The Manufactured Housing Institute arranged a collaboration between NREL and Schult Inc. Schult built two 16260 ft homes, one to a proposed "cold-zone" standard and one to a proposed "warm-zone" standard. Schult was allowed to choose any of their typical single-wide models as the base to which energy improvements would be applied. NREL didn't try to influence the design of the homes.

We also observed the homes being built in the factory, documenting "as-built" construction details and specifications. Schult calculated compliance with the new standards for their designs using HUD's new draft calculation manual. Later, we recalculated compliance with the new standards using the same method, but using the "as-built" specifications instead of "design" specifications. We then compared the calculated overall heat transmission coefficient to that measured in the environmental enclosure.

The proposed standard does not cover a number of factors that affect the energy efficiency of mobile homes (including infiltration and ventilation duct leakage). Many heat transfer paths were not taken into account by the proposed standard and calculation method. We tested for potential thermal anomalies. We investigated the degradation in thermal performance that might occur because of penetrations in the rodent barrier from field hookups and repairs, duct leaks, closing of interior doors with and without operation of the furnace blower, and exposure to winds.

Test Methods

We used a variety of techniques to test the mobile homes including coheating tests, infiltration tests, furnace efficiency tests, and infrared imaging. CMFERT is equipped with a computerized data acquisition and control system which maintains the warehouse and mobile home at a steady temperature and measures the electric resistance heater power required by the home. This method, known as a "coheating" test, allows a manufactured home to achieve a steady-state condition, with the warehouse temperature acting as the "outside" temperature. An overall heat transmission coefficient for the home can then be determined from the measurements. This coefficient includes heat loss from conduction through the envelope of the home as well as infiltration heat loss (heat loss from air leakage). To separate these two types of heat loss, we measure infiltration rates of the mobile homes with tracer gas and a blower door.

We also measure the combined furnace and heat distribution efficiencies of the homes. This is done by comparing the energy required to heat the home with electric resistance heaters in the living space (therefore 100% efficient) to that required when the home runs on its own (in this case gas forced air) furnace and duct system. The ratio of these quantities determines the overall heating system efficiency. These tests are done under both a still air and a simulated 3-mph wind condition. The wind is induced with a large bank of fans (see "CMFERT: Training and Testing of Mobile Home Retrofits," HE Jan/Feb '90, p.23, and "Mobile Home Retrofits Revisited: CMFERT Phase II," HE Jan/Feb '91, p.21).

`Apples to Apples'

The proposed standard and the draft HUD calculation manual do not include infiltration losses in the overall heat transmission coefficient calculations. For this reason we separated the measured heat loss into a conduction portion and an infiltration portion so that an "apples to apples" comparison could be made between the measured conduction heat loss coefficient, the heat loss coefficient specified in the standard, and the calculated conduction heat loss. We subtracted the heat loss caused by infiltration from the measured overall heat loss of the home to obtain the conduction heat loss coefficient. The maximum allowable conduction heat loss coefficient Uo is specified in the proposed standard as:

(cold zone) Uo = 0.079 Btu/hr ft²deg.F

(warm zone) Uo = 0.109 Btu/hr ft²deg.F

Both homes met the proposed standard with minimal design changes. The cold-zone home came in about 1% below the maximum allowable value of Uo and the warm-zone home passed by about 7%. In the warm-zone home, the roof, floor, and wall sections were all typical for this manufacturer, as were the windows and doors. In the cold-zone home, all components were typical except for the wall, which was increased from 224 to 226 in. stud-wall construction to accommodate a thicker insulation batt. Even the 6 in. thick wall is an option already offered by the manufacturer (for an additional cost). The only changes from the usual design were the amount of insulation specified in the floor and roof. Increased insulation fits into the typical cavity depth available in designs normally used by this manufacturer (see Table 1).

There's a Hitch

These results are encouraging, but the homes tested were built with the manufacturer's 82 ft² "minimum window" option. Had the window area been larger, substantive design changes would have been necessary to meet the standard. For example, if all other components of the homes were kept the same, but the window area was increased to 115 ft², Uo for the cold- and warm-zone homes would increase to about .084 and .122 Btu/hr ft²deg.F respectively, as estimated using ASHRAE heat-loss calculation methods. Therefore, the homes wouldn't meet the standard.

