I'm looking for pros and cons and cost of two wall systems; CMU with outsulation, covered by stucco or Hardiboard. Or SIP panel walls, again covered by either stucco or hardiboard. Location is zone 4, (california zone 11) which is Mediterranean, warm summer, mild winter, 25 -30 inch precip mostly in winter. Either version probably has a SIP roof, as we want open beams or timber purlins. Slab floor for some mass, north wall is against a hilside so it probably has to be a structural retaining wall. Title 24 only requires R8 on concrete wall. Your thoughts?

It's easier to air-seal the SIP, but at identical air tightness and R value the CMU wall would outperform the SIP and be somewhat more comfortable. In thinking about dynamic-loaded R for the mass wall, the fact that you have such substantial thermal mass elsewhere in the house the beneficial effective on the utility bills won't be as pronounced as in a low-mass building, and the lower absolute R may result in higher energy use than SIP walls in an otherwise high mass building.

No matter what, go for more than R8, even if that meets code under Title 24. R8 mass walls are the energy-use equivalent of 2x4 construction with a skim-coating of exterior foam. The extra ~40 cents per square foot of wall area it takes to bring that up to R12 would be pretty small in the grand scheme of things, and R16 would make a noticeable comfort difference at the temperature extremes, even with the mass of the wall. But your design temps are pretty modest.

I didn't see mass walls package listed in this document, bu t R19 2x6 is ~ R14 whole-wall:

http://www.pge.com/includes/docs/pdfs/about/edusafety/training/pec/toolbox/arch/climate/california_climate_zone_11.pdf

Thanks for the reply, Dana. This is pretty much what I had in mind about performance; need R14 or better outside boards on block walls,. I don't have price data though, so I am at a loss as to what the mass wall and insulation package might cost relative to SIP. Both walls have to have siding, so that seems like a wash. Air seal of the CMU seems easier to me; the SBC coat should be quite effective at that. Short of getting bids, does anyone have experience with one or the other methods, cost wise?

Construction costs are a local phenomena: specifically the number and eagerness of subs in your area who do CMU or SIP. Sips.org will lead you a members near you. If your experience is typical of this site, you will find the number and eagerness of sip builders to be lacking.

A caveat also about performance: I am pretty sure Calif code requires cmu walls to be filled with concrete, which results in a LOT of mass. You will read here an account of mass failing in a similar wall, sheathed in r14 polyiso, in Tucson zero energy house. (But now why. ARghhh!) http://www.toolbase.org/PDF/CaseStudies/TucsonZEH1Report.pdf The why is particularly intriguing here because there are literally hundreds of adobe houses within a hour's drive that rely solely on the daily buffering effect of thermal mass, plus an occasional blaze in the kiva, or the southwestern version of a masonry fireplace.

There a couple takeaways nonetheless. You need heat capable of taming the elephant you have invited inside your home. The tucson report implies/says that things might have/would have been different if radiant heat or passive solar had been available to heat the mass directly. Ergo, the most likely culprit was low-mass heat in the tucson house short cycling to the point of offsetting the buffering gains.

The second is muster some healthy skepticism about applying low-mass conventional wisdom to high mass houses. Presumably, the Tucson designers used all the latest and best design tools. There is a professor nearby (maybe) working on thermal mass. http://www.csuchico.edu/advancement/chico_stories/gkallio.php You might try to pick his brain.

Keep in mind that in some climates (probably not yours), you can get greater savings at far less cost using a programmable setback thermostat with low mass construction (like wood frame or SIPs).

This certainly could use more explanation: "in production housing where passive solar design is often not feasible, a home having high thermal mass may lead to additional heating energy use. The thermal mass may lead to occupant discomfort, ... it is not apparent that is has a benefit for the ZEH at APdS."

That Tucson house has a 10 inch slab, which in itself is a huge amount of thermal mass. Not designing it for passive solar was a mistake, but a radiant slab that thick is nearly impossible to control to match the diurnal shifts in load- it's too unresponsive, unable to change it's output quickly enough. If they'd gone with radiant floor they would have to cut the mass of the slab roughly in half.

