Editor’s note: This is the last in a series of guest blogs by Reid Baldwin about the construction of his house in Linden, Michigan. For a list of previous blog posts on GBA by Reid Baldwin, see the “Related Articles” sidebar below. You can read his entire blog here.
We have now been living in the house for about nine months. I would like to wrap up this series of articles by discussing the measured performance. I’ll start with the blower door test.
One of our goals during construction of the house was to make it airtight. Of course, no house is ever completely airtight. Some houses leak a little and other houses leak a lot. The standard way of measuring airtightness of houses is a blower door test.
A fan is installed into the front door so it blows air out. An equal amount of air comes in through whatever leakage paths are in the house. The technician adjusts the fan speed while measuring the pressure difference between the inside of the house and the outside of the house. That determines the rate of airflow required to establish a 50 pascal pressure difference. The air flow rate is compared to the volume of the house to calculate the number of air changes per hour, called ach50.
A low ach50 number is good. We had a blower door test performed on our previous house, which was built in the early 1990s. The ach50 was about 10. That house had many of the symptoms of poor air sealing: rooms that were uncomfortably cold whenever it was windy in the winter, for example.
New construction houses now usually have an ach50 of between 3 and 4. To be certified under the very rigorous Passive House standard, a house needs an ach50 of less than 0.6. Builders of Passive Houses go to great lengths to get to that level. For our house, the heating and cooling load calculations assumed an ach50 of 1.5, although I hoped for better.
Floor plan complicated air-sealing
The floor plan of our house is not ideal for air sealing. The fact that only some of the house has a second story and the ceiling in the hangar is midway up the second story meant that we had more wall/ceiling intersections to seal. We did a number of things to improve the airtightness.
The worst areas in most houses are the ceilings and the rim joist. To reduce air leakage through the ceilings, we:
To reduce air leakage around the rim joists, we:
To reduce air leakage through the walls, we:
- Taped the joints of the OSB sheathing.
- Caulked around the framing on the interior side
- Selected high-quality casement windows.
So, did these things work? Yes, they did. Our blower door test result of 0.82 ach50 proves it.
Verifying the heat load calculations
As discussed in a previous article on heating and cooling, the Manual J design heat load for my house is 34,000 Btu/hour at a design temperature of 7°F. To check this, I used a method of calculating the heat load based on fuel use as described in this GBA blog by Dana Dorsett.
The heating demand is satisfied by a combination of internal gains and furnace output. The contribution of the furnace is calculated based on historical fuel use data. To calculate the contribution of internal gains, this method relies on a guess at the home’s balance point, which is basically the outdoor temperature at which internal gains would keep the house at the design indoor temperature. A better insulated home will have a lower balance point than a poorly insulated home. Homes with a lot of solar gain or other internal gains would have lower balance points. Dana recommends guessing 60°F for a house with 2×6 walls and 65°F for a house with 2×4 walls.
Instead of simply guessing what the balance point was, I attempted to determine the balance point by applying Dana’s method to four different months of fuel use data. For each of the four months, I tried four different levels for balance point, 50°F, 55°F, 60°F, and 65°F. The estimated heat load gets lower as the presumed balance point gets higher. However, the slope is steeper for warmer months. (That is probably why Dana recommends using mid to late winter bills.)
The negative slope results from assuming that a lower proportion of the heat comes from internal gains as the balance point increases. (This does not imply that modifying the house to have a higher balance point would reduce the heat load.) I graphed the results (see illustration #3 below). In theory, the lines for the different months should intersect at the actual balance point and design heat load. In practice, noise factors like varying occupant behavior cause some divergence from the theory. Based on the graph, I believe the balance point is in the ballpark of 56°F and the heat load is in the ballpark of 28,000 Btu/h.
The lines for October to December seem to cross as expected. However, the line for January appears to have shifted. This could be due to noise factors or could represent an actual change in the heat load. I would expect some drop in the heat load because we improved the insulation in our attic around that time, as mentioned in a post I wrote in GBA’s Q&A forum.
In relying on the Manual J calculations, I specified a two-stage furnace with an output rating of 39,000 Btu/h on high stage, against the HVAC contractor’s recommendation. On mornings with single-digit temperatures this winter, however, the furnace ran continuously without satisfying the thermostat setpoint. I have not yet figured out why that happened. Is the heating load actually substantially higher than suggested by the Manual J calculation or the calculation based on fuel usage? Is the furnace not actually putting out its rated output capacity? (I did verify that the fan speed changes as expected when it is supposed to switch from low stage to high stage.)
One theory is that the small fan that goes with the small furnace is not able to distribute the heat throughout the relatively large house. I didn’t worry too much because the house never got uncomfortable and the temperature came up to the setpoint rapidly once the sun came out.
Another decision that the HVAC contractor pushed back on was my decision not to install a humidifier. On that front, the data say that I was right. The indoor relative humidity did not go below 30% over the course of the winter.
The Manual J calculated cooling load for the house was 18,000 Btu/h at 88°F. The smallest central air conditioner that was readily available was a 2-ton (24,000 Btu/h) unit. I haven’t attempted to systematically measure the cooling load. Anecdotally, I notice that the air conditioner runs about half of the time on the few occasions that the outdoor temperature has reached the mid 80s this summer. I wish I would have spent the extra money for a two-stage air conditioner. When only one zone is calling for cooling, it is sometimes uncomfortably cold close to a supply register. If we had a two-stage unit and a zone controller smart enough to use the low stage when only one zone is calling, that problem would be mitigated.
Out of curiosity, I installed a remote temperature sensor in the attic while dealing with the insulation issue. On a sunny day, it is not uncommon for the attic air temperature to be 20°F to 25°F warmer than the ambient outdoor air temperature. So, for cooling, the delta T through the ceiling is commonly three times larger than the delta T through the walls.
The relative humidity on the first and second floors seems to average about 50% in the summer without supplemental dehumidification. In the basement, the temperature stays in the low 70s without ever calling for air conditioning. The relative humidity in the basement often pushes 60% if I don’t run a dehumidifier.
During the process of designing, building, and living in this house, I have learned a lot about building science, HVAC, and the housing construction industry. Fortunately, I learned most of it in time to incorporate that knowledge into my house. There are a few things that I learned about too late in the process to apply. I got a much better result in terms of energy use and comfort than I would have gotten if I simply hired the usual cast of professionals to do what they usually do.
The cast of professionals that I hired also learned some things. It seems that only a tiny fraction of the industry is actively applying a lot of what experts know about how to build houses. Hopefully, my project made that fraction a little bit less tiny.