Long Term Deflection


Has there been any research or studies on long-term deflection of metal plate connected wood trusses? How much deflection is allowed?


For code compliance, the allowable amount of deflection is l/360, where l is in inches. For a stiffer floor it is common to use l/480. You can also use a length to depth ratio, which is typically 18 for good performance. The lower the number, the stiffer the floor.

Whether live load or total load is taken into account when determining deflection depends on the type of truss.

In 1973, the U.S. Department of Agriculture Forest Products Laboratory (FPL) began a study to determine the long-term performance of wood trusses. Trusses with seven different connection systems have been observed under loading for periods of 5, 10 and 15 years. At the end of each period, specimens were unloaded and evaluated for strength and stiffness under laboratory conditions. The trusses were designed according to the prevailing specifications of the time (TPI-65) which, for all practical purposes, assumed frictionless pin-connected joints. Bending moments were analyzed by assuming continuity of truss top chords. No allowance for continuity was assumed for design of moment stresses in truss bottom chords. Assumed to transfer only axial loads, bottom chords were designed as simply supported beams with superimposed moments. Nailed and glue-nailed trusses underwent performance evaluation along with metal plate connected wooden trusses for test result comparisons. As the control group, one truss of each connector system type was loaded to maximum in the laboratory. Three more of each type were loaded and subjected to sheltered or outdoor exposure. After each five-year interval, one truss of each construction was loaded to maximum in the laboratory. Each truss was supported and loaded independently. A load of 20 lbs/ft. (design dead load of 10 psf for a 24 in. spacing) was applied to the lower chords, and a load of 36 lbs/ft. (design dead load plus half of the design live load) was applied to the upper chords. This load represented that which the truss would experience during its service life. Out-to-out span (overall bottom chord) was 28 ft.

Over time, deflection under load (creep) occurred for all the trusses. The total amount of deflection appeared to be related to the rigidity of the joints. After ten years, nailed plywood gusset trusses had a total deflection of three times the initial deflection, metal plate connected wood trusses had a total deflection of two and a half times the initial deflection, and nail-glued plywood trusses had a total deflection slightly less than two times the initial deflection. Most of the creep occurred during the early stages of exposure. Except for nailed plywood gusset trusses, deflection values after 10 years were at an acceptable level, based on criterion for span divided by 360. After 10 years of exposure, there appeared to be no appreciable effect upon strength and stiffness as determined from laboratory evaluation, except for nailed plywood gusset trusses that had a 30% reduction in stiffness. All trusses met acceptable short-term performance criteria. The nailed-on plate trusses met the deflection criteria for span divided by 360.

Prior to this research, the manufacturers of truss plate connectors had performed load tests on hundreds of trusses, in some situations under circumstances far more stringent than those at FPL. Much of the industry research, however, was proprietary and served primarily to assure the manufacturer and the designer that design assumptions were supported by testing. The types of metal connector plates used on trusses fabricated in 1973 are no longer sold in the marketplace. While the connector designs used today differ from those used in 1973, the performance is expected to be similar to, if not better than, the results found in FPL testing.

SBCA Categories: