Yesterday, I made my first set of thermal images of the Mansfield House. It seemed like an opportune time to do so; we’d had rain and cold weather for the better part of the past week. But yesterday, the sun came out, and it got quite hot. The house was being warmed by the sun and drying out, so I thought I might get some interesting images under these conditions. Here’s one of the first infrared images I took, of the front of the house, from a position just on the opposite side of Jewett street:
Brief Primer on Infrared Imaging
For those readers unfamiliar with thermal imaging and its use in building analysis, here’s a very brief primer (I promise to be uncharacteristically gentle with the physics stuff ):
All physical bodies hold some quantity of thermal energy (or, heat), and have a tendency to either emit or absorb it, depending on their surroundings. What we commonly call temperature is simply a relative measure of the potential for thermal energy to be given off from, or absorbed by, some particular object (or more precisely, to be transferred between two or more objects, either directly or via some common medium).
One manifestation of thermal energy is radiation in the infrared range of the electromagnetic spectrum. An infrared thermometer is a device that can estimate the temperature of a small point (or “spot”) on the surface of an object without touching it, by measuring the infrared radiation emanating from that point, while also making certain assumptions about the material composition of the object itself (in technical jargon, the object’s emissivity).
A thermal imager (or infrared camera, or IR camera) takes this a step further by measuring many such small points and using visible light colors to denote different ranges of estimated surface temperatures at those points. When these colors are assembled into a spatial representation of the object, the resulting image provides a picture of the relative temperature differences across the surface of that object. For example, in the Mansfield House photo above, white-yellow (roof) and yellow (siding) depict relatively warm surfaces, while pinkish-blue (chimney) and blue (windows) indicate relatively cooler surfaces.
The key point here is that these images are not formed by reflected visible light, as in conventional photographs, but rather from the relative temperature differences across an object’s surface, as measured via the emission of infrared radiation. And yet, IR images often convey details very similar to those of visible light photographs. For example, in this second photo of the Mansfield House, there are very distinct “shadow lines” revealing the exterior cladding. However, they’re not caused by shadows, but rather by the fact that the bottom edge of each clapboard receives less sunlight than its face, and hence is slightly cooler. The same is true of the attic overhangs:
Why would anyone want to make thermal images of buildings? Well, there are quite a few good reasons. Being able to visualize and measure differences in the relative temperatures of building components can quickly reveal a wide variety of problems or inefficiencies. For example, a home inspector or remediation specialist can quickly find evidence of water damage and potential points of infiltration. A home performance professional can readily determine air leaks and insulation problems. And an electrician can discover circuits or equipment that are running too hot.
In my case, I’m interested in all of the above, but I’m also researching how very early American homes (17th and 18th century) performed “as systems”, in their original states (or what we best understand nowadays to have most likely been their original states), with the objective of creating normative models of this behavior. The thermal analysis of surviving early homes, therefore, is quite invaluable to this effort.
Returning to the topic at hand, let’s first take a look at what’s going on inside the Mansfield House attic, the hottest place in the house. [Note that in much of what follows, I'll give temperatures in Fahrenheit for the benefit of readers less subjectively familiar with the Celsius scale. But don't tell my good friend Allison Bailes; he's likely to take issue with my approach.] Now, here’s the Mansfield House roof system and chimney column:
I deliberately measured the temperature of the gap between two roof planks on the east-facing rear roof (which gets a bit less sun than the west-facing front roof), and got 124 °F. By comparison, a point on the chimney column, just a few feet below the roof, measured a mere 88 °F.
Measuring the front roof (which gets maximum sun exposure) revealed some interesting temperature differences between sheathing and rafters. A spot on one of the planks was a blistering 129 °F, while the bottom side of a nearby rafter only 102 °F, giving a notion of the thermal break afforded by the thick wood (these rafters are about R-6, for you serious energy geeks). What I like most about this photo is the obvious progression of decreasing temperature, from the top of the rafter to the bottom, as revealed by the changing coloration:
Here’s an image of the north gable framing. Again, the detail is pretty remarkable; even the BX cables running along the rafter are plainly visible:
The point where the rafter joins the overhanging tie-beam was only 84 °F, while the decking just above was about 107 °F. This area of the house has the least exposure to sunlight, given both its orientation, as well as the fact that it’s partly sheltered by a shade tree.
Here’s a final attic image, showing one of the rafters of the rear roof, in the vicinity of the chimney. The small object on the bottom side of the rafter, just under the crosshairs of Spot 1, is one of the temperature/relative humidity dataloggers I maintain through out the house. In fact, Spot 1 is pointed right at the datalogger’s sensor:
The temperature at the sensor, according to my IR camera, was 102 °F, while the bottom of the rafter itself just about 99 °F. On the other hand, the glowing little “hot spot” at Spot 2 is a hole left behind by a knot that had fallen out of the plank, and roofing paper behind the plank is exposed here. It measured about 131 °F (again, this gives one a notion of the insulating properties of about seven inches of oak). When I subsequently downloaded the temperature measurements of the datalogger, I found a recorded temperature of about 98 °F for this location, a little less than an hour before I’d made the image.