I've been in a number of those adobe-mass houses in winter- I'm not sure you'd want to use those as your paradigm. Yes, they rely on the mass of the wall to even out the load, but they still take a substantial amount of energy to heat in winter. (I've never been in an adobe house with a private kiva though.) The "failure" of the mass systems in the Tucson house is only a failure to live up to the expected performance as modeled with no-setback constant setpoint op3ration.

I didn't read any indication (between the lines, or directly) of short-cycling in that report, nor is there any reason to believe that the heating system WOULD short cycle in a high mass house simply because it was using low-mass air as the final heat transfer fluid. (It's all-hydronic up to the air handler, and well buffered with a water tank.) The mass of the house always evens out the load if kept at a constant temp, but it also makes it less efficient if setbacks are used. The heating system uses a tankless electric HW heater as the backup boost for an active solar thermal system but only had 220 gallons of water as thermal buffer- more than an order of magnitude less thermal mass than just slab, not including the walls. On any thermostat bump-up the buffered heat would almost certainly be depleted, causing the tankless to kick in, increasing the purchased-energy use. (Running the solar at 130F also cuts in to efficiency pretty dramatically too- massive but lower mass radiant slab would have allowed them to drop that operating temp considerably, boosting the collection efficiency.) The high current draw of the electric tankless also causes the inverter for to the PV array to kick off intermittently, reducing their PV intake and adding to the purchased power.

By trying to integrate too much into the system they ended up with a kludgy Rube Goldberg contraption with unforseen interactions, which is likely a larger issue in overshooting the energy use goals than simply controlling the mass of the building. But the simulation of the building assumed no temperature setbacks/recovery ramps, but the analysis implies that the occupants weren't living up to that assumption (how DARE they? :-) ), which undoubtedly contributed another layer to the increased energy use of their Rube Goldberg heating system. "Training" the occupants to use the mass and heating system as intended by the designers might improve performance, but they have a lot of interactive issues to weed out in that system before it would actually come close to the modeled performance. Using a heating & cooling heat pump (rather than AC-only) to provide the backup heat to the solar thermal may have been a better choice, but if it cut into the AC efficiency by much, maybe not.

CA zone 11 has about 2x the heating degree days that Tuscon does, and half the cooling degree days, but has mild enough heating design temps that it can be heated very efficiently with variable-speed compressor air-source heat pumps. It also has low enough latent cooling loads that one COULD use radiant slabs for both heating & cooling without risk of condensation on the floor. There aren't a lot of choices for air-to-hydronic in the US (there are more choices in Europe), but the Daikin Altherma has been here for a few years now, and is developing a track record. In CA climate zone 11 it would challenge geothermal on whole-system efficiency, but is substantially less money up front.

You might price out a minimal-R ICF against CMU construction too. It's more expensive materials-wise, but it goes in more quickly and is easier to air-seal reliably. In larger & more complects projects the scheduling costs can exceed the difference in material cost between the two systems. About the lowest-R you can get with ICF is ~ R16-18, (R22 has become the standard min for many manufacturers , abandoning the lower R versions in response to higher code-min requirements.) The interior foam of ICF isolates the mass from the conditioned space somewhat, but the thermal mas of the floor is fully accessible to the house, lowering the impact of that less-available mass in the ICF. The wall mass in an ICF still buffers the heat gains/losses through walls, even if it's not buffering the whole house as much as an exterior-foam only approach. Construction time wise it goes in a lot faster than CMU + exterior insulation, which may be worth the premium in a "time is money" analysis.

IIRC, the result in the tucson house was no net gain from thermal mass, which is substantially poorer performance than the adobe house next door. That leaves us two choices: the designers weren't capable of calculating heat loss, or they chose the wrong method. Yes, the slab design ruled out radiant. One suspects that today, understanding the importance of heating all that mass directly, they'd find a more accommodating foundation design.

On P27:

"The first year of operation, which had the home open as a model for nine months, reached 75 percent of
at goal. Thus far, since the home was occupied in October 2003, the net energy purchase
is about 30 percent with the cooling season remaining."

