This well-written article goes into detail with what we are talking about here, it was written by our resident forum member "Cameron Ware"

ICF MAGAZINE - THE ICF TIPPING POINT


I agree, there is somewhat of a "backlash" against ICF on this forum. The reality is that the "green building" realm does this to itself in many other areas as well. One should see the backlash when someone tries to get Passive House standard and uses "too much" slab foam or goes for a R-100 ceiling in order to get PH certified. Man, you would think they just sacrificied their firstborn to some pagan god. The green building blogs explode and lambast the homeowners for being poor stewards and the insurmountable "embodied energy" they wasted trying for PH certification. Or when someone builds a Net Zero Home but it has 4,000 square feet of living area. Oh boy, here come the "green gestapo" and round them up for a swift tongue lashing about how wasteful they are for building such a big home and throwing on PV solar to achieve Net Zero.

There is an unwritten rule in the green building realm that one human being can only occupy 800 square feet of living space per home. So two adults shouldn't see more than 1,600 square feet of home or they are being poor stewards of the earth and wasting resources living in a large home. These are just few examples but it goes on and on and on....

The point is that there is no one perfect building method and what's green to one person is being a polluting hog to another. Instead of the movement self-cannibalizing, it would be nice if the goal of building energy efficient homes was recognized and not turned into a slug fest.

As far as ICF goes. ORNL recognizes the DBMS and nobody is discounting physics/the laws of science. An R-21 wood framed wall vs. a R-21 ICF wall, perform differently, that is no doubt. Ideally, a hot box test using modern day ICF walls would be a great way to see what the data shows. ORNL's study was many years ago and a new study is desperately needed. Even Dana would agree that an ICF home in the desert SW or where vast diurnal temperature swings happen, is the way to go.

Besides the energy efficiency that ICF brings to the table. The superior strength, fire resistance, sound control, and longevity are also viable reasons for going with ICF rather than wood frame.

Now back to the regular scheduled programming...

Thermal lag is still working in a high mass wall even in winter. Thermal lag is the time it takes for a temperature event on one side of the wall to make it to the other side. Typically it never does as fluxes of heat and cold over 24 hour periods cancel each other out in concrete. In my part of the world, it means I can ignore the DD day temp of 10 degrees and size hvac for the average daily temperature of 30 degrees instead. But in winter, thermal lag has almost no effect on energy use. The average daily temperature is still 30 degrees no matter what, and as sailaway notes, correctly, if ambient temps never get above "comfortable", heat is flowing only one way. Out.

And no one is picking on ICF. The answer is plainly obvious, and has been offered in every instance. Run a wall section through a hot box test blessed by ORNL and publish the results. Then stop the BS anecdotal claims of R50 in Calgary. Alternatively, stop the BS anecdotal claims of R50 in Cal;gary. With bona fide results, Lbear could fine tune a TF wall for AZ. He might also find out that ICF is not the best wall choice in the desert SW. But writing finish to this topic would be a blessed event, too. No?

I dunno if I'm "grasping at straws" to discredit an R50-equivalence claim. I work as an engineer ( and have a degree in physics)- believe in taking measurements & using the math/modeling more than WAGs and armchair assertions. Very few heat load tools used by the HVAC trades model the mass & site factors well, though some are better than others. I have no insight to FBBPs heat load calculation methods, other than the assertion that he used R50 as the wall-R input, which doesn't imply a very sophisicated tool.

Thermal mass modeling effects of wall assemblies is a fairly well refined, and the actual performance is demonstrably climate, design, and site-specific. Assigning a single "R-equivalent" number to compare it's performance with a low mass wall of that R value, without the site, climate & design factors is pure BS, and ignores the true effects of the thermal mass. Even though it's an R23-ish wall at steady state, never have I stated the mass is irrelevant to either the peak or average loads. But under any site & design conditions it would be hard to support an "R-50 equivalent performance" on either average energy use or peak loads in a Calgary climate.

But I'd argue to sailawayrb that the mass effect is more than just "...the outdoor temp cycles significantly BOTH ABOVE and BELOW the indoor temp within a 24 hour period. ":

Outdoor & indoor temps are just two of the inputs to the filter. At night the radiated surface loss is determined by the temp of the surroundings, which on a clear night could average a degree or so below the outdoor temp. But that is more than made up by the radiational gains (even on the north side, and in the shade) during daylight hours. On the sides that get direct sun the surface temp of the siding easily hits 40-50F ABOVE the outdoor ambient, and in an ICF that's usually above the temp of the concrete, so that mass is getting "charged" with heat from both sides- the warm interior on one side, the passive solar gain on the other. This gain is measurable (if FBBP want's to instrument the house, you'd be able to show), and affects the both the peak load through that section of wall, as well as average daily/weekly load. When there's snow on the ground even the siding on the north side of the house can be a few degrees above ambient on a sunny day, due to the higher radiational temperature of the landscape. With a low-mass wall the passive solar gains on the siding only reduce the daylight hours load through the wall.

