light. The reason you see big heat sinks and venting on LED assemblies isn't because the LED is generating so much heat, but rather because LEDs are more sensitive to heat than other lighting types, drifting in color and intensity with temperature. Whereas 90% of the energy going into an incandescent bulb is dissipated as heat, only about 50-70% of the energy going into an LED fixture is dissipated as heat. Getting the same amount of light dissipates about 1/5 the amount of heat as with incandescents (at today's efficiencies) LEDs and CFLs are the solution, not the problem as far as keeping the temperature at the fixtures low. Linear fluorescents are still beating most LEDs on raw efficiency, but it's a moving target, with incremental improvements on both. The very best LED are at parity with T8 fluorescents on efficiency, but the upfront cost delta is still substantial. But over the lifecycle of the product (20+ years for best-in-class LEDs vs under 10 for fluorescents) LED can still have a financial rationale.

My current home is pretty old and actually had zerro insulation in it when I first moved in. I have watched over the years as recommended r-values have risen. So where is the hardest place to add insulation on my old house? The most difficult (if not impossible) area would be under the footings, next would be under the slab then along outer foundation walls. The easiest would be the house walls and roof.

So in my future home design I would think it would be prudent to insulate sub footing/slab and underground the most as a retrofit in 20 years would be the most difficult there. An extra few inches of foam are pretty inexpensive now compared to any future subgrade modifications. Also extra attention to detail on the drainage and exterior moisture control will be important. Wet ground will conduct more heat than dry ground.

LED lights are making big gains right now, it will be a few years till it filters down to the public. I have a small 1 battery XML or XMG flashlight that accidently got turned on inside my pocket -WOW - it got plenty warm. The tiny light seems as bright as the old policeman flashlight/clubs of the past.

Thanks all for your responses and knowledge sharing. I am a sponge right now.

Posted By kb on 29 Oct 2012 09:41 PM
...snip...
So there is no actual data about heat loss from a building envelope? Is it all theoretical? I am happy to work through the math of a heat load analysis with the energy modeling software or by hand. I am sure all of the software was put together by some very bright people, but still software depends on the value of the input and how it calculates.
...snip...

I have month-by-month natural gas usage for heating, domestic hot water (small because of solar hot water), and cooking (small) for my house at http://www.residentialenergylaboratory.com/rel_energy_use_natural_gas.html. The monthly average heating degree days are given at http://www.wrcc.dri.edu/cgi-bin/cliMAIN.pl?cosali. These measured results are close to what I computed using heat loss calculations that I performed before construction using REScheck combined with some hand calculations. (Details to be provided soon on website.)

The R-values that you came up with are close to those given by the Building Science Corporation that calls for cold-climate homes to have R-10 sub-slab foam, R-20 basement walls, R-40 above-grade walls, and R-60 ceilings. If you do the modeling for these values, you will probably find they are plenty sufficient, or maybe a little on the high side, for what can be paid off over 25 years. You might also consider high-solar-gain windows and solar PV if they fit into your overall package. They might have a better payoff than super-high insulation values.

Posted By kb on 30 Oct 2012 11:28 PM
My current home is pretty old and actually had zerro insulation in it when I first moved in. I have watched over the years as recommended r-values have risen. So where is the hardest place to add insulation on my old house? The most difficult (if not impossible) area would be under the footings, next would be under the slab then along outer foundation walls. The easiest would be the house walls and roof.

So in my future home design I would think it would be prudent to insulate sub footing/slab and underground the most as a retrofit in 20 years would be the most difficult there. An extra few inches of foam are pretty inexpensive now compared to any future subgrade modifications. Also extra attention to detail on the drainage and exterior moisture control will be important. Wet ground will conduct more heat than dry ground.

LED lights are making big gains right now, it will be a few years till it filters down to the public. I have a small 1 battery XML or XMG flashlight that accidently got turned on inside my pocket -WOW - it got plenty warm. The tiny light seems as bright as the old policeman flashlight/clubs of the past.

Thanks all for your responses and knowledge sharing. I am a sponge right now.

If you have enough headroom in the basement it's possible to add 1-2" of XPS on top of an existing slab and put a sub-floor on top of the foam to support a finish-floor.  If you insulate the interior side of foundation walls at the same time, you'll need at least an inch of XPS or EPS (more in colder areas) against the wall even if you're rounding out the R-value with an interior fiber-insulated studwall. Running the floor-foam and subflooring all the way to the wall foam (under the bottom plate of the interior studwall) there would be no point to insulate the footing. 

Limiting the thermal bridging at the foundation sill & band joist with spray foam that has continuity with your foundation wall insulation is usually pretty straightforward.

This is a pretty standard way to retrofit-insulate existing basement foundations- it works pretty well, and not difficult at all. Yes, it puts the thermal mass of the concrete outside the thermal boundary of the house, but the benefits of insulating far outweigh that of the thermal mass.

What is your ZIP code/location?  (Local climate makes a difference on the particulars how you insulate the foundation.)

The rate of progress of LED efficiency isn't nearly as interesting as the declining cost curve.  Most retrofit-bulb type LED assemblies are no better than self-ballasted CFLs on efficiency (50-60lm/w), but the light quality is usually pretty good- usually better color rendering than comparable-efficiency CFLs, but still 10x the price.  When it gets down to 5x the price the longevity of LEDs relative to CFLs make them far more attractive.  But don't count on 100lm/w 90CRI LEDs to show up at an attractive price any time soon.  Street price on the 90 lm /w 10W Phillips is still $40.   That's only ~50% more efficient than a $2 twisty running 13-14 watts for similar output, and you can buy dozens of twisty CFLs and several decades worth of "extra" power use with the $38 difference.  (And a T5-HO 14 still beats it on efficiency.) But when the Phillips hits $10 it'll be pretty attractive.  Efficiency gains on LED assemblies with good color rendering have been quite slow to materialize but the retail price has seen a steady year-by-year decline.

If you put any flashlight in your pocket and leave it there it'll get plenty warm, but the halogens of similar output will get a lot hotter (and a lot sooner), but the batteries required to run it for a reasonable time don't fit in your pocket the way an LED flashlight will. The color rendering of most flashlight LEDs is pretty g'dawful though, unlike the better edison-base LED retrofit assemblies.

So, I have found many references to the 5/10/20/40/60 ratio for windows/slab/foundation/walls/roof but I am still not sure why. I understand the theoretical opinions out there but I have not stumbled across much data to prove or disprove them.

Apparently the experts feel sub slab you are looking at a temperature differential of 65 degree basement to a 40 degree earth, so they don't think you need much insulation there. Is there no difference between conducion between solids that exists below the ground compared to walls/roof that are exposed to air which is much less conductive?

Why do roofs need R60 compared to walls at R40? Is it that much hotter near the roof to cause the temperature diference to be greater? Lets say the room/walls are at 70 degrees, is it now 85 degrees in the interior near the roof? If it is 20 degrees outside the temperature differences are now:
slab ~25
foundation ~ 35
walls ~50
roof ~65
Are these the basic assumptions? If they are, why aren't we sucking the hot air near the roof and putting it lower in the structure? If it is only a few degrees different near the roof then the roof/walls should be the same R value, right?

Posted By kb on 04 Nov 2012 11:32 AM
Is there no difference between conducion between solids that exists below the ground compared to walls/roof that are exposed to air which is much less conductive?

Do a search on convective and conductive heat transfer and read about the differences. You will find that convective heat transfer can be much higher than conductive heat transfer. The transfer of heat from the walls and ceiling of a house to air is dominated by convective heat transfer, not conductive.

Why do roofs need R60 compared to walls at R40? Is it that much hotter near the roof to cause the temperature diference to be greater?

It is a matter of what is physically and economically practical. It is inexpensive to pile up R-60 worth of insulation in an attic, while it is much more expensive to even reach R-40 in a wall. In a cooling dominated climate, the attic temperatures can be considerably higher than the outdoor temperatures, so that is another reason for extra insulation on the floor of the attic.

The heat loads are greater at the exterior, more especially at the ceiling where warm air stratifies. This is why we use a purpose-built radiant floor software.

Are you trying to learn thermodynamics via internet? If you have hired a radiant floor heating "expert", the design should be done and the questions few.

Thank you for your input gentleman,

Badgerboiler, you are right, what was I thinking? there are people on this site trying to make a living, why would they want to give out free information if they do not profit from it? I should be looking elsewhere I guess.

