Bob Holodinsky was hoping for a better outcome from the heat loss calculations he received for his new Peterborough, Ontario, home — calculations that appear to have upset his plans for heating with a ductless minisplit. “I thought I was on the right track,” he writes in a Q&A post at GreenBuildingAdvisor, “but now I am not so sure.”
The 1 1/2-story, slab-on-grade house will be built with 2×6 exterior walls, insulated with blown-in fiberglass and 2 1/2 inches of expanded polystyrene insulation (EPS). Conditioned space totals about 1,900 square feet.
There are lots of triple-pane windows on the northeast side of the house, which faces a lake. The living-dining-kitchen area alone is responsible for a heat load of 17,000 Btu/hour, and the house as a whole shows a heat loss of 42,000 Btu/hour.
Holodinsky’s builder says that insulation in the roof will equal R-75. Exterior walls will have R-24 insulation between the studs, with an additional R-10 in the exterior foam sheathing. The slab is insulated with 3 inches of EPS.
“I was hoping to heat it with a Fujitsu minisplit on each level,” he writes. “Is that still a viable option?”
Holodinsky’s misgivings are the topic for this Q&A Spotlight.
Heat pumps are not the right option
Although he can lower his energy demands by air-sealing and increasing the amount of insulation in the house, minisplits may not be the best fit for this Climate Zone 6 location, writes a GBA reader who calls himself Flitch Plate.
“I would be concerned with minisplits as the sole heat source in Peterborough, no matter what the BTUs are,” Plate says. “Heat pumps are the poorest performers and least efficient at the times you need them most: very low temperatures. I would design in a backup heating solution, such as a small wood stove centrally located on the main floor or an electric heat source in each room.”
Jason Hyde says he’s not aware of any Fujitsu units suited to Peterborough’s winter design temperatures. Fujitsu says that its XLTH series works down to 15 degrees below zero, Hyde says, but these units produce only 18,000 Btu/hour at 47 degrees F. “I can’t find anything telling me what they will put out at lower temperatures,” he adds. “In any case, two units is still only 36,000 Btu/hour.”
Given these numbers, and without seeing plans or heat-loss details, Hyde doesn’t see any single upgrade that will put him within Fujitsu range.
Too much glass, not enough foam
One problem, says Matthew Nolette, is that 2 1/2 inches of EPS aren’t enough exterior foam for that climate zone. He points to an article by GBA senior editor Martin Holladay on calculating the minimum thickness of exterior foam. However, extruded polystyrene (XPS), with a higher insulating value than EPS, would work at that thickness.
Another problem is all that glass on the northeast side of the house, which Nolette says “would be indicative of design desires conflicting with energy goals.”
Passive solar building guidelines suggest minimal glass on the north side of the house because windows contribute no direct solar gain but plenty of heat losses. Windows on the east and west sides of the house also should be limited.
“You’re in a tough site with northeast views in a cold climate; there will be some performance compromises if you cannot determine how much northerly glass is a lot and how much is enough,” he writes. “Has your builder completed a house with these design goals before? Are they designing it? Do you have access to the designer and are you using a third-party energy modeler?”
Interpreting those pesky heat loss calculations
Yet another complication could be the energy loss calculations themselves, which were produced with a program called Wrightsoft. The report is complex and “more than a bit confusing for me,” Holodinsky says, but it seems to predict 0.48 air changes per hour.
“The assumption is way high and suggests a very leaky house that would have a blower door test of 6+ [air changes per hour at 50 pascals of pressure],” Jerry Liebler replies. “With minimal attention to air sealing, you should be closer to 1.5 ACH at 50 Pascals. Ask your builder to do a blower door test and ask for a guarantee that it’s below 2 ACH @ 50 pascals. Then ask that the infiltration rate for the heat loss calculations be 2/12 = 0.1666 ACH .”
In fact, Liebler adds, actual air infiltration is only a fraction of the test number because the test “artificially creates a greater pressure difference than actually will occur.”
“With the 2 ACH test data the heat loss due to infiltration should be about one-third of what was calculated,” Liebler continued, “By the way, the proverb ‘garbage in = garbage out’ applies. I’d be very suspicious of all the heat loss calculations and assumptions. For example, were the U-factors of the windows you are paying for actually used, or did they use something else? Triple-pane windows should be U=0.2 or less.”
We’re smack dab in the middle of the Information Age, adds Sonny Chatum, so Holodinsky should consider going online and getting a copy of the BEopt software from the National Renewable Energy Laboratory and running his own calculations. Chatum used the software himself to predict the heating and cooling loads at his own house after an energy retrofit (losses dropped from 40,000 Btu/h to 14,000 Btu/h, which Chatum handles with two ductless minisplits).
“It is useful to note that sometimes experts and professionals get hung up in their traditional ways,” Chatum added. “Sometimes, and in some ways, the Information Age can possibly make you as much or more of an expert than ‘the expert’ … All the traditional contractors were astonished with what I had done, and did not believe that the minis would be adequate. They were wrong.”
Sharpen your pencil, indeed
When Holodinsky tracked down the energy modeler who worked on the report, he found that a series of miscalculations, sloppy inputs, or erroneous assumptions skewed the numbers significantly.
The overworked technician, working on four computer screens at once, listened to Holodinsky’s concerns and made some adjustments to keep program inputs much closer to what the house would actually look like when complete. By raising the R-value of exterior walls, adjusting specs for window performance, and changing a program setting for airtightness, heating loads dropped by 10,000 Btu/h, to a total of 32,000 Btu/h.
“Much better,” Holodinsky said.
But maybe not good enough.
“The heat loss was high and the designer was sloppy from the details imparted here,” writes Morgan Audetat. “DIY heat loads for modern construction with ERV, triple-glazing, and rigid foam in cold climates is not for the amateur or the harried professional.
“If you are building your home around your heat source, you need a proper ACCA Manual J. If your builder will agree to a blower door and a specified ACH — good luck with that — then that number can be plugged into the Wrightsoft program and all your worries are over…”
Our expert’s opinion
Here’s what GBA technical director Peter Yost said:
There is not much left for me to deal with here, except to echo these points:
Additional or “backup” heat source. Taking a look at some historical and recent winter weather data for this neck of the woods, it’s clear you need to have another heat source in addition to the cold climate minisplit heat pump you have cited.
From the WeatherSpark website: “The coldest day of the last 12 months was January 22, with a low temperature of -26°F. For reference, on that day the average low temperature is 8°F and the low temperature drops below -11°F only one day in 10. The coldest month of the last 12 months was February with an average daily low temperature of 1°F.”
And for a “down Maine” confirming testimonial, take a look at the info here from Efficiency Maine.
Heat load calculations. These are very dependent on the assumptions regarding data you might have on airtightness. Simple programs like Manual J may have a three-level category for airtightness — leaky, average, and better — that is wholly inadequate for higher performing building enclosures when 25% to 40% of the heating load in a climate like this may be due to air leakage. The long and the short of it: get a blower door test to nail down this portion of your heating load.
Undersizing any space conditioning equipment isn’t good. But oversizing is less of an issue for heating equipment than it is for cooling equipment. Heating appliances that are a little bigger than necessary provide a quick recovery from deep overnight setbacks, and that’s not such a bad thing.