Historic Windows

Problems and Solutions

November 01, 2011
November/December 2011
A version of this article appears in the November/December 2011 issue of Home Energy Magazine.
Click here to read more articles about Windows

Windows are widely viewed as being essential to retaining the integrity of historic structures, and most historic structures have single-glazed, wood-frame windows. Yet in most climates, these windows waste energy and let in cold air. Modern windows that can improve energy performance fourfold while keeping the occupants comfortable are now widely available. However, unlike enhanced attic insulation or condensing furnaces, replacement windows are visible — sometimes very visible. So the question is, Should historic windows be restored or replaced? This is a major issue in much of the United States.

That's putting it as gently as possible. The truth is that there is a good deal of debate around this issue, much of which produces more heat than light. It was clear to my colleagues and me that there is a need for unbiased information to help guide historic preservation policy, and energy-saving decisions by owners of historic homes. So we decided to form a research team and enter the fray. The Center for Resource Conservation and the Synertech Systems Corporation teamed with Phoenix Window Restoration and Alpen Windows to work on a home in a beautiful historic district in Boulder, Colorado. Our goal was to demonstrate how to preserve the historic character of the building's windows while increasing the home's energy efficiency and comfort. We secured modest grants from several parties, including the city of Boulder and the Colorado Historic Society, and set to work.

We started with the hypothesis that the best decisions flow from accurate data representing a range of practical options. The 'accurate data' part involved testing a variety of windows, both in the field, in an occupied home, and in a laboratory facility designed to test both leakage and heat transfer coefficients. The 'range of practical options' part involved testing existing windows and old aluminum storms under a variety of conditions. We tested them (1) in the condition in which we found them; (2) after restoring them in several different ways; and (3) in combination with several varieties of storm window. These included storm windows fitted with state-of-the-art insulating glass. We also tested a new replacement window both in the lab and in the field. Then we took key findings and ran a series of simulations to determine how various combinations of windows would perform on each of four sides of an identical home in five American cities, each in a different climate area. This yielded lots of interesting data — and a potentially promising solution to the restore/replace dilemma.

Insulating and Air Sealing Counterweight Pockets in Double-Hung Windows

Unlike newer window systems, doublehung window systems have counterweight pockets on the inside of the frame. These pockets, which are not insulated, contain iron weights, ropes, and pulleys. They are frequently the source of substantial convective and conductive energy loss in summer and (especially) in winter. The combination of a large cross-sectional area and a thermally leaky frame means that even if the glazing is replaced with excellent insulating glass, the overall performance of the window system will be modest at best.

The study team explored two options for retrofitting counterweight pockets. The first option was what we call the spring solution, which eliminates the function of the counterweight pocket, replacing the weights with a spring mechanism and stuffing the pocket very tightly with fiberglass insulation. We employed this solution on the test home.

The second option was what we call the column solution. This option retains the function of the pocket and its weights while eliminating their shortcomings. It substitutes new rope for old and inserts a plastic tube into the pocket, inside of which the weights slide freely. It adds urethane foam to the remaining space around the tube. In both cases, the window is air sealed and insulated, and if the work is carefully done, the retrofit is invisible to the naked eye.

Counterweights are attached to the windows with ropes that pass over pulleys and down theinside edge of the window frame and sash. A knot is tied in the end of each rope. Holes are drilled in the two vertical outer edges of the sash to accept the knots, thereby securing the weights to the sash.

A plastic tube with an inside diameter just larger than the outside diameter of the counterweight is then inserted into the counterweight pocket. The weight slides inside the tube, while the rest of the space in the counterweight pocket is filled with an expanding-foam insulation, such as urethane. Cardboard and packing tape are used to seal up both ends of the plastic tube to prevent the urethane from penetrating the area where the weight slides.

The urethane is applied to the inside of the counterweight pocket in layers to ensure that the plastic tubes are completely surrounded. It expands considerably as it cures. Urethane sticks to almost everything, but not to plastic sheeting. Technicians wrap boards covering the counterweight pocket in plastic to facilitate postcure inspection. Then they clamp the boards down and test the movement of the ropes. Cure is complete in a few hours. Clamps are removed, the technicians verify that the ropes are functioning properly, and cracks are sealed.

Here's the Story

For the study, we chose a 108-year-old, 2,700 ft2 brick dwelling with patient owners who care deeply about both energy efficiency and their beautiful historic home. The home has numerous original double-hung, wood-frame, single-glazed windows, many of which still have the original glass, ripples and all. The study focused on three windows on the south facade and three on the north facade. Two of these windows measured 9.1 square feet each, of which 71% was glazed. Two windows measured 15.7 square feet (glazing 78%); and two measured 21.5 square feet (glazing 80%). Each of the six was equipped with a triple-track aluminum-frame storm window and a screen, which had been added several decades ago.