We analyzed several design alternatives to bring units having a window area of 115 ft² into compliance. Homes with increased window area could meet it many ways. For instance, a warm-zone home with 115 ft² of window area could meet the standard with added self-storing storm windows or an increase from R-11 in the floor pan and wings to R-22 and R-19 respectively, and an increase from R-11 to R-19 in the walls. In a cold-zone home with 115 ft² of window area, increasing the average roof insulation from R-16 to R-26 by substituting flat-roof-truss construction for vaulted roof scissor-truss construction could bring the home within compliance. Vaulted ceilings could be retained in the cold-zone home by increasing their average insulation from R-16 to R-26 by redesigning the scissor truss to be deeper in the heel and deeper overall (this would also require redesigning the flat-ceiling truss to match the pitch of the new scissor truss in homes with combined vaulted and flat ceilings).

Schult typically uses wood siding and triangular truss roofs with blown-in insulation for both their single and double-wide models. Roughly half of the manufactured housing industry now consists of "single-wide" units, and many of these still use (less energy-efficient) metal siding and bow-string truss galvanized roofs. This will probably make it harder for them to meet the new standard.

Testing the Tests

The new HUD compliance calculation method proved quite accurate for these homes coming within 2% and 8% of the actual measured values for the cold and warm zone units, respectively. However, these homes contained relatively few of the problematic heat-loss paths found in older homes. The calculation method should not be expected to work as well for homes with ventilated walls, mostly empty interstitial cavities, and cavities with large leaks to the outside. Neither the standard nor the compliance calculation method account for natural infiltration, convective bypasses, duct leakage, heat losses from operation of forced-air blowers, or ventilation requirements. The better constructed and insulated the home, the more accurate the calculation method will be. This trend is seen in the data where the calculation method was more accurate for the cold-zone home than for the warm-zone home.

Real-Life Conditions

We tested the homes for sensitivity to various real-life situations. Rips in the rodent barrier were made to simulate the types of holes likely to result from routine installation and maintenance of a home (for instance, field hook-ups of gas, water, electrical, and waste lines). Cuts were made and insulation was shifted to the side at strategic places (for instance under the toilets) as if repairs were made to those parts of the home. To characterize the effects of air pressures induced by the forced-air circulation system, for some tests, the furnace fan was hard-wired on, without the furnace actually producing any heat.

Both homes were more resistant to degradation in performance from wind, rips in the rodent barrier, and operation of the furnace blower, than older homes we have tested in the past. Ducts and floors were relatively tight (compared with older homes) and under-floor cavities contained more insulation, thereby suppressing convective bypasses. However, with interior doors shut and the furnace blower operating, the heat losses for both homes increased by 39 and 36 Btu/hr deg.F (approximately 15%) respectively. In this case, the overall heat-loss coefficient exceeded the standard for both homes. This is extremely important because occupants are likely to close bedroom doors, especially at night when the furnace operates the most.

In a forced-air heating system of this type, air is delivered to the individual rooms via air ducts under the floor (in the belly). The air then returns to the furnace through an air intake at the furnace itself, positioned in the central kitchen/living room/dining room area. When all interior doors are open, the return air can pass freely through the doorways to the furnace. If the interior doors are closed, the air must pass through relatively small cracks under the doorways to return to the furnace. This results in higher pressures in the delivery ducts and end rooms, which are partially sealed off from the central room by the doors, and relatively lower pressures in the central room even if the doors are properly undercut as they were in these homes. These pressure differences increase the overall infiltration rate of the home and the ductwork, thus increasing heat loss (see "Mobile Homes: Small Zones, Big Problems," HE Sept/Oct '93, p. 34). Even small holes in the duct work can result in large heat losses under these conditions.

Furnace Efficiency

Generally, the heating system was more efficient in the cold-zone home. This is probably because the cold-zone floor cavity contained R-22 insulation under the heating duct, whereas the warm-zone home contained R-11. We also tested the furnace efficiency under the same set of above mentioned real-life conditions. Again, the largest effect was from the combination of closed interior doors and operation of the furnace blower. In the cold- and warm-zone homes this decreased heating system efficiency from the base case by 5% and 9% respectively. The greater decrease in furnace and duct efficiency for the warm-zone home may also have been caused by the smaller amount of floor insulation in that home. In both homes, efficiency decreased when holes were cut in the rodent barrier and when the interior doors were closed. An approximate 3-mph wind, however, did not appear to affect the delivered heat efficiency of either home, indicating that the combination of an intact rodent barrier and either R-22 or R-11 under the duct, render the homes relatively impervious to duct heat loss from wind. It is possible that decreases in efficiency could be detected under greater wind speeds, or if holes in the ducts and/or rodent barrier were larger.

