II. BAIHP Technical Assistance (F)

• Fleetwood Homes
  Category D, 500 Homes Auburndale, Florida factory

FEMA Homes

In September of 2004 BAIHP researchers tested and inspect single-wide homes built by Fleetwood under contract with the Federal Emergency Management Agency (FEMA) to identify possible areas of moisture-related damage and provide recommendations to mitigate problems.

These homes are destined for victims of hurricane Charley in Southwest Florida. Various singlewide floor plans are being constructed with the typical size being 14x66, several of which were tested for duct and envelope tightness. Other construction specifics include:

• In-line, metal floor duct system with 1 or 2 short branch ducts
• Duct risers sealed with mastic
• Branch duct joints sealed with mastic, then covered with metal tape
• Down flow gas furnace installed in central hallway
• Large door undercuts plus small door-mounted return vent in bedrooms
• Central exhaust fan ventilation strategy
• Vinyl interior wallboard throughout
• Vinyl exterior siding

FEMA-required specifications that differ from typical Fleetwood design include:

• Vinyl flooring throughout
• Double floor decking (½-inch OSB over ½-inch plywood)
• R22 floor insulation
• “Chicken wire” installed below the belly board
• 80% AFUE, 70 kBtu gas furnace with no cooling installed
• FEMA provides a 2.5-ton split system (coil & condenser) to be installed on-site
  • Goodman CKL30-1L condenser & Mortex 96-842J-OP A-coi

Cooling System and Air Handler Issues

The immediate concern with these homes is the FEMA-provided cooling system that, at 2.5 tons, may be oversized for the application. This, coupled with the fact that a vapor barrier is located on the wrong side of the exterior wall and floor assemblies, increases the potential for moisture damage to those surfaces. Other issues that can impact the moisture durability of these homes are addressed below, but initial envelope and duct test results indicate no immediate cause for concern.

A properly sized cooling system should be an integral part of any strategy to mitigate moisture damage in a hot humid climate. We recommend using the latest version of Manual J calculations to determine proper cooling system size and it appears these homes may be oversized by as much as one ton. Oversized systems are prone to short-cycling for much of the year which tends to cause higher indoor humidity levels than properly-sized systems.

Another issue with an oversized system is it allows homeowners to maintain lower indoor temperatures than might otherwise be possible. Maintaining indoor temperatures below the outdoor dewpoint can lead to moisture damage over time especially in homes with interior vapor barriers (vinyl floor and wallboard). Average summer ambient dew point temperatures in Southwest Florida are in the low to mid-seventies.

Beyond reducing the cooling system size, some benefit can be gained from adjusting the air handler fan speed in cooling mode and adding outdoor air ventilation. Lower airflow over the coil will remove more moisture, help to reduce indoor RH levels and possibly encourage higher thermostat settings by the occupant. Adding a passive supply (not more than 40CFM) of outside air to the return side of the air handler will promote positive pressurization of the home which may lessen the likelihood of moisture damage to wall and floor assemblies.

In-Plant Construction

Metal duct fabrication was observed during production where mechanical fastening and sealing methods appeared suitable for a tight durable system. Duct ends and branch duct joints were first fastened with screws then mastic was applied by tube. Metal tape was placed over the mastic (shown below at top right). This method produced tight duct systems as demonstrated by the 3 to 4% leakage rate found in four completed homes.

The continued use of mastic is encouraged for a long-lasting, positive seal. While there is little harm in using metal tape over mastic it does not provide much additional sealing. One possible drawback of tape over mastic is that it may hide gaps that could otherwise be seen and corrected by workers. Applying mastic alone by brush should prove adequate and less costly. A fabglass mesh is useful when applying mastic by brush to cover any large gaps that may occur.

Figure 16 Metal duct fabrication on FEMA homes,
Fleetwood plant – Douglas, GA

A bead of mastic was applied to supply risers (Figure 16 bottom right photo) prior to being attached to the trunk line with screws. Once the riser was attached an opening in the trunk line was cut out. The same method was used for the return plenum riser. This method can provide a positive seal when adequate mastic is applied – not always certain from observations on the production floor. Although testing showed four such systems to be fairly tight, some leakage at the risers was evident at the interface of the thin metal of the trunk and riser collar where unfilled gaps where found.

To prevent leakage at risers, mastic should be visibly squeezed out at the interface when attached. The mastic bead should be 1/2 to 5/8 inch in diameter (size of your little-finger) to allow full contact between surfaces.

Post-Production Testing

Four newly completed singlewides (all 14x66) were tested at the Douglas plant. Total duct leakage was measured on all homes but only two homes were measured for envelope tightness and duct leakage to out.

Blower door testing showed the envelope on the tighter side (0.73 CFM50/ ft 2) of the airtightness range typically found in new homes (0.75 to 1.0 CFM50/ ft 2). Of greater importance is where this leakage occurs. With sheet vinyl flooring installed throughout these homes, air leakage through the floor is the biggest concern. A history of floor moisture damage has been documented in manufacture homes located in hot/humid climates where vinyl products are installed. Increased air leakage between the floor and belly has greater potential to force outside air into the belly should a negative pressure situation arise in the home (caused by duct leakage and/or inadequate return air transfer). Both the interior floor surface and the exterior belly board should be sealed as tightly as practicable. Plumbing penetrations make up most of the holes through upper floor surface and can be difficult to seal. One simple option currently being used by the Fleetwood plant in Washington state involves the use of a EPDM rubber sheet cut to fit plumbing pipes and stapled in place prior to vinyl flooring installation, providing a durable, flexible seal (Figure 17).