Now, here are some interesting images of the rear roof that reveal something about how these early homes had once performed thermally (and perhaps still do, to the extent they haven’t been overly modified).
The first image below shows the south end and center of the rear roof, just below the chimney. You can easily discern up to four rafters under the deck, including two on either side of the chimney, and one butting up against the center of the chimney (the small object protruding from the central rafter is a roof jack left behind by the last contractor who worked on the roof). This image reveals that the roof is slightly cooler between these three rafters than it is on either side of the chimney:
But why is this the case? Well, it turns out that this is a section of the house known as the chimney bay, which is mostly enclosed by walls on either side, and houses the massive, stone chimney column. The chimney column doesn’t simply descend vertically below the roof, but also slopes backwards, roughly following the contour of the rear “lean-to” roof, so as to provide a flue for the kitchen fireplace, the largest fireplace in the house. An example of this can be seen in the following illustration:
The chimney column, together with the large stone basement upon which it rests (this is where the term “basement” comes from, by the way), collectively function as a huge thermal mass. In the past, it provided considerable radiant heat to much of the house, once it was brought up to temperature via its fireplaces. And in the summer, its close connection to the ground meant that it could usually remain cool enough to draw some heat away from the rest of the house, dissipating it via the ground. I believe that (to a reasonable extent, anyway) the above thermal image reveals the cooling effect that a large, dormant chimney column can have on the rest of the home.
Two things I currently cannot account for in the above image, however, are the rather large hot spot (Spot 1) toward the bottom/left, and the small blue, cool spot (Spot 3), at the bottom/middle, of the image. I suspect Spot 3 might indicate a roof leak, or perhaps a small area of missing sheathing. As the roof dries over the next few days, I’ll check to see if this spot shrinks or disappears.
This next image shows the center and north end of the Mansfield House rear roof:
What’s interesting here are the large, rectangular warm areas. This portion of the roof covers the north garret, which is a small utility room with a sloped ceiling defined by the roof. There’s a garret on either side of the chimney bay. The north garret’s current “ceiling” is comprised of 2″ foam panels, a relatively recent modification to the home (though by whom, and when, I know not). The foam panels are fastened directly to the roof planks, appear to begin at the top of a small, exterior knee wall, and terminate at the rear top plate, a 6×9 timber beam that also seats the rafters at their mid sections.
Obviously, the foam panels are inhibiting the transfer of heat downward, into the north garret, so the roof surface remains quite warm. The ceiling of the south garret, by comparison, is comprised of plaster-on-lath, with the lath fastened to the rafters. The resulting space between the plaster ceiling and roof sheathing is about 4″-6″, and it’s unfilled and uninsulated. As a result, the roof over the south garret is warm, though not quite as warm as it is above the north garret.
Water Damage and Mold
Next, I turned my IR camera toward the heavily water-damaged portions of the rear wall (some of you might recall I wrote a post about this situation, and its causes, not too long ago). The image below is of the kitchen door. The IR camera reveals that the water damaged lower panels and bottom rail of the door have a temperature of about 73 °F, in contrast to the 86-90 °F temperature range of the drier upper panels and siding. I claim that this lower temperature of the damaged areas is due to evaporative cooling of the saturated wood. Same goes for the mold on the siding on either side of the door:
This next image is a close-up of the bottom rail, lower panels, threshold, casings, and siding. Extreme evaporative cooling, all around:
Here’s yet another IR image of the damaged kitchen window sill, and the excessive mold on the siding just below it. There’s a 7 °C / 15 °F temperature difference between the rotted sill and the dry siding right next to it.
A nice capability of my particular model of IR camera (a FLIR E60bx) is its ability to compose picture-in-picture images, where a portion of a visible light photograph can be rendered in infrared. This capability is particularly helpful in fixing the exact locations of problems revealed by the infrared imagery. Here’s an example, using the same scene as in the previous photo.
Pop Quiz on Infrared
I hope all of you have found this article both interesting and informative. If you’re intrigued and would like to read further, I heartily recommend a number of excellent articles on infrared imaging and testing, authored by my good friend and advanced infrared thermographer, Sean Lintow Sr. of SLS Construction.
Now, here’s a pop quiz for you. I’ll award two extra Scooby Snacks to the first reader who posts a comment providing a correct answer to the following question:
Since infrared thermal imaging doesn’t involve visible light, how is it that the Mansfield House sign is so clearly legible in infrared?
As a hint, you might want to take a look at a visible light rendering of the same sign: