Use of actual windows as a test platform has the disadvantage of being dependent on cold and varying ambient temperatures, so quick testing of many different materials is impossible.   The use of air temperatures in the first two studies is another problem.  These temperatures are always subject to convection currents, and the proper location for an average temperature between the window and the shade changes with circumstances.  One of the internet studies showed the significant effect of sensor location.

My previous tests used surface temperatures for most sensors. Each sensor was attached to a small aluminum plate to eliminate the effect of the wires, and to partially average the temperature. These sensors were attached to the plastic cover of the “window” and on both sides of the hot face foam “standard” panel. This system works well and will be used on the new design.

My previous testing, while repeatable, had several shortcomings.  Since I had no heat source except the room, the warm side temperature facing the test piece (behind the standardized foam plate) was below room temperature.   In addition, the heat inflow to the window frame was not originally accounted for, and contributed to significant errors at higher R-values. This must be designed out of the new design.

Proposal for a new “Standard”

No window covering test method exists because there appear to be too many variables, and the role of convection currents is the paramount source of confusion.  This is a proposal for a standard test method using a cold face (simulating the window) and a warm face (simulating the room).  The idea is to introduce a standard method that can provide a base line for R-values and a process using Computational Fluid Dynamics (CFD) to add correction factors for the many variable factors that occur with windows coverings.  It should be noted that this process will NOT generally be useful for face-mounted curtains, which can easily have negative R-values depending on their installation.

The role of boundary layers and convection currents are unique with window shades.  For most construction materials the R-value must exclude these effects because they disappear as soon as they are brought in contact with another material.  With window coverings these boundary layer effects are a major part of their effectiveness.

For most materials a “guarded hot-box” is used to measure the R-value, and it effectively eliminates the boundary layer effects.  For the window-covering test-unit the cold walls of the window frame on the cold side of the test unit are the major source of heat inflow that causes problems.  This region is about 2.5 inches wide (depending on installation of the test piece).  A “guarded hot-box” (actually a guarded cold-box) around this area may be necessary to get accurate R-values with high R-values test samples.  The problem with high R-values (>8 or so) is that the amount of heat flow through the test unit is small. The cold side walls of the test unit are then very cold so the delta T to the outside is high, and it takes only a little outside heat to affect the calculated R-value.  So, the guarded-cold-box around the cold area of the window frame is needed.  The problem can be reduced somewhat by making a bigger “window” area (area versus circumference), but a bigger test unit is more awkward and expensive to build and test.