Creating Windows of Energy-Saving Opportunity

Andrew M. Shapiro is an energy engineer with the Vermont Energy Investment Corporation in Burlington, Vermont. Brad James is a master's degree candidate at the University of Vermont. S. Flanders of the U.S. Army Cold Regions Research and Engineering Laboratory assisted with project oversight.

Brad James sets up a test rig to measure air leakage around a window.

From their handmade glazings to their crafted muntins, old windows add much to the character and charm of historic homes. But just looking at their loose jambs and leaky sashes can make an energy auditor shudder. Although the tendency among some contractors has been to replace the windows in older homes, until recently, there has been very little data available to guide renovators in choosing the most energy-efficient window rehab option (See Energy-Efficient Window Retrofits, HE Jan/Feb '97, p. 35).

To help fill this data gap and supply additional guidance to renovators, we evaluated the thermal efficiency of more than 150 windows in 29 old New England homes and one municipal building. We determined the energy savings and costs associated with different renovation strategies, from simply weatherstripping to replacing the entire sash. The study was funded by the National Center for Preservation Technology and Training and the Vermont Division of Historic Preservation.

The windows varied widely in age and condition; a few were at least 100 years old. Many of the original windows had storm windows installed. Some renovated windows still had the original sash, while others had been retrofitted with a new sash. For the retained sashes, the contractors used a variety of weatherization and renovation methods; these are described below. The retrofitted windows had received either a new sash in the old jamb or a new vinyl or wood window insert (also known as a secondary frame) with a new sash. On some homes, the contractors had installed new storm windows.

Although the windows we tested varied in size and shape, we were able to make comparisons across sizes by normalizing the data to a typical window 36 inches wide by 60 inches high.

Figure 1. On the left, typical air leakage sites with Sash Leakage, S, and Extraneous Leakage, E. On the right, a cutaway showing the parts of a typical single-hung window.

We calculated thermal losses using WINDOW 4.1, a computer model developed by Lawrence Berkeley National Laboratory's Building Technologies Program. We based our infiltration test method on ASTM E783-93, performing two air leakage tests on each window. We constructed a simple measurement device around the windows by taping a plastic sheet onto the interior trim and attaching an air hose, blower, and pressure tap. First, to test total leakage, we drew air through the window using the blower and measured the flow rate, in ft3 per minute (CFM), at various pressure differentials across the plastic sheet. Then, to test extraneous leakage, we attached a second plastic sheet to the exterior trim of the window and repeated the test. By subtracting the value obtained in the second test from that obtained in the first, we were able to estimate sash leakage.

Some building designers think of window infiltration only in terms of sash leakage. But significant leakage can also occur between the window frame and the rough opening. (Note that window manufacturers report only sash leakage in product data.) To estimate how much rough opening leakage contributes to total window infiltration, we measured the temperature of the indoor air, the outdoor air, and the air being drawn through the window during 33 of the extraneous leakage tests.

We found that, on average, the air drawn through the windows in the study was approximately 30% cooler than the indoor air. Based on this difference, we assumed that approximately 30% of the extraneous leakage was outdoor air coming through the rough opening. We thus estimated total infiltration as sash leakage plus 30% of extraneous leakage. While this method was not very precise, it did allow us to estimate the relative contribution of rough opening leakage to heating load.

Air leakage rates of the original windows varied widely, reflecting the wide variation in the condition of the windows. We derived three baseline ELA values--typical, tight, and loose--for use in the comparison with the renovated and retrofitted windows. The loose baseline value was the mean leakage value of all the original windows without storm windows. The typical baseline value was the mean leakage value of all original windows that had storm windows in place. The tight baseline value was set at one standard deviation lower than the typical baseline value.

Table 1 is a summary of the ELAs for the original and renovated windows. The vinyl window insert was clearly the tightest option, having one-fifth the leakage of the baseline average.

Tightening Up with the Original Sash

For those houses where the sash was retained rather than replaced, we found that contractors used a variety of methods and materials to tighten the sashes. These methods and materials are summarized in Table 2.

Figure 2. First year heating cost per window, pre- and post-treatment.

Lead Abatement

The problem of lead paint often arises when dealing with old windows. As shown in Table 3, lead abatement can add significantly to renovation costs (see Getting the Lead Out). For example, weatherizing a loose baseline window with weatherstripping, sealing the top sash, and rehabbing an existing storm costs $75 if lead abatement is not needed. At the annual savings rate of $15, this represents an annual rate of return of 20%. But when lead abatement is needed, the cost jumps to $200, which approaches the cost of replacing the sash.

