After many different experiments and configurations, I think I have reliable data on the performance of insulation in a typical Airstream shell configuration. Some of you may have read the thread where I tried to do four test panels at once--that data (or the analysis strategy) didn't work out. So I built the single panel test rig here:
All the panels were identical--1-1/2" thick and 12x12". The insulation was installed as you would expect, the glass mat was placed to ensure there were no gaps in the corners and the sheets of foil were glued in place, ensuring an airtight seal and supporting the foil directly between the face sheets, with a 5/8" air gap on each side.
The test setup was an old PC-XT running a single chip 12-bit A/D converter through the parallel port. The sensors were Kelvin devices so there was very little electronic noise--precision was 0.3 degrees F and accuracy was on the order of 0.5 degrees F. Samples were taken every two minutes.
The charts below show the absolute temperatures of representative test runs on the left. The right hand charts in each case show how close the panel temperatures were to their environment (the inner panel to the water and the outer panel to the outside air) divided by the temperature difference between the water and the outside air.
The data was analyzed in two ways. First, how fast did the water bottle lose heat? This should provide some idea of the R-value of the insulation. However, the heat lost throught large surface area of the five other insulated walls of the box was, in total, more significant than the heat lost through the test panel, so the data show only small differences between the panels. In any event, a smaller absolute value of the slope means better insulation performance.
Second, what was the ratio of the inner and outer panel temperatures to the total temperature difference between the outside air and the water? It seems intuitive that if the inner panel is relatively warmer, then the insulation is doing a better job. What this means is that lower numbers are better (if the percentages were zero, that would mean zero heat flow through the panel--the inner panel temperature would be equal to the inner environment and the outer panel temperature would be equal to the outside air).
You have to get the full size images (click on them) to see all the lines on the charts.
When you look at the charts, it's obvious that an empty panel (air only) has worse performance all the way around--higher heat leakage and colder inside panel.
The glass mat performance is good. Relatively warm inner panel, but the heat loss number is higher than for both of the foils. Remember, lower numbers in the right hand chart are better.
The Home Depot foil (in visual appearance it has two layers of large bubbles between the foil) performance as well as the glass mat and had lower heat loss.
The "Prodex" (
Prodex TOTAL Insulation - Aluminum-Foam-Aluminum from Insulation4LESS.com) was a puzzle. The inner panel was relatively cold, but the heat loss was very close to that of the "HD foil." I dissassembled both of these panels to check their construction, reassembled them, and got the same data in a second test.
The bottom line is that I still have open questions. First, it's obvious that the inner air temperature isn't the same as the water--lots of heat (relatively speaking) is going out through the walls of the box, so the air temperature is quite a bit different than the water.
The next series of tests will include an inside air temperature sensor, which will give more precision to the temperature ratio test (in about a week--it usually takes one night per panel test, unless it's overcast and cold during the day).
Second, will the foils perform even better when the summer sun is heating the outer panel to 160+ degrees? I'm more interested in hot performance than cold, so that test is coming this summer.
Zep