Tucker Hill Bridge

The Mad River Path Association (MRP) manages a series of trails in the towns of Warren, Waitsfield, and Fayston. The trails, nearly 10 miles in all, range from the flat, pastoral Mad River Greenway along the bank of the eponymous river to the steep Vista Trail that gains an overlook behind Warren School.

Last year MRP removed a failing bridge on the Millbrook Trail, which runs 2.1 miles from the Mad River Barn to Tucker Hill Road. The missing bridge breaks the trail in two for everybody except those hardy hikers and mountain bikers willing to ford a stream. Before proceeding with construction of a replacement bridge, MRP needed a Professional Engineer to approve a new design.

Proposed footbridge location marked with stakes.
Proposed footbridge design by architect.

In this case PERCH did not design the bridge – an architect did. But PERCH did what an architect is not qualified to do, which is to verify that the design is structurally sound. The starting point was a set of sketches by the architect, Charlie Hosford, which show the bridge from several perspectives and label the materials and dimensions. Conversations with Charlie helped PERCH determine what was needed: an analysis of the structural components of the bridge, and a design for key connections.

The primary loads on the bridge are its own weight (dead load), pedestrians (live load), snow, and wind. According to NCHRP Guide Specifications for the Design of Pedestrian Bridges, the standard live load on a footbridge is 90 pounds per square foot, which equates to adults standing shoulder-to-shoulder across the entire span. This load case might occasionally happen in a city park, where pedestrians might crowd onto a bridge to watch fireworks, but it’s ludicrous for a rural trail. MRP proposed that they could post the bridge for a load limit and bypass the NCHRP provision, which in any case is a guide rather than a legal requirement. This allowance makes a slimmer (and cheaper) design possible.

Part of girder splice design by PERCH.

PERCH confirmed that the decking and girders are adequate as designed for a live load of 30 pounds per square foot, or up to 25 persons spaced equally across the bridge. PERCH also issued a connection design for the girder splices, using Simpson metal plates and a specific nailing pattern. An Engineering Report was submitted to the Town of Fayston, and I’m proud to report that they issued a building permit right away. Construction will proceed this summer.

The permit!

Mind the Gap

Jeff bought his home in the middle of summer several years ago. The first winter, a gap opened in the drywall between the gable-end wall and the vaulted ceiling. The gap seems to close every summer and open again every winter. Jeff hired PERCH to diagnose the problem and recommend solutions.

What on Earth was going on here? My initial investigation ruled out several possibilities. The gap is seasonal, not progressive (although Jeff does think it gets worse each year), so it doesn’t indicate a problem with the superstructure but something environmental. Frost heave seems unlikely, as Jeff has no uneven floors or major cracks elsewhere in the house.

The gap between gable-end wall and ceiling.

For a little while I was wooed by the idea of truss lift. Roof trusses are known for expanding and contracting in cold climates: in winter, the bottom chord stays warm and damp because of its exposure to inside air, while the top chord and webbing gets cold and dry right below the roof. The top chord and webbing contract and pull up the bottom chord. But Jeff with his vaulted ceiling clearly doesn’t have roof trusses. They must be rafters, no deeper than 2×6, for which any differential contraction would be barely visible.

The best-fitting explanation was not frost heave, but a different kind of heave. Certain soils (typically soils with a lot of clay) are known as expansive soils because they collect groundwater and expand during wet seasons, then lose the groundwater and contract during dry seasons. If the gable end wall was built on an expansive soil, it would tend to drop in the winter and rise in the summer.

But wait. Wouldn’t the rest of the house move, too? I found evidence to the contrary during my site visit. On the first floor, a wall adjacent to the offending wall had a few minor cracks in the drywall. Directly below, in the basement, a crack ran across the plaster covering the concrete foundation. Aha – the walk-out side of the basement is framed by a stud wall, which is much lighter than the concrete walls on the other three sides. The soil under the stud wall hasn’t compacted as much as in the other locations, so it’s more susceptible to subsidence when the soil contracts.

Cracking in the first-floor wall, about 6 feet from the corner.

Sometimes structural engineering is like solving a mystery. I search for clues and weigh possibilities against the evidence, and hitting upon the right answer is very rewarding.

“Once you eliminate the impossible, whatever remains, no matter how improbable, must be the truth.”

Sir Arthur Conan Doyle

All’s Well With an LVL

Chris is renovating a 120-year-old house and wanted to open up the floor plan. Chris hired PERCH to design a pair of beams that could replace the existing loadbearing walls, spanning the 18-foot room width and picking up the existing floor joists. Chris wanted to procure the beams inexpensively and install them easily, with little or no disruption to the ceiling height.


Right away, using a steel wide-flange or T shape was out. Steel beams are expensive and heavy, and connecting them to the stringers would have required either complex details or temporary removal of the stringers. Ordinary dimensional lumber wasn’t strong enough for such a long span within the confines of the ceiling height, so PERCH looked to laminated veneer lumber (LVLs) for more strength. A pair of 3.5-inch-wide, 9.5-inch-deep LVLs were strong enough to get the job done. Joist hangers allowed the LVLs to share depth with the 2×8 joists, so the lumber only protruded down an inch from the existing finish ceiling.

The new header beam on the right requires support down to the foundation at the location marked with a red circle.

PERCH also investigated a hybrid solution called a flitch plate: a steel plate sandwiched by two pieces of lumber. In theory the steel increases the assembly’s overall elastic modulus while the lumber keeps the beam light and enables nailed connections. But in the final analysis a flitch plate was too complicated. The beam would need to be assembled first, with through bolts ensuring the wood and steel share loads, and then it would be prohibitively heavy to lift into place.

PERCH developed an installation procedure (which required temporary support of the floor joists) and checked that a column of dimensional lumber built into the existing exterior stud walls would provide sufficient support. Chris followed the procedure and reported that everything went up perfectly.