On Monday, the Pennell crew erected the ell by hand. They had a roustabout on-site, which is like a more portable, telescoping gin pole, but the bents were light enough to raise with a crew of four. The ell, a drop-tie frame built in the mid-1800s, was dismantled earlier this spring during the first phase of Pennell House repair. The frame parts were transported back to the shop in Berwick, repaired, and test fit. Our most recent North Bennet Street School intern, and newest employee, Joe McAllister devoted his final school project to the cutting and joining of two additional bents to prepare the frame for re-use as a contemporary kitchen.
Following the ell, repairs to the house were extensive. The frame was lifted on steel I-beams in order to replace the foundation and completely rebuild the undercarriage. Seven of the eight house posts needed repairs, two of which required full replacement. The first floor studs of the north, south and east walls all required lap repair or replacement. Along the north eave, all three second floor girts and eight of their associated braces were replaced. Ultimately, the entire east gable bent was completely dismantled, repaired, and rebuilt, while the rest of the building was left standing. See “before” picture, above, and “after”, below.
Revisiting the job-site this week, I realized that the diversity of joinery matched the broad scope of repairs. The decision to use a particular scarf, spline or lap joint is dependent on a number of factors including location, level of deterioration, difficulty of installation, historic significance, and whether or not the joint will be in tension, compression, or subject to twisting. For joinery enthusiasts, I’ve recommended Historic American Timber Joinery, by Jack Sobon, and I’ll recommend it again; it is the definitive reference manual for those pursuing traditional repair of historic timber framed buildings (I’ve linked to a PDF, if you want a hard copy, it’s worth ordering one from the Timber Framer’s Guild). On a hybrid job like this one, combining preservation, energy retrofitting and adaptive re-use, we used both traditional scarf techniques like those in Sobon’s book, and contemporary approaches, like splines and free tenons. Paradoxically, sometimes the newer repair techniques are able to preserve the most original material.
If you visited our site before, job or web, you’ll know that our bladed scarf is an old standby. It works well for post fixes, because the keys prevent the joint from slipping or twisting under outward pressure. The outward thrust of the rafters from above, in combination with the possibility of a rolled sill, and the inward tension of the tie beams and tie girts, means that a post scarf should have some means of “locking” to prevent slippage. This could also take the form of an under-squint (see below) but in this instance, we prefer the square-bottomed keys of the bladed scarf. These multi-directional forces are what make a simple lap joint inappropriate for post repairs. We expect our repairs to last for as long as the building has already been standing. Over the course of 150 or 200 years, there may be shifting in the foundation, or deterioration in a sill, that would complicate the pressures acting on our post fix. A bladed scarf joint is designed to withstand those forces, so that in 100 years, if a sill needs to be replaced, the post foot and associated repair can remain intact. (I hope Athena, protector of woodworkers, notices that we strive only to double the lifespan of a building, we don’t expect to triple it.)
A bladed scarf is also used to repair an unsupported section of sill. When a sill, summer beam or floor girt is supported on posts or piers, rather than a full foundation wall, it needs a repair that can support itself and prevent sagging without introducing metal brackets or plates. The introduction of big plates of metal, especially in potentially moist environments, like a basement, risks the danger of condensation and its dreaded associate, rot. While a lap joint may be sufficient for many sills, on stable, continuous foundations, the keys in a bladed scarf give it compressive strength perpendicular to the joint.
In a timber under considerable tension, such as a tie beam, a bladed scarf joint may not be appropriate. The joint has considerable resistance to compression and twisting, but relies on pins and friction to prevent spreading. We’ve long used a tabled, wedged joint to prevent spreading in tie beam repairs, but at the True-Randall Farm in Montville, we encountered a stop-splayed, under-squinted and wedged scarf that had been used to extend the length of tie beams by more than ten feet. The barn was moved over a hundred years ago, and the joints had loosened, but held up considerably well under the strain of a crumbling mid-century concrete block foundation. The biggest threat to a barn’s frame is water infiltration. When tying joints fail, allowing plates and rafters to spread, roof leaks can result, leading to water infiltration that will accumulate on any available horizontal surface: plates, girts and, often, all the way to the bottom of the frame, at the sill.