Operated as-designed during the open house it nearly met the net-zero energy use goals. Operated by real occupants it only made it to 30% below the paradigm case next door, which is CMU, not adobe. The authors explain in some detail some of the systems-reasons for the shortfall, and speculate that by using setbacks and ramp-ups the extreme mass of the building probably increases energy use and reduces comfort levels (see thermal mass discussion on p.29.) With the systems they have in place, as-used by the occupants the high mass is a net-negative. Higher R and constant temp would likely fix those issues evenwithout heating the mass directly, but it's true that a lower mass slab radiant floor would make the place more comfortable even at lower temps, and would likely further decrease the energy use of the building. (The active solar would have much higher output too.)

Using an electric tankless as backup heat to medium-temp active solar in a high mass house is kinda brain-dead design, when lower temp higher-efficiency solar could have been used for charging the slab, with air source heat pump hot air as the backup default. That sort of backup makes sense for domestic hot water, and in some low mass houses, but proved to be a not well thought out misapplication here. For about the same money on mechanical systems it could have been far more efficient in heating mode. But I'd have to step into the time-machine to see if the appropriate heat pumps were available at the right output & efficiency back in 2002 or earlier when they were designing and simulating the place. The used a high SEER 2-stage AC, and (apparently) figured that being a cooling dominated climate, that was more important than the heating system efficiency in terms of the total energy use impact. But relying on resistance electric heat (the tankless) and operating the active solar in a less-efficient mode proved to be an Achilles heel in the design.

Nowhere do they describe an adobe house with energy use that's more than 30% below the CMU house paradigm though.

Like Jonr quoted, "it isn't apparent that (thermal mass) is a benefit at ZEH at APdS." Thermal mass is what we're discussing here, rather the collectors bolted to the roof. And unless the builders and owners of adobe houses have suffered mass delusion for centuries, thermal mass is in fact quite beneficial in Tucson if you don't get too cute. The report said thermal mass worked better than expected in cooling season so the heating strategy must have been well beyond "kinda brain dead." If it worked on paper, either the math was wrong or the system was far less efficient than expected.

The rest of it is hysterical. The ZEH designers rework thermal mass for modern set-and-forget sensibilities. So the first thing the owners do is race over and twist the thermostat dial. Can't blame them On night 3 of bitter cold Tucson style (45?) the missus is saying I want major heat now, like ... .like ... setting some logs afire and snuggling up. Perception is everything, eh?

I've been in Tuscon when there was frost on the ground- that clear desert air is good for quite a bit of radiational cooling overnight for fairly high diurnal temperature swings. The 97.5h percentile heating design temp is +32F (99th is 28F). The January mean low temp is about 40F- the peak-coolth is colder there than you might think, even if the mean temps are pretty mild. (I've seen frost at altitude in the Santa Catalinas in August, on days when the highs were in the high 80s.)

But the solution is to sleep-in until it warms up. Winter only occurs between 3-7AM, after that it's pretty spring like! :-)

They explain in some detail (complete with graphs) how operating the tankless was bringing down the inverter for the PV, increasing the amount of purchased energy, and it's clear they gave the whole heating setup short-shrift in the design phase. It looks like they grafted a standard design for solar DHW onto a mid-temp hyrdro-air situation without fully simulating the "what if"s of what that means if the occupant chooses to bump the house up a degree or two when there's an order of magnitude more thermal mass in the conditioned space than in the solar buffer. At constant temp the active solar probably handled a significant fraction of the heating load despite being sub-optimally high temp for good collector efficiency, but that could get killed that pretty fast with regular bump-ups/set backs. The also had to peel back ~20% on collector area (8' panels instead of 10s) to cater to local aesthetic sensibilities ("Thou shalt have no solar panels visible from the street") and it's not clear that they re-simulated after that adjustment. With a radiant floor and consequent lower operating temp they could probably close to double the solar uptake during the heating season, which would have a dramatic reduction on the duty cycle of their electric tankless. Even in a higher-R house (at those levels) in Tucson the annual heating energy use would far outstrip the DHW energy requirements, and (with 20-20 hindsight) the active solar design wasn't even close to optimised for effiecient space heating.

I've seen similar active solar combi systems using a gas or propane fired tankless as backup with hydro-air output and those too are sub-optimal for space heating but common as retrofits to homes with pre-existing duct. Low temp hydronics delivers higher solar efficiency, but at lower than DHW temps. The fact that they were pouring a slab in the first place seems like they missed the boat, but they state the document it was considered & rejected- they didn't want to pay the per-square foot cost of going that route and they were already committed to the 10" slab (for reasons unstated), which is clearly too thick.

The alternative to sleeping in, in a place where winter happens between 3 and 7 am, is to live in a high mass structure that buffers wide swings in diurnal temps. If thermal mass reduced cooling demand more than expected kn Tucson, why would it fail so miserably in the (minor compared to cooling) heating season that unexpected energy consumption wiped out the summer gains? Short cycling is my guess.

There is no loss of efficiency with short-cycling an electric on-demand hot water heater, or the air handler for that matter. The explanations lie elsewhere (many of which are in the document- read it again!)

And the high mass of the house REDUCES short cycling rather than increasing it.  The thermostats used for low & mid-mass heating systems have a 1-2 degree F hysteresis (some have adjustable hysteresis, most do not.) The thermal mass of the house being fully inside the insulation increases run times- it takes longer to increase the air temp by the amount of hysteresis with the massive floor & wall surface areas are soaking up the available heat from the air.

There aren't a lot of choices for air-to-hydronic in the US

I agree, although I haven't seen an explanation as to why one cannot use a water->air geothermal heat pump in reverse. Ie, run outside air through the duct.

Vexingly, the explanation about the failure of thermal mass is not in the report. That the tankless ran more than expected and chewed up PV output does not explain why "it isn't apparent that thermal mass was a benefit" when, in fact, it actually worked better than expected in the season that counted. Gotta say it helps to have some experience in playing catch-up in a high mass structure. While the mass may keep the air handler running longer, in the end, it has heated a millimeter or two of the eight inches in question, and five minutes later the blower is running again. And again. And again. And again. I doubt that the solar hot water part of this system would fare well in this regimen.

All that aside, my initial caution was to make sure you had hvac equal to the elephant in your house. And because you WILL play catch up at some point, because of power outages, vacations, or in my case leaves that didn't fall until mid Nov, I would recommend a radiant slab as well.

Thanks for the input, especially Dana. RE; the NM slab; I haven't read the documents, but if a slab is doing a good job cooling, but not heating, my first guess is it isn't insulated well. Of course, if there isn't passive solar input, nd the slab is thick, it is going to be cold all the time. High mass designers would do well to read Hait's booklet about PAHS. It talks about the balance required between windows, sunlight and mass. My immediate take is better to have too much window than not enough. You can shade windows in summer, and use thermal shutters at night in winter. As they have found in NM, it is hard to add windows after the house is finished.

Posted By McFish on 13 Jan 2012 08:58 PM
Thanks for the input, especially Dana. RE; the NM slab; I haven't read the documents, but if a slab is doing a good job cooling, but not heating, my first guess is it isn't insulated well. Of course, if there isn't passive solar input, nd the slab is thick, it is going to be cold all the time. High mass designers would do well to read Hait's booklet about PAHS. It talks about the balance required between windows, sunlight and mass. My immediate take is better to have too much window than not enough. You can shade windows in summer, and use thermal shutters at night in winter. As they have found in NM, it is hard to add windows after the house is finished.

They went with 2" of iso (~R12) as oppose to their standard (and near code-min) 1.5" (~R10) on the walls.  The ceiling was R41 in the form of fiberglass batts (batts are always a suspect choice when shooting for high-performance, since they demand perfection to get any performance out of them) plus radiant barrier with an air gap of unspecified depth.   They had simulated with R42 cellulose, but the builder balking citing anxieties about "moisture issues", but they didn't elaborate. I'm sure it would have performed better with the denser blown ceiling/attic insulation than the batts they went with, but the size of that factor can only be speculated, but it's probably not the straw the broke the camels' back.

This is definitely not a high-R building- they figured they'd be able to get the performance by managing the thermal mass and active solar inputs well, all of which had issues.  The inverter kept resetting and timing out when the tankless HW heater they were using as an electric boiler kicked on hard, knocking huge chunks out of their net PV input during the heating season, and for local-community aesthetic reasons they cut the size of the active solar thermal array by 20% to keep it less visible from the street.   The list an infiltration assumption of 0.35ACH/natural, but no test data to support it- I'm assuming they didn't bother to test & remediate. That could also be a wild card in the heating season performance.