Posted By kb on 08 Nov 2012 05:01 PM
Thank you for your input gentleman,

Badgerboiler, you are right, what was I thinking? there are people on this site trying to make a living, why would they want to give out free information if they do not profit from it? I should be looking elsewhere I guess.

As one who does not currently make any part of my living from the building, heating, or energy efficiency biz I gotta say that statement is a bit short sighted. There are LOTS of professionals who will pass on free information on this website and others, and if it ultimately helps their biz, great.

But nobody is going to spend the time to fully engineer a solution for you over 27 posts, even if you WERE going to pay them for it.

And there are many people elsewhere perfectly willing to take your money who may or may not have a clue, eh?

I agree with Dana1's comments.

A lot of us give without expecting anything in return.  Since I retired in 2004, I donate most of my free time to helping homeowners save money on building and energy.

Thanks guys, I was being a bit facetious, a poor attempt at humor on my part. I suppose I was also trying to seperate the wheat from the chaff. I apologise to the "wheat" out there. I am a knowledge sponge. Even if I would hire the most expensive "expert" out there I would have a million questions. The more I learn the more questions I have.

Some people are content to read the last page of Consmer Reports ..."just tell me what blender to buy". That is not me. I want to know every detail possible on why I am doing what I am doing.

So my post was to to determine if there is any data on R-values. It appears that most of the information out there is generated by theoretical calculations more than actual data. Different countries seem to have taken a different path than North America, and have a completely different building system to learn from. Theories are good, but unbiased real comparative data is better.

So far it seams that throwing another layer of EPS below the slab/footings/foundations is the easiest to do while building and the hardest to retrofit later. That is why I am questioning the standard 5/10/20/40/60 ratio for windows/slab/foundation/walls/roof when I see some deviations out there with positive results.

Lee Dodge, Thanks for pointing out the obvious to me regarding convective heat losses. I really like thought experiments as a way to find answers so, regarding convective losses I am reminded of the following examples. An ice cube will melt faster under a running faucet than sitting still in a glass of water, I typically cool off a hot cup of tea or bowl of soup by "blowing" on them, convection ovens work by "blowing" the heat around.

These thought experiments lead me to believe that the velocity and volume of air will be a factor in convective losses above grade. Inside the house, these convection currents can be minimized and relatively low, but outside they will vary by the prevailing wind velocity. So of course that brings me to question the r-40 wall and r-60 roof relationship. Suddenly where the wall becomes the roof I need an extra r-20? Isn’t the wind the same speed on the outer walls as it is on the roof? And inside aren’t the convective currents/losses close to the same for the walls and ceiling? (This is backed up by the 1988 Gebhart literature)

I believe that similar conditions exist for the turbulent flow at the boundary layer of the interior walls and the ceiling. With a cathedral ceiling it would seem the natural “heat rises” buoyancy effect is even less than the buoyancy at vertical walls. Also, in experimental conditions Holford and Woods reported in the “Journal of Fluid Mechanics in 2007 that “the fluctuations of the temperature of the thermal mass lag those of the interior air, which in turn lag those of the environment.”

Currently I am leaning towards radiant floor heating so there will be less “blowing” and less convection inside. In other words I am trying to go more with natural convection vs forced convection thus lowering the velocity and volume of the internal air in order to lower the internal convective losses. Of course a tight house would necessitate the use of an HRV which will introduce more convection.

I am planning on using SIPS, so I an adjust my r-Value with thicker or thinner panels.

Posted By kb on 10 Nov 2012 09:17 AM
...snip...
So of course that brings me to question the r-40 wall and r-60 roof relationship. Suddenly where the wall becomes the roof I need an extra r-20? Isn’t the wind the same speed on the outer walls as it is on the roof? And inside aren’t the convective currents/losses close to the same for the walls and ceiling? (This is backed up by the 1988 Gebhart literature) ...snip...

You apparently have a different objective from many of the rest of us. You are looking for a constant heat flux (heat loss per unit area per unit time) through each part of your house. That is, you want the heat flux through the walls and the ceiling to be the same (as well as to the ground), independent of the cost of getting there. That is different from my objective, anyway, which was to minimize the heat loss of the whole house given a certain budget to work with. So when cost is factored in, reducing the heat fluxes through walls is more expensive than reducing heat fluxes through the ceiling with "conventional" construction, so I chose a higher R-value for the ceiling than the walls.