We used this stop-splayed, under-squinted, and wedged scarf joint to repair the east gable tie girt. The east gable bent contained two tying timbers: a tie beam above, which runs from plate to plate and required full replacement, and this tie girt, which supports the second floor joists, and is fully supported by studs from below. The under-squinting is the little angled cut two inches from the top and bottom faces of the timber; this angled cut also helps to mechanically “lock” the joint, and prevent twisting.
Elsewhere in the house, we found another instance of under-squinting, in a slope-shouldered bolster used to repair a post. This was a really cool fix that was probably installed sometime after a renovation that involved hacking out the interior faces of the posts, so that they wouldn’t intrude on the interior wall plane (how dowdy and old-fashioned!).
This assault on the frame resulted in some posts being sliced in half, immodestly revealing their joinery. Unfortunately, this hackery also removed the bearing shoulder of the post which formerly supported the ends of the second story floor girts. The bolster above was installed around 100 years ago to support the end of one of these girts.
Three second-story girts along the north eave were rotted and needed full replacement. The second floor joists fit into cogs cut into the interior faces of the girts, and stayed put during installation of the repairs. Likewise, the four north eave posts could not be moved (the east gable post was replaced in full, but needed to be installed before we replaced the girts). The conditions created by a standing frame required that we use a free tenon or spline connection to repair these elements. We cut the girt to length, shoulder to shoulder, and cut a slot in the underside of the timber the full length of the free tenon.
We install the girt between the standing posts, and then insert a free tenon into the slot. Then we slide, or pry, the tenon laterally so that in engages with the accompanying mortise in the post. Lastly, we plug the gap that is left in the girt. In those instances where the posts are in better condition, the post mortise can be extended. The free tenon is inserted below the girt and slid up into the slot. Then the extended mortise in the post receives a plug (see above).
Seven of the house’s eight posts required repair or replacement. In each case, we preserved as much original post material as possible, resulting in some fairly idiosyncratic fixes. In the one pictured above, an interior corner of the post had been removed in the previous “renovation” and required a slightly more complicated version of the bladed scarf joint.
The twin threats of squirrels and rot wreaked havoc on the east ends of the north and south plates, requiring a scarf repair for each of them. Plates endure considerable torque, created by the outward thrust of the rafters and inward tension of the tie beams. Lee used a halved and bridled scarf on the ends of these timbers in order to retain the most material, and prevent twisting or rolling.
The original ell plate was full length, and in good condition. The plans required an extension of the ell by two bents to accommodate a contemporary kitchen, but we didn’t want to remove any more original material from the plate than was necessary. The plate had its own interesting joinery, worth preserving, in the form of a rabbet along the top interior edge, that caught the birdsmouth on the rafter tails. A traditional scarf joint would have required the plate to be cut back as much as two feet. Instead, Ed designed a spline-joint, that connected the original plate, the new plate extension, and the post, all in one (above).
Three quarters of the first story studs required repair or replacement. Where feasible, we used a simple half-lap repair on the athlete’s feet of rotten studs. The half lap, instead of a full length stud replacement, allowed us to replace studs with tenons on either end, even when the stick was captured by a sill below and second-story girt above.
Often, we encounter bolts and L-brackets employed to little effect. Sometimes, these metal band-aids do more harm than good, due to the introduction of large plates of metal, against which water can condense and be held against the timber, or by creating a tensive or compressive force where it does not belong. There are instances, however, where a metal bolt or bracket is the best solution. The tie beam ends at the Pennell House were one such case. Each of the 5 remaining tie beams showed various levels of rot at either end, outside of the plate and the rafters’ birdsmouth. Other than the east gable tie, none of ties were rotted enough to require a scarf repair. However, the joinery on the end of the tie, the angled cog capturing the rafter’s birdsmouth, needs to resist considerable force, especially from the rafters that carry the cupola.
We wanted to ensure that the tie beams would continue to prevent the bottoms of the rafters from spreading. Ultimately, we used a combination of 3/4″ threaded rod, and Simpson-brand L-brackets to create an economical solution to this pervasive, but relatively minor, problem.
Preservation work can be frustrating, because every building is unique, and every problem is interconnected with others. The lack of a universal solution makes preservation work almost as difficult to estimate as it is to execute. Fortunately, it is the very same combination of variety, unpredictability and creative problem-solving that makes this work so much fun.
For more photos of our process at Pennell; please visit our Flickr page.