There's nothing mysterious about how the local climate, site factors, and house design can affect the magnitude of this potential benefit by an order of magnitude or more, which is why making use of the available models is useful for predicting the actual performance. An ICF house that's always in the shade, with minimal direct sun won't get nearly the solar-gain input to the walls of an ICF oriented east-west in full sun, and the outdoor temp input would dominate. With low mass walls the direct solar gain benefit though walls is measurable, but since it doesn't get stored, it has very little effect on the peak loads, since the peak loads typically occur before dawn on nights of high radiational cooling, more than 12 hours after the last solar input.

Prescriptive R-values under building codes are about average (not peak load) energy use characteristics of typical/average home designs, and are based real measured performance of typical buildings.  Calgary's climate is approximately a US climate zone 6 climate, but with more winter sun than some parts of US zone 6. The IRC 2012 prescriptive values for zone 6 in Table N1102.1.1 are:

R20 + R5  continuous insulation, for 2x6 construction  (whole-wall R= ~R18-R19 at typical framing fractions)

...or ...

R13 + R10 continuous insulation for 2x4 construction (whole-wall R= ~R19-R20 at typical framing fractions.)

...or ...

R15 continuous insulation in a mass wall with 50% or more of the R on the exterior of the mass (a 2" + 2" bottom of the line ICF)

...or ...

R20 continuous insulation in a mass-wall with less than 50% of the R on the exterior of the mass (CMU with rigid R20 continuous  insulation on the interior side.)

So R15 ICF is the rough equivalent of an R19-R20 whole-wall low-mass wall, from an energy use perspective.

This will scale approximately linearly with higher R, which put put the typical performance of an R23 ICF at rough-equivalece with an (R23 x R19/R15 =) R29 or maybe R30 whole-wall low mass wall. (R20 2x6 16" o.c. with 3" of exterior XPS or polyiso gets you there)  That's from an energy use point of view- the ICF will perform better than that from a peak-load point of view.

How much better really depends on the aforementioned site-factor/house-design/local-climate aspects, but the models predict it pretty well.  In most home designs that come in at R23 whole-wall (independent of mass effect benefits) the windows losses have a bigger effect on the peak load than the walls losses, and the window gains have a bigger effect than the passive wall gains on the average energy use.  Even the inherent error in unmeasured ventilation rates can as big an effect on peak load than the difference between R30 vs. R50 wall performance in most home designs.

But if you're in love with the concept of R50, who am I to argue with love?

In regards to the OP question. Couldn't one just install an ICF wall between the garage and home instead of wood framing the partition wall? Pouring a separate footing for the ICF garage partition wall and thermally breaking the garage and home slab from each other?

In this instance a SIP roof would probably work better in that you don't have a shared attic or any attic in that case.

Posted By Dana1 on 27 Nov 2013 11:43 AM

This will scale approximately linearly with higher R, which put put the typical performance of an R23 ICF at rough-equivalece with an (R23 x R19/R15 =) R29 or maybe R30 whole-wall low mass wall. (R20 2x6 16" o.c. with 3" of exterior XPS or polyiso gets you there)  That's from an energy use point of view- the ICF will perform better than that from a peak-load point of view.

How much better really depends on the aforementioned site-factor/house-design/local-climate aspects, but the models predict it pretty well. 

So it wouldn't it be a far reach to say an R-23 ICF wall performs like a R30 whole-wall wood framed wall?

Personally, would you ever build or recommend an ICF wall system home?

ORNL's study was many years ago and a new study is desperately needed.

I second that. There are numerous brief references to studies that don't seem to be available. And various important assumptions aren't disclosed.

Any claims about internal mass savings that don't specify how much the occupants were willing to let indoor temps vary is suspect.

Posted By Lbear on 27 Nov 2013 01:13 PM

Posted By Dana1 on 27 Nov 2013 11:43 AM

This will scale approximately linearly with higher R, which put put the typical performance of an R23 ICF at rough-equivalece with an (R23 x R19/R15 =) R29 or maybe R30 whole-wall low mass wall. (R20 2x6 16" o.c. with 3" of exterior XPS or polyiso gets you there)  That's from an energy use point of view- the ICF will perform better than that from a peak-load point of view.