Table 3. Renovated Windows Annual Heating Savings and Renovation Costs per Window Renovation Cost Cost with Lead Abatement Annual Savings (Tight) Annual Savings (Typical) Annual Savings (Loose) Retain sash: Vinyl jamb liner $175  $300 none $0.80 $14 Weatherstripping 75 200 $0.20 1.70 15 Replace sash: Single-glass sash 200 200 0.30 1.80 15 Window inserts 250-500 200-500 1.90 3.40 16 Low-e, double-glaze inserts 250-550 250-550 5.30 6.80 20 Storm Windows New exterior 100 225 1.00 2.50 16 New interior 115 240 1.30 2.80 16 Interior low-e 155 280 4.70 6.20 19 Note: Costs for inserts varied with the material, which ranged from medium-cost vinyl to high-quality wood. Full-sash lead abatement adds $125 to other rehab costs. Savings were based on 7,744 degree-days and oil heat at 90¢/gallon with 75% overall heating season efficiency. Note that the samples of most windows tested were very small. Cost estimates for window upgrades were based on interviews with housing developers and/or builders. Estimates were normalized to a $20-per-hour labor rate, and included contractor markup.

Figure 3. First year heating costs, infiltration only, renovations retaining sash.

Table 3 also shows the costs and savings associated with adding storm windows. A new exterior storm window added to a loose baseline window has a first-year savings of $16 at a cost of $100 (excluding lead abatement), or a 16% annual rate of return. Adding a low-e interior storm to a loose window saves $19 at a cost of $155, a 12% annual rate of return.

Figure 3 shows first-year heating costs due to air leakage only. The variation in the cost of air leakage in the first four columns suggests variability in workmanship in installing the jamb liners. Also, windows where the jamb was out of square had the poorest seal.

Sites C and D both incorporated weatherstripping at the meeting rail but the much lower leakage rates should not be attributed to that difference alone. Jamb liners require that the sash be fitted precisely to the liner and to the jamb to prevent leakage between the jamb and the jamb liner and between the liner and the sash.

The strategy used at site E resulted in quite low sash leakage. However, the total leakage was approximately the same as that at sites C and D, due to high leakage through the rough opening. Similarly, even though site G had a very low sash leakage, performance was undermined by the rough-opening leakage.

Is a New Sash Worth It?

Figure 4 compares the associated heating costs for infiltration and thermal losses for windows that retained the original sash. The chart shows that, compared to the thermal losses, infiltration accounted for a small part of the total heating cost, regardless of the strategy used.

Table 4 lists the heating costs for windows with new sashes or window inserts. With the exception of the low-e sash and the poorly fitted sash, the total heating costs were very similar, ranging from $12 to $14.

Figure 4. Total first year heating cost, renovations retaining sash.

Table 4. Annual Heating Cost for Upgrades that Include Replacing the Original Sash Upgrade Description Non-Infiltration Heating Cost Infiltration Heating Cost Total Heating Cost Vinyl window insert $12 $0.37 $12 Wood window insert 12 0.70 13 Sash & storm 12 1.68 14 Sash & storm (poor fit) 12 4.83 17 Insulated glass sash 12 0.60 13 Insulated low-e sash 8 0.60 9 Note: Savings were based on 7,744 degree-days and oil heat at 90¢/gal with 75% overall heating season efficiency. We did not observe the use of low-e glass in the field and calculated the thermal loss for the low-e retrofit. We estimated the added savings of low-e glass over other sash replacement strategies at $3.40 per year.

Brad James, who worked on much of the testing and analytic work for this study, covers a window with plastic in preparation for the extraneous leakage test.

The difference in annual energy savings between renovating an old sash and replacing it with a new one was very small--retrofits saved only a few dollars.

For preservations, the good news is that with a proper choice of renovation strategy and good workmanship, historic sashes can be almost as energy-efficient as replacements. Window renovators and homeowners can give more weight to comfort, maintenance, lead abatement, egress requirements, durability, ease of operation--and historical value--without sacrificing energy savings. For those of us who work with old windows, this is very good news indeed.

Grimsrud, D.T., M.H. Sherman, R.C. Sonderegger: Calculating Infiltration: Implications for a Construction Quality Standard, Proceedings, ASHRAE/DOE Conference on Thermal Performance of Exterior Envelopes of Buildings II. 1982.

James, B., et al. Testing the Energy Performance of Wood Windows in Cold Climates, National Center for Preservation Technology and Training, 1996.

Lawrence Berkeley Laboratory, Windows and Daylighting Group, Berkeley, California. WINDOW 4.1: A PC Program for Analyzing the Thermal Performance of Fenestration Products (1994).
 
 

Home Energy can be reached at: contact@homeenergy.org
Home Energy magazine -- Please read our Copyright Notice