Although we had been very meticulous about checking the trailer and subfloor for squareness by measuring full-length diagonals across it, we didn’t - for whatever reasons - do likewise with the raised framing walls.  I think we were lulled into some complacency by several things: (1) the fact that the floor kept appearing to be both remarkably flat and level in all directions - as far as we could tell from a 4-foot level and from shedding of rain water, (2) all the framing members were cut VERY precisely (my work) and fitted together tightly (Mikey’s), and (3) occasional checks for plumb (with our aging, cloudy, difficult to read 4 ft level) indicated that studs were vertical. Also, (4) when wall sections were raised, the top plate pieces met up with each other almost perfectly and did not either leave gaps or collide.  What else could be more reassuring?

Well, our negligence became apparent when the 4x8 sheets of sheathing did not seem to fit together quite as tightly as we expected while mounting sheets around the back door, while trying to keep both the horizontal and vertical edges of each sheet in their proper places over corresponding framing members.  The discrepancies noted were on the order of 1/4 of an inch or less, but they were mystifying nonetheless, and annoying because they were in a whole ballpark larger than the tolerances we had been trying to work to thus far.

So, a hiatus in construction was declared and a full investigation ensued, taking place over the last two days at least (with me doing measuring and texting Mikey at work with my curious and sometimes reversed findings).  We had at first made a lot of ad hoc measurements with a long tape measure, inside and outside the framing, and concluded that the worst problem seemed to be that the 24’ back wall was not a true rectangle but instead was a very regular parallelogram, with opposite sides equal in length but apparently skewed somewhat in the direction of our (big) house.    We couldn’t tell just how much the opposing, parallel sides were offset from each other, but measurements of the diagonals proved to be a full 1/2” different.

The fact that sheathing had already been applied with a very high quality, practically indestructible adhesive (PL Premium) - partially for the very purpose of ensuring that the wall would never become skewed out of square - now worked ironically and embarrassingly against us, as we contemplated in hindsight how best to correct a problem that would have been fairly easy to avoid earlier.  (Another opportunity to demonstrate the appropriateness of Grandpa’s comment about more sense being shown getting out of a problem than getting into it.)

We thought quite a bit about doing whatever would have to be done to remove the five sheets of sheathing already in place, and we contacted Chris, our informal and unpaid consulting engineer about what to do.  He spoke within a few hours to a contractor friend who described how to remove the glued plywood, and it did not sound like a process either rapid or pleasant.  The plywood would not be re-usable, and the framing would be damaged if not treated carefully. A lot of sanding of leftover dried glue would have to be done.

Obsessed as I sometimes can be with any process of collecting lots of data both useful and useless, I hit upon an idea two days ago (Thursday, 9/1) for a good way to get some very accurate measurements by myself of all four walls and possibly the overhead framing that will be the roof before long.  I proposed making measurements of the insides of various wall areas and openings, marking out the four corners of each area to be checked by using the inside corners where vertical and horizontal framing members - studs and plates - come together.  To facilitate measuring accurately and without Mikey’s help (he being at his day job), I cut a bunch of short 1x4 stock pieces to tack tightly into the inside corners that I wanted to measure from.  Measuring an inside dimension accurately and alone then became a simple matter of hooking my tape measure onto the protruding end of a 1x4 block and unrolling the tape to the protruding end of an opposite block (vertically, horizontally, or diagonally).  The blocks were tacked in place with 6d nails, in holes pre-drilled at useful angles to make hammering in (just part way) and removal easier, without splitting the blocks.

I measured areas which corresponded closely to each of the four full walls; then Mikey suggested measuring the areas formed by various groups of several adjacent stud cavities to check on local distortions from squareness in specific smaller areas.

After compulsively recording lots of juicy numbers, I made a whole science out of analyzing the degree of distortion in each area measured.  I hit upon determining (1) any difference between the two diagonal measurements of the area and (2) the horizontal run of the area, and then (3) the quotient of (1) by (2), which I multiplied by 1000 and dubbed the “distortion ratio.”  Values of this parameter were calculated for each area measured and came out to numbers between 1.9 and 10.5, with lower numbers corresponding to less skewing. (A value of zero would indicate perfect squareness).

Then yesterday, at Mikey’s repeated urging, I ceased compulsive tape measurements and started checking what could be learned by suspending a plumb bob (which I dubbed Robert the Plumber) at various points on the framing.  Wall area measurements, had, as mentioned above, indicated that the 24-foot back wall was a parallelogram leaning away from the Grove and toward our house, and its distortion ratio overall was about 2.0.  The biggest finding of Robert, however, was that the vertical end of that wall was indeed just that - quite vertical!!   That meant that for the whole back wall to be skewed, the FLOOR must not be level but must be sloped slightly away from the big house!

Apparently, if you are attempting to build a really square structure on a trailer, you need to be aware that the trailer itself is going to bend or bow somewhat in the middle, where the supporting wheels have springs which compress as the structure gradually becomes heavier.  In our case (and likely in others) the four corners of the trailer are supported and stabilized by scissor jacks.  Before construction began, Mikey adjusted those four jacks as well as possible to assure the trailer would be both level and flat, and as construction has progressed, we have occasionally “eyeballed” the trailer frame from front to rear to see if it is remaining straight.  It has seemed to be very slightly bowed upward in the middle over the three axles.

Clearly, in order to avoid problems with sheathing and siding fitting properly (not to mention interior finish components) it is desirable to ensure continually that the floor stay as nearly as possible both level and flat.  We have not tried but have thought it might help to work somehow with a 25 foot piece of clear plastic tubing and some water to make an effectively very long level and check the floor with it.

It would seem that when a full-sized house is built on a poured concrete foundation, the issue of ensuring that the foundation surface is level is resolved by the simple tendency of the semi-liquid concrete to level itself when poured.  Then you could be confident that framing is square if vertical members are all kept perpendicular to the foundation; but in a house built on a trailer, things are more complicated.

We are probably going to go forward working with the discrepancies that exist now (they are really a long way from being showstoppers), but if we build another house sometime we will (1) keep checking the floor for level and flatness, (2) square up each wall section as it is assembled and put diagonal bracing on it before raising into position, where its fit at top plate level will verify whether the entire wall being put together is square, (3) take diagonal checks on each fully framed wall just before installing sheathing, and (4) buy a nice big level that can be easily read.