The consciously and specifically opted NOT to make any design changes to optimize passive solar (so as not to change the exterior appearance of the house from others in the neighborhood?) assuming that in a cooling dominated climate optimized cooling season performance was paramount.  That probably accounts for much of the heating system kludgy-ness (that's the precise engineering term for it ) and lower/less-predictable heating season performance.

More wall R may have improved things, along with some minor tweaks on glazing for solar gain, but the $64,000 question is what would have happened if they'd air sealed the place to a specified (and lower) level than their stated assumption number. (And it's not too late to improve that 0.35ACH/natural number even as retrofit- it's not very low number at all. It could even be leakier than the IRC 2009 code min of 7ACH/50 and still be 0.35ACH/natural.)

All things considered they did pretty well to come within 75% of meeting the design goal during the demo-period, but it's not surprising that it underperformed by an even deeper amount during the heating season with those pesky humans fiddling with the thermostats and messing up their design. :-)

The house couldn't be passive solar because it was in a subD with other houses close by. I would guess that a lot suitable for passive solar is relatively rare. The overactive tankless was clearly symptom rather than cause. The pesky humans could well be symptom as well. (i.e. they twisted dials because they were cold, or because the power bill was too high.) From what Dana writes, it's possible that the designers COULD NOT calculate a heat loss. They aren't telling, and we don't know.

At any rate, don't finger the slab, McFish. Because houses have 1.5 to 3 times more wall surface than floor surface, there is more concrete in the walls than the slab, even at 10 inches. Adding them up, you have a LOT of mass.

The overactive tankless was a direct cause of reduced PV input due to unhappy interactions with the PV inverter. When you're going for net-zero shutting down the inverter taking the PV off line regularly cuts holes in the modeled NET energy use by cutting into the available output. They claimed that loss to be "relatively small", but didn't quantify it.

I suspect the 20% hit in thermal collector area was a bigger factor in poor heating season performance. If you look at table 11, the modeled solar energy uptake was 80-90KBTU/day , but the measured was about 56 K. This is significant in face of the fact that the modeled heating energy use was ~30KBTU/day and the measured was ~47KBTU/day, a 17K overshot on energy use, coupled with a ~30K undershoot on energy uptake. Better air sealing & attic insulation could be enough to make the difference on the energy use end, but something was definitely amiss on the active solar thermal design. The 20% undersizing as-installed relative to the model sure didn't help, but doesn't account for the total shortfall, which was more than 30%. (Arguably the collector efficiency went up slightly due to lower operating temps when the array size shrank, and a 20% reduction in area should have been only a 15-18% reduction in thermal output, not 30%+.)

The paper implies that thermostat tweaking was a problem, but they don't quantify that either. The modeled energy use was "set and forget" on heating & cooling T-stats.

And yes, concrete filled 8" CMU is a lot of thermal mass, and probably 2x that of the slab, but they don't break it down anywhere in the document. It's only a single-story which limits the wall-area/slab-area ratio, but it appears to have a relatively low fractional area of glazing & door, looking at the pictures.

Not that it matters much at this point, but my curiousity is aroused; do these documents mention any slab insulation at all? If cooling performed well, but heating was poor, that seems like an obvious possible reason. I spec'd a vertical insulation several years ago, forgot to be there when the work was scheduled. Subs had never heard of it before, figured it was going to be covered up anyway. They only installed 12 inch , only about 3 inch below grade. Worked great all summer, why not, earth soaked up all the heat that went into the slab. Heat gain in winter was much less than expected, until I had a sewer problem and discovered the short installation. It worked pretty good after extending the coverage another 12 inches into the ground, with a wing of 6 feet.

McFish said:"Not that it matters much at this point, but my curiousity is aroused; do these documents mention any slab insulation at all?"

For the house in Tucson, it looks like ground water temperature is about 68 or 70 F there (http://mb-soft.com/solar/soilmap.gif), so not sure what the custom is in that area, but I guess slab insulation would not be real critical there.