Air Tightness

To see if the homes became less tight from the stresses encountered on the highway trip from the factory, we blower-door tested the cold-zone home before it left the factory and again when it arrived at CMFERT (about a 300-mile trip). The home did not become leakier as a result of the trip.

In addition to the two homes we tested at NREL, we blower-door tested three other homes during our visit to the factory (see "Factory Observations," p.25). All of the mobile homes we tested were extremely tight--perhaps too tight. ASHRAE recommends an average infiltration rate of at least 15 cubic feet per minute (cfm) per person. At the factory the homes averaged 10 cfm per person. It may be that a mechanical ventilation system should be required, or that the new standard should include maximum and minimum fresh air requirements.

As expected, both homes had increased leakage and infiltration rates after holes were cut in the rodent barrier. In the warm-zone home, we sealed off the furnace ducts to observe any change in equivalent leakage area, an indirect way to measure duct leakage know as "the subtractive method." We measured only a 3 in² reduction. Visual observations revealed approximately a 6-in² gap between the furnace plenum and supply duct. These holes are small compared to ones we have observed in homes in the low-income Weatherization Assistance Program. However, even small supply duct leaks can cause large heat losses and reductions in heating system efficiency when the furnace is on and interior doors are closed. Also, these ducts were new. The tapes used to seal the fiberglass folding-board duct sections are known to degrade over time, and these ducts are extremely difficult to repair. It is probable that duct leakage will increase over time with this type of duct.

The Shape of Mobile Homes to Come

Generally, the proposed HUD standard represents significant progress with respect to the 1976 standard. However, the emphasis on the overall heat transmission coefficient (Uo) should be complemented by equal concern for other issues.

• Maximum and minimum air leakage and ventilation criteria should be part of the standard. Blower doors could make determining leakage simple for manufacturers and compliance inspectors.

• A requirement should be included in the new standard for duct and furnace plenum integrity, a potentially large source of heat loss.

• A requirement for balancing air distribution, return air systems, and forced ventilation systems should be included in the new standard.

• A loophole in the proposed standard as worded was the definition of Uo. This value is normalized by the total surface area of the home. This allows a designer, when necessary, to meet the letter (but not the spirit) of the current and proposed standard by increasing wall height. This increases the ratio of low heat-loss surface area to that of high heat-loss components such as windows. Uo is decreased, but the total heat-loss of the building actually increases. Everybody loses: the manufacturer incurs greater materials costs, the building uses more energy, and consumers pay higher fuel bills. This method of increasing the wall area to decrease the Uo was used on the warm-zone home when it failed to meet the proposed standard with its usual 84 in. wall. By increasing the wall height to 90 in., the designer met the standard. Apparently, this is a common practice. This loophole should be rectified.

Scan Results

We used infrared cameras to look for potential thermal anomalies. Although we found a number of them, only one of these, the connection of the furnace to the heating duct via the floor plenum, was of major thermal significance in these two homes. We know this from the tests at CMFERT. While infrared imaging is extremely useful for locating potential problems, ordinarily it is almost impossible to quantify the energy significance of a problem just from the scan. For example, a scan might indicate a duct leak problem behind a wall, but it does not indicate how much energy would be saved by its repair.

The ability to conduct the infrared scans under controlled, repeatable steady-state conditions, and in conjunction with other testing techniques, proved extremely informative, and could eventually lead to improved interpretation of infrared images taken under field conditions. In this process we learned as much about the infrared technique as we did about the homes. This is illustrated by differences observed between sets of scans taken under both steady-state and non-steady-state conditions.

Under non-steady conditions, many structural elements appeared to have severe thermal shorts when observed from the inside because of the temperature history of the units. This was because of their thermal capacitance. When the home was rapidly heated, these structures remained cold longer than the insulation cavities and interior finishes, thus giving the appearance of thermal bridging. Once the homes reached steady state, many of these temperature differences became much less pronounced. Auditors using infrared cameras in the field need to be aware of the inside and outside temperature history of the home. For example an infrared scan could be very misleading shortly after occupants switched on the thermostat.

We also observed that the rodent barrier material was slightly reflective in the infrared spectrum. Therefore in some cases heat reflections from the body of the cameraman appeared to indicate large heat leaks in the belly cavity.

Running the Gauntlet of Regulators

Mobile homes represent roughly 15% of the nation's new homes. Occupants tend to be poorer--less likely to be able to afford high energy bills. One estimate is that new standards will cost consumers $10 per month in increased mortgage payments, yet save them $18 per month in lower energy bills.

