In late 2019 I wrote a post called A Philosophy of Outbuildings, attempting to capture the lessons from 40 years of making and using utility structures on the land in Five Islands, and now a bit here in Gorham. In the summer of 2020, as part of our garden expansion I took down our oldest, smallest ground-mount PV array of four modules, which was mounted on a rack of untreated 2×4 that was rotting up from the ground. We needed a new rack for the PV, and we have also talked for a long time about eventually getting chickens here. I had already worked up a mostly-effective technique for using quality aluminum foil tape to weathershed the seams between modules on a ground-mount array, making a pretty nice garden shed that might become a third essay in this series. So a solar chickenhouse seemed like a decent idea.
One of the principles of the philosophy is that Small Outbuildings Should be Portable – because they can be, because it increases the value of the time put into building them, and because otherwise they often get in the way of bigger plans that come along later. In the case of grazing livestock, portability is a primary requirement, to expose the critters to fresh pasture, and the Chicken Tractor is at this point a classic DIY project. My parents have taken that concept to the extreme in Five Islands with two small barns for laying hens, built on the four-wheeled frames of retired haywagons from the Holbrook dairy operation in Woolwich. These are truly impressive structures, complete with large winch-up flyways, and those hens get moved every 2-3 days, but a conventional stick-frame rectangular building on a long, spindly four-wheel wagon will come under some pretty intense wracking strains when pulled over uneven ground, and indeed the first iteration ended up somewhat diagonal due to this fundamental challenge of kinematics.
The wracking of the four-wheel chicken tractors offended my engineer’s sense of propriety, much as an out-of-plumb outhouse grates on a carpenter’s, so I resolved that our portable solar chicken house would be kinematically correct, with a three-point stance to avoid wrenching when transported. This complicated things significantly, but the end result was satisfying enough to document in this post, with some lessons learned that could be useful in designing similar buildings.
The size of the building was roughly set by the four PV panels to be mounted, older 185W Chinese mono modules that measure a bit over 2×5′ each. To minimize footprint and provide a bit of standing headroom inside, I arranged them 2×2 in portrait format, leading to a collector a bit over 5’wx10’h, set at 45 degrees (we’re at 44 degrees north here), giving a right-triangular wedge for the main mass of the structure.
When tilted up, the run of the collector array fell just short of 8′. But for efficient use of materials, and to provide a bit more floorspace, I went for a full 8′ in the north-south dimension. To cover the gap at the north end I put in a small north-facing roof facet which I covered with twinwall polycarbonate to let in a bit of light; this also echoes the design of the (decidedly non-portable) solar garden shed that sits just to the north. In the end this resulted in a lot of picky carpentry; it might have been best to tilt the array slightly flatter than 45 degrees and deal with the reduced headroom, but I am happy with the result. With our without the north-facing rooflet, the north wall is the only one big enough to receive a door, and is also a convenient place to implement daylight/ventilation.
Inspired by a coop Gerry Carroll built in NJ, I was keen to try accessing the laying boxes from outside, and the building could use a bit more width to give it stance against blowing over, so I slung two rows of laying boxes on either side running north-south, designed to have hinged access hatches on the side walls and de minimis shed roofs to cover. I clad the laying box sheds with salvaged mobile-home skirting in a nice gray-brown faux woodgrain color, scored from Dave’s salvage collection. We’ll see how the external-access laying boxes work out if we ever get around to populating it with chickens.
The kinematic foundation is the most unique element of the structure, and fortunately I have some decent pictures. Because the shed would be located within our large fenced garden area, it would probably move relatively infrequently (on an annual or slower-than-annual rotation) to allow the chickens to fertilize a fallow area, so I decided to use skids rather than wheels. The kinematic principle could (and I would say should) be applied to a wheeled structure with minor variations. Like the portable tractor shed I built upriver, the skids are made of ordinary treated lumber, 4×6 flatwise in this case. Here are some shots of the substructure under construction, showing how the floor can pivot freely and be adjusted independent of the base:
Built upward from the skids are a front crossbeam (2×10, sculpted top and bottom to increase clearance) and a back crossbeam (C-channel construction of PT decking, similarly sculpted below for ground clearance), plus 2×6 diagonal braces to keep the substructure square. The front crossbeam has metal plates sandwiching it fore and aft, through-bolted (in this case the plates happen to be leftover tail vanes from a prototype wind turbine). The curved metal plates project well above the crossbeam, and form a trunnion that accepts one of the joists of the floor deck with a large pivot bolt. This gives the superstructure a roll degree of freedom relative to the substructure, which is key to preventing the building from wracking as the skids move semi-independently to conform to the surface contour of the soil. Additionally, the flex in the metal plates, the wooden joists, and general slop in the system allows sufficient pitch degree of freedom such that the floor can be leveled.
The ability to level the floor is not necessary to the fundamental goal of a kinematically appropriate structure, and the chickens probably won’t care, but as the son of a carpenter I feel that certain standards must be upheld. So the north end of the structure is supported off the rear crossbeam by two salvaged scissor screw jacks from light cars, which allows the floor attitude to be adjusted, taking advantage of the roll and (modest) pitch degrees of freedom of the front pivot point described above.
Between the solar panel roof and the specialized kinematic base, I tried hard to keep the structure from getting too heavy. The floor is 1/2″ PT plywood framed with a 2×6 perimeter, but joisted with 5/4 decking. The wall framing is similarly light, mostly decking ripped in half lengthwise, and the side walls of the structure are diagonally planked with Hammond Lumber’s thinner, nicer-grade shiplap that is rough one side but measures an actual 3/4″ thick (where regular shiplap is actual 7/8″ thick). I also incorporated some hardware cloth and twinwall polycarbonate for light and ventilation. Here are some photos of construction:
We got an early snow in the fall of 2020 that then melted, and I took the opportunity to skid it into the garden to its final (for now) home:
The structure moved nicely, and it was almost a letdown how trivial it was to level it in place with a couple cranks of the screw jacks. Because the superstructure is intentionally built light, it is secured downward to the skids and the jacks with light chain and preloaded by turnbuckles left over from a fencing project. I bolted on the PV modules, and it sat through the winter and spring while other projects took priority, including growing and harvesting a nice crop of rye (which deserves a post of its own):
I finally got to ‘finishing’ it, roofing the laying box sheds with the trailer skirting and building a door on the north wall this summer. All in all, a satisfying project, currently storing gardening sundries and awaiting a shipment of chicks someday.
Naturally in the course of this type of exploratory project, I came up with some things I’d do differently next time. Most basically, the construction makes the superstructure a few inches taller than I think is necessary, and fitting the kinematic degrees of freedom into a shallower package would be cool. This might be accomplished by dispensing with the metal trunnion plates and fitting the front crossbeam between two closely-spaced floor joists. This would also require a more compact arrangement for the adjustable elevation on the north end, and while the scissor screw jacks were free from Dave’s dump collection and well-suited, they have a temporary, inelegant feel about them; in building another one I might try to substitute some stumpy homemade turnbuckles with welded end plates, fabricated from a couple large-diameter left-hand threaded nuts and bolts. This would obviously also provide downforce, eliminating the need for the turnbuckle/chain.