How much better really depends on the aforementioned site-factor/house-design/local-climate aspects, but the models predict it pretty well. 

So it wouldn't it be a far reach to say an R-23 ICF wall performs like a R30 whole-wall wood framed wall?

Personally, would you ever build or recommend an ICF wall system home?

That would be, on-average, in US climate zone 6, and only from a whole-house energy use point of view, probably pretty close.  In a much warmer sunnier climate it'll usually do a bit better (on average), in colder climates, worse (on average).

But when you drill down to the particulars rather than the averages it's really all over the place, which is why modeling it makes a difference. If on a particular site/designe/climate if, say, you rotated the orientation of house 40 degrees or changed it's shape and it does a whole lot better, wouldn't you want know that before you poured the footings?

Posted By Dana1 on 27 Nov 2013 01:45 PM

That would be, on-average, in US climate zone 6, and only from a whole-house energy use point of view, probably pretty close.  In a much warmer sunnier climate it'll usually do a bit better (on average), in colder climates, worse (on average).

But when you drill down to the particulars rather than the averages it's really all over the place, which is why modeling it makes a difference. If on a particular site/designe/climate if, say, you rotated the orientation of house 40 degrees or changed it's shape and it does a whole lot better, wouldn't you want know that before you poured the footings?

In the desert southwest climates like Northern Arizona that has almost 300 days of sunshine and diurnal swings of 30-40 degrees in 24 hours. ICF walls should perform better (on average), is a fair statement

I agree, modeling is important and I really believe ORNL needs to do another in-depth testing on ICF walls to finally put this debate to rest. The last test left too many questions and open ends.

Personally, would you ever build or recommend an ICF wall system home in such a climate like the desert SW?

Dana, I think we are saying the same thing although it is complicated to fully explain to someone without the technical background and any explaination is also totally lost on someone who simply believes otherwise. I view and model thermal material R-factor and thermal material heat capacity in a similar way that I model electrical resistance (R) and capacitance (C). In short and as you know, the RC of an electrical circuit determines it’s frequency. The electrical circuit frequency determines how the circuit can be used. Electrical capacitors require alternating current to operate...they do NOT pass direct current. High heat capacity walls require alternating temperatures to operate. The thermal lag effect created by a high capacity wall has to be aligned with the intended use of the wall in order to achieve the desired benefit. This is why even uninsulated concrete walls work very well in climates that are very hot during the day and very cold at night…if you get the wall thickness correct and hence the RC frequency characteristics of the wall correct. By the time the daytime heat reaches the inside surface in the evening, the evening outside temp has cooled down that this wall heat is greatly welcomed. By the time the daytime heat comes back, the wall has already cooled off and keeps the inside cool.

Lbear: The desert southwest is an optimal location for getting the performance out of ICF, both in terms of the preponderance of high diurnal temperature swings, and the high insolation even during winter.

sailawayrb: You're right an electronic R-C-R filter analog models ICFs well, but the temperature/voltage polarity doesn't need to switch (a true AC signal) for the filter to have an effect, it only needs to vary with time. A uni-polar blip will not only affect the average, but also the peak. It's important to take into account that the surface temp is not identical with the outdoor temps on cold but sunny days, and that the temperature gradient between the concrete and exterior surface can change directions even when the outdoor temp is still well below the indoor temp.

Take an IR thermometer reading of your siding in full sun on a calm 20F day- don't be surprised if it's as warm or warmer than your interior wall temp. In an ICF that input gets integrated by the thermal capacity of the concrete, lowering the peak heating load even 15 hours or more later, whereas in a low mass wall the effect is negligible.

Posted By sailawayrb on 27 Nov 2013 03:44 PM

High heat capacity walls require alternating temperatures to operate. The thermal lag effect created by a high capacity wall has to be aligned with the intended use of the wall in order to achieve the desired benefit. This is why even uninsulated concrete walls work very well in climates that are very hot during the day and very cold at night

What the above describes is what the Native Americans used for shelter hundreds of years ago in the desert Southwest. Adobe and masonry homes were used by the Indians in areas like Arizona. Some of the shelters still stand today, over 800 years later (Montezuma). Concrete is man-made rock/stone and like natural stone it takes into account the basic principles of strength, fire resistance, termite resistance, and of course thermal mass.

The first time someone personally experiences the desert climate is sometimes an eye opener for them. To see 80F highs in the daytime with the intense sun beating down on you. To when the sun sets and temperatures drop into the low 40's. Bake you in the day, freeze you at night, that is the desert climate.