In 1987, Congress passed legislation requiring HUD to revise its energy standards for mobile homes, which have not been changed since they were instituted in 1976. At the same time, standards for site-built buildings have vastly improved and a large number of factory-built homes already exceed current standards. But the revisions were delayed (see "HUD Standards Overdue," HE Sept/Oct '92, p.9).

Exasperated with the delay, Congress (in an amendment to the Energy Policy Act of 1992) ordered HUD to issue new standards within one year or relinquish to the states its authority to set the standards (see "Unfinished Business at HUD," HE Nov/Dec '92, p.2). HUD contracted with Pacific Northwest Laboratories to develop revisions to the standard and PNL's recommendations were included (with few changes) in HUD's proposed rule. With time running out, HUD was expected to issue the final rule in late October.

-- Cyril Penn

Component             Cold-Zone Home                        Warm-Zone Home        
___________________________________________________________________________________________
Floor Wings           R-20 (2 3.5" batts compressed         R-11 (1 3.5" batt)
                            in a 226 joist cavity)        
Floor Pan             R-26 (1 3.5" blanket doubled over)    R-11 (blanket)        
                            over)
Floor Average         R-23                                  R-15        
Ceiling: Vaulted      R-17                                  R-17        
Ceiling: Flat         R-26                                  R-26        
Ceiling Average       R-18                                  R-18        
Walls                 R-14 (2" 2 6" framing,                R-10 (2" 2 4" framing, 16" o.c.)
                            16" o.c.)
Windows               R-1.4 (1 pane + interior              R-0.8 (1 pane)
                             self storing storm)
Heating Duct          R-4 board inside R-22 blanket         R-4 board inside R-11 blanket
Furnace               Gas-forced air                        Gas-forced air
Return Air            Through living space                  Through living space

        
Internal Dimensions                        
___________________________________________________________________________________________
Length 2 Width        55.35 ft 2 14.7 ft                    55.7 ft 2 15 ft
Wall Height           7 ft                                  7.5 ft        
Vaulted Ceiling Area  636 ft                                653 ft2
Flat Ceiling Area     178 ft2                               183 ft2
Window Area           82 ft2                                82 ft2
Floor Area            814 ft2                               836 ft2        
Volume                6173 ft3                              6580 ft3        
___________________________________________________________________________________________
Note: R-values are average R-values for the entire assembly including framing,
insulation, compressed insulations, air films,and so on.

A one-week factory visit allowed us to become familiar with factory operations, production-line procedures, and quality control methods. Here are some observations and suggestions from the visit:

• Floor joist spacing was not always consistent. Most joists were 16 in. on center (oc), but some were 18 in. The insulation batts were precut for the 16 in. spacing, which caused insulation gaps in the floor wings, requiring production to be slowed down while insulation scraps were hand cut to fill the gaps. Modular design changes to a 16 in. structural floor module would eliminate odd joist spacing, simplifying the installation of insulation batts and allow standardization of the design of the chassis wing supports.

• The plastic strapping that supports the insulation batts between the floor joists was not stapled at each joist., causing batts to sag and leaving voids behind the band joist, which created a thermal short at the floor edge. This problem was especially evident in the warm-zone home with R-11 (3.5 in.) batts. Stapling the strapping at every joist would help. A better solution would be to always fill the joist cavity with a full-depth batt (R-19 for these homes) and staple at every joist regardless of the zone for which the home was designed.

• A flange detail should be developed to seal the connection between the furnace plenum and the heating duct where gaps were observed in both homes.

• A template or other quality control procedure should be instituted to ensure that the rough opening for the furnace plenum is properly aligned with the top of the heating duct. Otherwise, a duct leak is unavoidable.

• The fiberglass folding-board ducts look flimsy and are difficult to repair, and the durability of the tape joints is questionable. Additionally it is difficult to design good durable connections between the floor register sleeves and the heating duct. Alternative heating duct materials and connection systems should be investigated.

• The scissor trusses are quite narrow, especially at the heels, limiting the depth of ceiling insulation in the ventilated-roof design used with these homes. A deeper truss design would solve these problems. An alternative would be to eliminate the ventilated roof design so that the attic could be completely filled with insulation. However, this would need study to assess the potential for long-term moisture condensation problems in the attic in humid climates.

• A flange detail should be developed to seal all rodent barrier penetrations.

• The use of "truss-studs," common in Swedish manufactured housing, should be considered. Truss studs use less material, simplify wiring and plumbing, and allow for higher average insulation levels in walls.

• The use of structural insulated panels should be investigated.

Related Articles