For example, increasing the R-values of my walls from about R-19 to R-29 by adding 2" of XPS rigid foam cost me $5000 (including labor), with a payback of 42 years, while increasing the ceiling insulation from R-38 to R-60 cost me only $1120 (including the cost of raising the raised heel), with a payback of 28 years. Details are given here . It was not practical, in my case, to increase the wall insulation beyond R-29 because the return on investment (ROI) was already quite long. It would have required a different type of wall construction to reach very high wall R-values. The cost per unit R-value of rigid foam is quite high compared to that for cellulose, so that is one reason behind the costs

Since your objective is to balance heat fluxes through the walls and ceiling, then indeed, you should choose similar R-values for the walls and ceiling. In fact, since the walls are vertical and the ceiling or roof is flatter, the natural convection will be higher for the walls than the roof. Therefore, you might need to choose a higher R-value for the walls than the ceiling/roof. On the other hand, if you are using A/C, then you need to consider solar heat loads on the ceiling/roof. You will also need to figure in the natural stratification of temperature in the house, with higher temperature near the ceiling, but the stratification should be low for a tight house. Obtaining the same heat flux through the basement and slab is more difficult, requiring complex 3-D calculations, but you seem to be familiar with the literature.

Since you plan to use SIPS that incorporate the relatively expensive rigid foam for insulation in both the wall and ceiling, then matching the wall and ceiling insulation levels should be easy. Indeed for SIPS the recommended 10/20/40/60 guideline may not apply.

Thanks Lee, that makes sense. I wasn't aware that the 10/20/40/60 was based on cost/labor. If it is, then the cheapest labor install would be the EPS in the subslab area - just lay it down! I first used SPIS over 20 years ago after I saw an article in FineHomebuilding and I made my own SIPS. I found the ease of istallation and reduced labor to be quite significant as well as less thermal bridging due to lack of framing. It has stood the test of time so far even though SIPS technology has come a long ways since then.

Your comments have also led me to realize that walls and ceiling should be rough and not smooth to minimize surface boundary losses by lowering the velocity of the air currents near the surface. I am not planning on putting shag carpet on the walls but a rough plaster will be more optimal than a shiny gloss painted surface.

So I will continue to research the origin of the 10/20/40/60 mark and balance of heat fluxes. Thanks for your input.

Posted By kb on 10 Nov 2012 12:24 PM
...snip...
Your comments have also led me to realize that walls and ceiling should be rough and not smooth to minimize surface boundary losses by lowering the velocity of the air currents near the surface. I am not planning on putting shag carpet on the walls but a rough plaster will be more optimal than a shiny gloss painted surface.

Perhaps I have misled you. My comment about convective heat losses being high relative to conduction was just to address why it is recommended that you use higher R-values in the wall and ceiling than the below-ground basement and slab. Heat losses from the wall and ceiling are related to the convective heat transfer from the outside surfaces to the air. Heat losses to the ground are related to the conduction from the outer surfaces through the earth, and the ground (1) does not get as cold as outdoor air temps, and (2) the ground does not offer convective heat transfer. Therefore, less R-value is recommended for the below ground areas.

Concerning the heat losses from the walls and ceiling, the heat losses from the inside of the walls to the outside is governed (mostly) by conduction through the walls. There is convective heat transfer from the air to the inside surface of the walls, but the R-value for that air film is roughly R=0.68, and for the ceiling R=0.61 (http://www.coloradoenergy.org/procorner/stuff/r-values.htm). (There is also convective heat transfer from the outside surface to the air, but the R-value here is roughly 0.17, lower due to forced convection.) These numbers are small compared to the R-value for an insulated wall or ceiling. Bottom line, don't worry about he wall roughness, or about the velocity of air inside the house.

So I will continue to research the origin of the 10/20/40/60 mark and balance of heat fluxes. Thanks for your input.

Rather than using this guideline, especially since you plan to use SIPS, you need to do your own heat loss calculations. I like to use REScheck, a free download, which will allow you to compute UA for each section of the house, where U is the conductivity, and A is the area. The heat loss is proportional to UA. That will allow you to see the relative contributions of different parts of the house to overall heat loss. REScheck will not compute heat transfer through a slab, though.

Thank you Lee, you have been very helpful.