This is why ICF modeling is not straight forward and black and white. A wood-framed R-30 wall is still a wood-framed R-30 wall whether it is built in Phoenix or Minnesota. An ICF wall built in Phoenix behaves differently than an ICF wall built in northern Minnesota. This is where wood and ICF cannot be compared like apples to apples, it turns into an apples to oranges comparison. The thermal mass is where the "wrench" is thrown into the R-Value equation. It's not a simple calculation of the thermal resistance of insulation. While concrete doesn't contribute to the R-Value, it sure does contribute something to the way to thermal lag and mass effect the temperature swings. Hence the reason why ORNL termed the "DBMS" Dynamic Benefit for Massive Systems or "effective R-Value" for mass/ICF walls. The scientists and researches knew from their studies that the concrete is doing something but how to quantify what it's doing is where the test came short. For now, they termed it DBMS and effective R-Value. Until a deeper study is done, that is what we have to live with for now.

The Native Americans of 1,000 years ago figured it out, in the sense that thermal mass works for shelters in desert climates, even though it technically doesn't have any significant R-Value to it. We just need to put the theory into recognizable data so it can finally stop this confusion and debate.

The theory is sufficiently well modeled (albeit in 2D only) in DOE2.2 to be predictive of average energy use, and within the statistical noise of construction method error.

There isn't really any confusion or debate until/unless you try to factor out the dynamic effects of the assembly given it's site-factors & climate by assigning it a dumbed down single number like "equivalent R". It's a meaningless number, since the actual performance will vary dramatically with how & where the ICF is used, and whether you're talking about average energy use vs. peak load vs. the detritis build up of corn broom on curling sheets degrading or enhancing the game, or something even less relevant. With an ICF the heating/cooling loads of the intrior are affected by the delta-T between the interior and the concrete. The temperature of the concrete is affected by several other factors most of which are filtered/damped by the exterior R.

There is no way to assign a single comparative performance value, since the other inputs are unknown, even if you know the performance characteristic you're most interested in (average energy use, peak load, induced spin on the rock from corn stalk fibers.) The performance varies with context, but that doesn't mean it's unknowable- the models work pretty well.

But I'll admit it's hard to SELL the stuff with all of the necessary fine print without making the customer's eyes glaze over. A single number is easier to sell, even if any single number is complete BS since it's free of the all important context, but it leads to some really bad habits and over-the-top claims. Anybody who really cares about it should make use of the available models. But since an 2.5" + 2.5" ICF beats code in all US locations, the "really cares" aspect is usually lacking. It's better than code min, and the dynamic affects of the thermal mass will give it another slight kick above what a low mass wall of equal steady-state R would be. But it's only one aspect of the house's ultimate performance, and it's rarely the best price/performance deal out there on energy grounds alone. (The structural aspects of concrete construction is also definitely worth something, more in some regions than others.)

Exactly right, Adobe is a wonderful use of high heat capacity wall “technology”. It was actually the Syrians who first used this “technology”, but the Native American Indians likely did so independently too! We have actually done quite a lot of work refining our thermal mass models to actually allow proper engineering of passive solar buildings and more so to allow proper engineering of integrated passive solar and hydronic floor heating systems in buildings. For clear sky solar conditions we can now accurately forecast the thermal mass temp (and hence it’s heat gain to the living space via convection and radiation) within 0.5F (and 1.25 BTU/H-SF) for any date and hour. We have even made free DIY thermal mass software available although you can only do so much with JavaScript as compared our MATLAB simulations. We found the “rules of thumb” that are typically used to design passive solar buildings to be complete rubbish.

Borst Passive Solar Thermal Mass Performance Software

A wood-framed R-30 wall is still a wood-framed R-30 wall whether it is built in Phoenix or Minnesota.

Even the mass in a typical wallboard lined home is significant. And as always, how significant depends on things like opening the windows and allowed interior temperature range.

As the models show, it's possible for external mass to cause an increase in energy usage in some climates.

I should also add that while I am a huge fan of ICF buildings, I have little love for Passive House standards. My feeling is that any building that looks that ugly will never be loved and maintained long-term no matter how well it performs. Just to set the record straight...

Todd - what do you think I’ve been saying? If you prefer to raise your DD temp by 20fº or I say use R50 instead of raising my DD temps, its going to be the same general idea, except that increasing the R value for the ICF is much more accurate because it just effects the ICF portion of the envelope.

