Roof Framing Defects; Case Study 1-2004
by John F Mann, PE
Villages of Hamilton
Villages of Hamilton is a development of 130 townhouses built around turn-of-the-century in Hamilton Township (Mercer County), New Jersey.
End unit; April 2004
As is usual practice for a new development, the architect worked directly for the builder.
In early 2004, structural defects were found at one townhouse, during follow-up consultation after general home inspection for a real estate transaction. This led to discovery of extensive structural defects with roof framing and supports of more than 20 townhouses throughout the development.
Below are several photos showing roof framing defects.
Based on available information, the builder never did make proper repairs for most of the major structural defects identified, even after at least two meetings of the homeowner association attended by numerous owners, code officials and builder management.
Support for I-beam Type Roof Joists
For many of the attached townhouses at Villages Of Hamilton, building plans specified sloped wood I-beam type roof joists to form a long sloped ceiling. However, the architect failed to include (in design plans) basic support and connection details for the I-beam type joists.
Wood roof trusses were also specified. Architect failed to provide any details for hangers, tiedown connectors or permanent lateral bracing.
Of course, at this point, you may wonder why code officials did not flag these major deficiencies with building plans during plan review. Although detailed discussion of that general issue is for another article, the short answer is that; (1) Code officials are not fully qualified to review and understand design plans since they are not architects or engineers, and (2) Architect (and engineer if involved) remains completely responsible for building design, whether code officials perform any plan review or not.
Roof framing must be designed to resist downward (gravity) loads and upward loads due to wind uplift pressures on roof surfaces, as specified by the governing building code. Design for wind uplift is often performed incorrectly, or, as in this case, even completely ignored.
Photos below show variety of structural defects with roof framing and supports that occurred primarily due to lack of any design details on building design plans, along with failure of builder to obtain necessary details from architect (or other design professional such as qualified structural engineer).
High ends of long (20+ feet) wood I-beam roof joists supported on edge of a horizontal 2x4 nailed to short end-verticals of roof trusses. Roof joists, spaced at 16 inches, do not align with roof trusses, spaced at 24 inches.
54 Kingston photo 7 042204.jpg (JPG — 299 KB)
Bottom flanges of I-beam joists bear only on the edge of the 2x4 (photo above). Such method of support does not comply with standard specifications of the joist manufacturer (Trus Joist). Most important is that there is no lateral bracing at end of roof joist. Also, there is no tiedown connection to resist wind uplift forces.
During investigation, joist manufacturer was asked to comment on some of the defects shown here. However, no response was obtained.
Twisted / tilted roof joist between roof trusses.
Method of support varied from townhouse to townhouse. Builder did not provide any documentation showing as-built support details, as required by New Jersey UCC regulations.
Many I-beam roof joists are also supported by a long LVL beam. In many townhouses, framers made large notches in the roof joists, completely cutting away bottom flange of joist, to provide a flat surface for the joist web to bear on top of the support beam. Such method of support may be grossly defective without specific design by a qualified professional, in accordance with joist manufacturer specifications.
High ends of I-beam roof joists supported on top of an LVL beam, with bottom flange of joist cut away. Web is reinforced with plywood. This "detail" was made-up by the framing contractor. Plans did not show any support details.
As-built conditions at high ends of I-beam roof joists vary greatly.
There is no tiedown connection at high ends of roof joists other than (perhaps) one or two toenails. Ends of roof trusses are supported in hangers. However, there is often no nails through the truss end-vertical.
High end of roof joist notched to fit over beam. Note gap between beam and reinforced web of joist. Method of support is not clear. However, nails have been installed through flange and web of joist into end vertical of roof truss (also supported by beam).
At edge of roof, where edge joist abuts inside end of the gable endwall truss, high end of edge joist (center top, behind truss connection plate) clearly has much reduced depth, with web bearing on top of beam.
At this townhouse, high end of roof joist at end of attic space (up against end of endwall-type roof truss) has no support underneath at all. The horizontal 2x4 supporting interior joists is run across opposite face of the roof truss end-verticals. Drywall is part of a gypsum fire-separation wall. Edge joist is likely supported by wall studs below, however wind resistance along top of wall is then provided only by the ceiling drywall which may be grossly deficient without adequate blocking. Plans show no details.
High end of roof joist bearing on horizontal 2x4 nailed to end-vertical of roof trusses. Severe notches apparently made for fitting over beam (see next photo).
High end of roof joist notched to fit over LVL beam. Joist is bearing on the cut web, partially reinforced with plywood.
122 Kingston photo 38 040904.jpg (JPG — 554 KB)
Support For Roof Trusses
The code (TPI) governing design of roof trusses specifies that, when a truss is supported by another truss, the connection must be designed by the truss designer. In practice, such connections (such as hangers) should be specified on truss diagrams, with complete nailing. However, this essential code requirement is very often ignored by the truss designer, architect, builder and code official. Even when the truss designer specifies a hanger, the required nailing is often omitted.
At Villages of Hamilton, many roof trusses are supported by a "truss girder", also termed "girder truss", that is designed by the truss manufacturer. Yet, required hangers were not specified by truss designer on truss diagrams.
Hanger supporting end of roof truss, up against side of truss girder. Note that many nails holes are not filled since there is no wood behind the hanger. Also, two nails are installed at the edge of truss members (web & bottom chord), resulting in lack of nail shear capacity and risk of splitting wood.
Interior end of roof truss, barely on very edge of hanger seat. Hanger is nailed to LVL beam. End of another truss, along left side, is not supported in this hanger, or by any other obvious method, other than being nailed to the first truss.
122 Kingston photo 24 040904.jpg (JPG — 464 KB)
Ceiling Drywall Cracks Under Main Roof Support Beam
In almost all townhouses, high end of I-beam roof joists and inside ends of roof trusses are supported by a long LVL wood beam (seen in photos, typically with a medium to dark brown coloration).
A drywall joint occurs directly under the beam, at upper end of the long, sloped ceiling formed by the I-beam roof joists. In just about every townhouse inspected, this joint is cracked. Several owners reported numerous attempts to repair the cracked joint, without success.
The problem, due primarily to deficient design, is that the LVL beam is excessively flexible to prevent cracking of ceiling drywall, even for ordinary loads. The fact that repeated cracking of drywall has occurred throughout the development, even without full design load applied to the roof, demonstrates the design defect.