It is mostly professional craftsmen, working with specialized tools and equipment, who do the skilled work of restoring old windows. Some, like Phoenix Windows Restoration, whose technicians did most of the restoration work on this project, have a portable shop that they bring on-site, so that they can rebuild several windows at a time. The restoration process can breathe new life into old windows and substantially improve comfort and energy efficiency. However, because the work is painstakingly conducted by skilled craftsmen, it is much more expensive than buying a window system manufactured off-site and having it installed by the homeowner or a local carpenter.

For our study, Phoenix Window technicians carefully rebuilt five of the six windows in ways that retained their historical character. A custom-built wood-frame window that matched the aesthetics of the original but was carefully air sealed and designed to provide ventilation through a screen replaced the sixth, a small window in poor condition in a bathroom. Blower door testing revealed that counterweight pockets in the other five windows were quite leaky. After the tests, Phoenix Window technicians removed the weights and insulated and air sealed the pockets. Then they equipped the newly sealed windows with hidden channel balances, which provide the functionality of counterweights but cannot be seen. (See "Insulating and Air Sealing Counterweight Pockets in Double-Hung Windows," p. 36 for another possible approach.)

Phoenix also filled holes in and around the frames, removed old paint and glazing compound, restored the functionality of the original sliding mechanisms, installed weather strip, sealed wood surfaces, installed new glazing compound, and adjusted or replaced lock mechanisms. Post-retrofit blower door testing at the home revealed an average of about 6 therms of natural gas savings per window per heating season from diminished convective losses alone.

In addition to the restoration work described above, insulated-glass units (IGUs) were retrofitted in a 16 ft2 window on the south facade (the original glazing had been replaced years ago). Finally, at the request of the homeowners, Phoenix Window technicians built three traditional-style wood-frame storm windows to replace the old aluminum-frame storm windows. One of these new storm windows was equipped with single glazing and a sliding ventilation panel. The other two were equipped with IGUs. These IGUs were constructed with a thin plastic film between two sheets of glass, special coatings on the inside surfaces, and krypton gas; the latter was supplied by Alpen Windows. One of the storm windows with IGUs also had extruded polystyrene (R-value of 5 per inch) integrated into hollowed-out portions of the window frame.

Energy losses or gains through windows vary directly with their size, the temperature difference between inside and out, and the heat transfer coefficient (U-factor). U-factors are the inverse of R-values, where R is the resistance to heat flow. In practice, it is difficult to measure U-factors in the field, so the project team built and instrumented a testing facility to measure U-factors and study air leakage. The testing facility features a superinsulated 390 ft3 inner chamber designed to measure fenestration samples of up to 20 cubic feet. It is surrounded by a well-insulated outer chamber. The inner chamber is heated by electric-resistance radiant panels, and the outer chamber is cooled by chilled air. The result is a difference in temperature that averages about 70ºF. Heating energy required to maintain this temperature difference is measured precisely, as is temperature from a number of probes inside the hot chamber and between the hot and cold chambers. Data loggers record energy use and temperatures each minute for subsequent analysis.

In addition, we used a recently calibrated variable-speed fan (a Duct Blaster from Minneapolis Blower Door) and associated DG-700 two-channel digital manometer with the inner chamber to quantify the extent of air leakage through cracks between fixed and moveable portions of the window frames.

The solar heat gain coefficient (SHGC) is the fraction of solar radiation admitted through a window that is directly transmitted and absorbed, and subsequently released inward. In this project, the team used a factory-calibrated pyranometer to take readings of the SHGC in both the test home and the laboratory.

Table 1. R-Values and U-Factors for Ten Fenestration Systems
Table 2. Savings from Retrofitting Double-Hung Window with a Low U-Factor Storm

Table 1 shows some key results of laboratory testing on ten fenestration systems. Window 5 is the middle window on the south elevation.

Low standard deviations indicate consistency in measuring techniques and lend credence to the relative results of U-value testing. The wind-adjusted U-factors and R-values reflect canonical parameters of the difference between the insulating value of an air film under still-air conditions and an air film with a 15 mph wind on the exterior of a structure.

Each of the storm windows was tested on its own and in various combinations with existing and retrofitted double-hung windows. Note that the single-glazed storm (R = 0.81) outperformed the existing aluminum storm (R = 0.53) by a factor of 1.5; the new storm without insulation in the frame (R = 3.15) outperformed the aluminum storm by a factor of 5.9; and the storm with insulated frame (R = 3.61) outperformed the aluminum storm by a factor of 6.8. Retrofitting the old double-hung yielded an R-value of 1.57, an improvement by a factor of 2 over the existing old double-hung (R = 0.79), whereas replacing the old double-hung with a new vinyl window yielded an R-value of 2.25, an improvement by a factor of 2.9. Retrofitting the old double-hung window and then installing a new storm window with a partially insulated frame yielded the best overall performance. This yielded an R-value of 5.33, an improvement by a factor of 6.8 over the old double-hung in its original condition.

Estimating Performance in Five Cities

How a window performs in a given city depends on many variables. These include window area, SHGC, and U-factor, as well as solar radiation and temperature data for a typical meteorological year (TMY). These variables can be used to calculate hourly energy gains and losses for virtually any fenestration system for which TMY data are available.

Table 2. Savings from Retrofitting Double-Hung Window with a Low U-Factor Storm
Table 2. Savings from Retrofitting Double-Hung Window with a Low U-Factor Storm

We used RESFEN (for 'residential fenestration') software developed by Lawrence Berkeley National Laboratory to estimate the summer and winter energy performance of two fenestration systems in five American cities (see Table 2). The two systems in question are (1) the old double-hung Window 5 and (2) the old double-hung Window 5 combined with the new low U-factor storm. We chose the cities to show a range of climates. Table 2 shows energy and economic performance based on 100 square feet of each of the two fenestration systems, installed on each of four facades. The calculations assume a retrofit cost for the low-U storm window of $25 per square foot, a rough average of the cost of this type of retrofit nationally. The data also reflect the cost of gas and electricity, which can differ substantially in different areas. For example, homeowners in Anchorage pay $0.441 per therm for natural gas, while those in Phoenix pay $1.43 per therm. Homeowners in Boston pay $0.17 per kWh for electricity, while those in Atlanta pay $0.071.

We compared results for the old double-hung window and for the same window with a new low U-value storm window in each of the five cities. Absolute savings are expressed in millions of British thermal units (MBtu), where 10 therms of gas = 1 MBtu and 293 kWh of electricity = 1 MBtu. The absolute savings realized by adding a new low U-value storm window to the old window average 44.4 MBtu, and the relative savings average 87%. Dollar savings average $480. The absolute savings in Anchorage were more than 77 MBtu, yet the simple payback is 28 years, because natural gas costs so little in Anchorage. These analyses are based on the following assumptions and caveats:

  • The assumed cost of a highly energy-efficient storm window reflects no economies of scale and assumes a wood frame, which is labor intensive to manufacture and has an R-value that is one-third that of fiberglass or vinyl.
  • The analysis does not take local, state, federal, or utility incentives into account.
  • The analysis does not take into account such intangible benefits as increased comfort, yet for many consumers this is the most important reason for upgrading windows.
  • The analysis does not take into account any increase in the lifetime of the primary window caused by rehabilitating that window or by adding an energy-efficient storm window.
  • Cost-benefit calculations assume that the rate of inflation in energy costs is identical to the overall rate of inflation. For this reason, the analysis is probably conservative.

Before drawing inferences from these findings, it is useful to take note of the key role played by frames in the energy performance of windows. Manufacturers tend to use the same shape and size of framing material for both large and small windows, so small windows have relatively more framing material than large ones. Many modern IGUs are much better insulators than the frames that surround them (the wood frames found in historic windows have an R-value of only about 1). So if an IGU has a center-of-glass R-value of 10 (U = 0.1), and if it is mounted in a wood frame the area of which is 25% of the whole, the net R-value of the window system is only 3 (U = 0.33). If the same IGU is mounted in a fiberglass frame the area of which is 13% of the whole, the net R-value of the window system is 7.75 (U = 0.133). Since this system will also improve SHGC (because less sunshine is blocked by the frame), the overall improvement in energy performance for southern exposures on winter days will be improved by a factor of 3.


  • it is important to improve the R-value of frames, as well as the counterweight box of an old double-hung window;
  • it makes sense to raise the glazing-to-frame ratio of any window system if the frames have an R-value of 1 or less;
  • the smaller the frame and the higher its R-value, the more cost-effective it is to invest in high R-value IGUs; and
  • the smaller the frames and the higher their R-value, the more cost-effective it is to retrofit storm windows with high R-value glazing.

The Storm Window Option

Adding an efficient IGU to an existing window yields good energy savings, if we air seal and insulate the existing window and its weight pockets. However, if a wood frame is large relative to the glazing, the only good way to save energy while retaining the historic window is to install an energy-efficient storm window.

Adding storm windows has aesthetic, as well as thermal, advantages. When we added a storm window to our original window, performance increased by well over 400%, even though we kept the original wood frame. Switching to an energy-efficient fiberglass frame would improve the performance a great deal more, probably increasing overall efficiency by a factor of 6 or more over that of the original single-glazed wood-frame window. It is possible to fabricate fiberglass frames that look virtually identical to wood frames. Such a storm window could bemanufactured at a lower cost than a wood-frame storm window. It would last longer, save more energy, and achieve better comfort than a wood-frame storm. One advantage of storm windows is that many homeowners associations and local historic preservation groups view them as removable, and so acceptable.

The storm windows studied on this project ranged from single-glazed clear-glass units to high-performance IGUs 3/4 inch thick. The latter were fixed units that can be opened to provide ventilation (or to serve as an emergency exit) only by unhooking the bottom of the storm window and pushing it out from the wall, enabling the hooks on top to function as hinges. We also made two storm windows in the same fashion, but with IGUs on the bottom in a removable frame that could be replaced by a screen and frame for summer use.

A handful of manufacturers now produce single-glazed storm windows with low-e hard-coated glass. Although we did not include these windows in this research, we estimate that combining them with a conventional single-glazed, double-hung historic wood window would probably result in a fenestration system with an R-value of slightly better than 2 (and a U-factor of 0.45 or so).

At least one manufacturer is about to bring into the market high-quality storm windows akin to those tested in this research. Phoenix Window Technologies now offers a range of storm windows and related energy-saving products for the retrofit residential and commercial markets. In view of our research results, a mass-produced, semicustom fiberglass-frame solution may find strong commercial demand among owners of historic homes. My analysis suggests that storm windows 10 square feet or larger, with fiberglass frames, could be sold in bulk quantities (20 or more) for a price ranging from $14 to $18 per square foot. In the cities that we analyzed, the payback period on installing such a system would be about the same as the payback period on installing a lower-end vinyl replacement window — with more efficient, comfortable, and attractive results.

Apropos of comfort, a thought experiment may be instructive. In general, the temperature between the inside and outside of several fenestration units like primary windows and storms divides in proportion to their respective R-values. Imagine a single-glazed wood-frame window with an R-value (and a U-factor) of 1. Imagine that an aluminum-frame storm window with an R-value (and a U-factor) of 1 is installed a couple of inches away. On a day when the outside air temperature is 10ºF and the inside air temperature is 70ºF, the temperature between the two windows will be around 40ºF, roughly halfway between the inside temperature and the outside temperature. Further, the inside surface of the wood-frame window will be quite cold, because of thermal bridging. This will cause drafts. It will also cause substantial radiant heat loss from the bodies of the occupants.

learn more

For a copy of Historic Window Research Final Report, on which this article was based, go to www. SynertechSystemsCorp.com under Historic Preservation.

The Window Attachments web site, which is maintained by BuildingGreen, Incorporated, and Lawrence Berkeley National Laboratory's (LBNL) Windows and Daylighting group, has a number of well-written fact sheets on the single-pane low-e storm windows mentioned in this article. To download, go to www.windowattachments.org.

For more on the Phoenix Window Technologies windows mentioned above, go to www.phoenixwindow.net.

Now imagine that we replace the aluminum storm window with a high-quality storm window with an R-value of 5 (and a U-factor of 0.2). Given the same inside and outside temperatures as we saw in the first case, the temperature between the two windows will be 60ºF. And the inside surface of the wood-frame window will be quite close to the ambient temperature in the home, which will very substantially limit radiant heat loss and drafts.

Finally, imagine that the wood-frame window is somewhat leaky. Unless indoor relative humidity is pretty high (which is rare in old homes), there is less likelihood of reaching dew point between the glazings than would be the case if the wood-frame window did not leak. This is because the space between the glass won't be cooled very much. The potential for fogginess between the windows will also be reduced. Working to ensure that the wood-frame window is at least as tight as is the storm will help as well.

History and Energy Efficient

It is possible to improve the overall energy performance of historic window systems fourfold or more by repairing and sealing the existing window and installing an excellent storm window. All this can be done without altering the historic character of the original window. This strategy also protects the original windows, gives them new life, and makes them more functional. Insulating and air sealing the envelope and the duct system as well could improve efficiency by 60 - 80% over historic buildings that are leaky, have little insulation, and have wood-frame sin-gle-glazed windows. It would also increase the functional lifetime of the home and lower energy bills for the duration.

Might it be possible to please both historic preservation buffs and conservationists?

Our research suggests yes.

Larry Kinney has been involved in energy efficiency research and development for 35 years. He is president of Synertech Systems Corporation in Boulder, Colorado, and was the technical team leader on this historic-windows research project.

Discuss this article in the Best Practices (Residential) group on Home Energy Pros!

Add a new article comment!

Enter your comments in the box below:

(Please note that all comments are subject to review prior to posting.)


While we will do our best to monitor all comments and blog posts for accuracy and relevancy, Home Energy is not responsible for content posted by our readers or third parties. Home Energy reserves the right to edit or remove comments or blog posts that do not meet our community guidelines.

Related Articles
SPONSORED CONTENT What is Home Performance? Learn about the largest association dedicated to home performance and weatherization contractors. Learn more! Watch Video