When ORNL builds a lab large enough to house a 6000 sq. ft. three storey house then maybe they will get the same results that people living in ICF home have been getting for twenty years. An actual house where all you are doing is switching the r value of the wall section of the envelope when you do heat loss calc and then comparing the actual results is much more accurate then subject a small section of that envelope to a lab test. But lets look at the Hot Box concept for a bit. When you run a trial in a hot box, you set the parameters you expect, then lock the subject into the box. You run the cold side down to the temps you set and then start pumping heat into the hot side till you reach your preset and then measure the amount of energy it takes to hold that inside temp. From this measure you see if your original assertions where correct and with it you can rate the thermal conductivity of the subject mater.
With my house I had a professional using one of the best software programs available to heating designers to set the parameters in his heat loss calc’s. I then built the house and tested it. As you know from previous threads, the first two years I heated the house with two electric water heaters with all four elements working. Then I waited for a week of very near design temp conditions. The week had lows around -33 c and daytime highs around -26ºc. It was heavy cloud and stormy with medium to high winds. While I did not have control of the exterior temps, they turned out to be very near to the conditions you would set the Hot box for, plus wind. The first year the house was not occupied so no occupant introduced heat sources. No domestic hot water heater so no showers etc. No appliances and no lights accept the small amount of time each day that I worked in the house. So inside conditions where also very near to what you would have set in a hot box. Also no siding on the ICF so reasonably high reflectance if the had been any sun. Ceiling and floor slab insulation as per the heat calc’s. The heat calc’s indicated that the home should consume 52,000 btus. Now its true, I can’t say exactly how much energy was put into the house, but it is easy to calculate the maximum energy that was available. I don’t think anyone is going to claim that those four 3000 watt elements put out anywhere near 50,000 btus. So for all intents and purposes, I have replicated the Hotbox but did it quite a bit better in that I was testing the whole assembly rather then just an 8 x 8 square.
The second year was again heated with the water tanks and the third year with a 50,000 btu HTP boiler. And again each year had good results. True, the second and subsequent years had more intrinsic heat then year one as by then we were living in the house. So people can claim my comments are just hype and without merit but the facts prove otherwise. And at the risk of repeating myself, mine is far from the only ICF house in the Calgary area that had their heat loss calc’s based on R 50

BS wither buffalo or bull has been used to heat homes for centuries but I can assure you that none is being used to heat this house. On the other hand Webster defines FACT as “the quality of being actual” or “something that exist or occurs.” A better definition give by some is “something that can be measured and can be reproduced in the same manner.” All of this applies to ICF houses in Calgary. I may not be able to correctly explain it but I can for sure substantiate it and I can replicate it. It is not unusual for an assembly to perform different then the sum of its parts.
Have we ever had a complaint on GBT entitled “I can’t heat my ICF home”??

Fascinating conversation, guys. You've definitely confirmed for me that insulated mass walls are the way to go in the desert southwest, and I think you've convinced me that ICFs make more sense. My only remaining hesitation concerns the interior insulation of most ICF walls which seems like it will substantially reduce the ability of the concrete mass to absorb interior heat and solar irradiance, since it will be intentionally insulated from those heat sources. I know some folks have mentioned wall systems that have extremely thin interior insulation, but has anyone made an IFC product that uses a non-insulating mass material like drywall or something as the one side of the form? The Albuquerque area is ridiculously sunny so it would be a shame not to use the available mass to store some of the heat from all that sunshine.

Posted By ILikeDirt on 28 Nov 2013 01:49 PM

The Albuquerque area is ridiculously sunny so it would be a shame not to use the available mass to store some of the heat from all that sunshine.

Mass doesn't just have to be the wall structure. One can have exposed concrete slabs on the south elevation and they work great as thermal mass. Also, you can install an interior finish on the drywall like PlasterMax.  It's not a high mass wall but it does provide some thermal mass absorption. It can also be applied directly to the EPS interior foam on the ICF wall. Even 1/2" drywall has some mass to it.

I am in Arizona so I am also going with ICF also. I will have exposed interior concrete floor slabs on the south end to take advantage of the thermal mass storage. The climate of Albuquerque is probably a Zone 5, I assume, or is it still Zone 4? Either way, it is the ideal climate to utilize ICF due to the abundant sunshine and vast diurnal swings.

Also, if you are in a wildfire area (as is most of the desert southwest), ICF (6" core) has a 2-3 hour fire rating vs. wood frame which is around 30 minutes. Ideally, don't install gable, ridge or soffit vents and use a closed conditioned attic but if you do go with an open attic design, get ember resistant vents. That is how 80% of wildfires start, by embers landing inside the gable, soffit, and attic vents and catching your roof on fire. I would recommend a steel SIP for your roof but that again is design dependent.

Aha! Found what I was looking for: http://www.nudura.com/en/icf-wall-technology/oneseries.aspx

Albuquerque is zone 4, but just barely. Areas immediately north and east are zone 5.

Thanks for the roof suggestions. Great info.