Figure 17 Rubber seal – Washington Fleetwood plant.

One 14x66 home was tested for interior pressure imbalances by turning on the air handler fan. Depressurization of the interior space can occur if duct leakage is excessive or insufficient return air pathways exist between rooms with closed door. No detectable depressurization was measured during the test indicating sufficiently tight ducts and adequate return air pathways from closed rooms.

Duct system airtightness testing showed four systems in 14x66 singlewide homes to have duct leakage rates to out of between 2 and 4% of conditioned floor area at 25 Pascals. A value of 3% is generally considered sufficient to inhibit negative pressurization of the conditioned space. Leakage to out was directly measured in the first two test homes at 2 and 3%, while the last two homes were judged to be slightly higher as inferred by the measured total leakage rate. While these leakage numbers are good, only a small amount of leakage is necessary to dramatically increase the leakage percentage in homes of such relatively small size.

There are three general areas in these duct systems where leakage is likely to occur:

• End of duct runs
• Trunk to branch connection
• Supply risers and the air handler supply plenum

The first two of these areas were isolated and tested by duct blaster in the plant on a newly fabricated system prior to installation in the home. This particular duct system had only one branch connection whereas the four previously tested homes had two branches. Results showed a leakage rate of about 8-10 CFM at 25 Pascals, attributed to two closed duct ends and one branch to trunk connection. This would indicate that on the four duct systems tested earlier (with two branches each), roughly one-half to two-thirds of the leakage to out (20 to 30 CFM50) occurs at duct ends and branch connections with the remainder occurring at the risers and plenum.

Fleetwood Factory Visits in 2002-05

In 2002, researchers visited four Fleetwood factories in southern Georgia to investigate the cause of moisture-related building failures when units were installed in a hot-humid climate. The factories are located in Douglas, Alma, Pearson, and Willacootche. As a result of FSEC recommendations, the factories have changed their duct construction practices and are now constructing airtight ducts with mastic.

Six Fleetwood homes, all in Florida, were tested for moisture and mold damage from April 2002 through March 2003. All of the homes had damaged flooring due in part to a lack of ground cover and poor crawlspace ventilation. Damage to the floor in one home was exacerbated by a plumbing leak. Only one home had moisture damage to the wallboard material, and this home showed a history of thermostat settings below 72° F. A report for each home was submitted to Fleetwood for corrective measures. One additional high bill complaint in Cobb, Georgia was investigated during that period. Between April 2003 and October 2004 ten Fleetwood moisture damaged homes were investigated by BAIHP, seven in Florida, one in Texas, and two in Georgia.

In May 2003, FSEC researchers were asked by Fleetwood and Coleman to travel to Fleetwood's five southeastern plants and test three homes built by each factory to get their plants certified for building EnergyStar Homes. A sample of the data collected is shown in Table 13.

At the Auburndale, FL plant, BAIHP researchers conducted the tests in houses set up in the factory's parking lot. The houses did not have air handlers, but total duct leakage was within range to achieve Fleetwood's goal for this plant which was to build houses according to the EPA EnergyStar Building Option Packages (BOPs) for manufactured housing, Climate Zone 4, and to attain a less than 5% duct leakage rate (Qn,total #5%). The houses showed some need for additional envelope sealing which was implemented after the first house was tested. The other two houses showed marked improvement in whole house air tightness. Recommendations and test results were provided to Fleetwood via email (no formal trip report). Similar testing was conducted at the Georgia Fleetwood factories in Willacoochee, Pearson, Douglas, and Alma.

Woodland, Washington
Category C, 222 homes

Industry partner Greenstone has been working with BAIHP staff and SGC/E-STAR manufacturers to evaluate a hybrid floor insulation system. These systems, composed of one R-11 belly blanket and R-22 blown cellulose insulation eliminates over-compression and reduces the chance of leakage during transport and set-up, while minimizing material and labor costs. Fleetwood Homes of Washington adopted this system for all of their homes in 2001. Other manufacturers have adopted the hybrid floor insulations system, which provides less insulation voids and reduces first cost of R33 floor system over 3-R11 fiberglass batts. One potential consequence of using the hybrid system is increased moisture in the belly; in 2003, BAIHP staff installed data loggers in two homes to determine whether this is a problem; after the data loggers were retrieved in 2004, BAIHP staff submitted a report to Fleetwood suggesting no dew point problems within the floor system (Figure 18).

Figure 18 Temperature and Dew Point Under Hybrid Floor Decking

• Florida International University, 2005 Solar Decathlon
  Miami, FL

FSEC provided technical assistance to FIU (Florida International U.) for the 2005 Solar Decathlon (http://www.eere.energy.gov/solar_decathlon/). An introductory meeting was held at FSEC in October 2003. Subsequently, a design competition was held among FIU students and the team, comprised of architecture and engineering students, to merge the 10 winning designs into a single conceptual design. In April, the team met with BAIHP researchers at FIU to review the schematic drawings and model.

Researchers discussed strengths, weaknesses and technical needs of the schematic design including cooling loads and strategies for mitigating each (reflective roofing, advanced glazing, shading, ventilation, point source moisture exhaust, etc.), building integrated solar (PV) systems, solar water heating, mechanical system design, energy storage, construction challenges, and the aesthetics of energy efficiency. Students plan to use ray tracing capability of the CAD tools they are already using to study shading and daylighting and will schedule another review with BAIHP researchers this summer as they move into design development.