It shouldn’t surprise me how often I see someone check an internal corner for square using a basic 12″ square or worse still a speed square, but it’s not something to rely on most of the time. Even carpenters do it and whereas today they think it’s a good and positive test, in the past, furniture makers and joiners generally relied on it only as a quick and rough check in lieu of a more accurate one yet to come. Imagine testing the corner of four foot by six-foot frame with a speed square or even a larger framing square. It could possibly be a quick check to see how close it might be but overall it is not a good way of testing for square at all. If this surprises you then the following will help you better understand.
Mostly it’s to do with distortions that occur in the wood due to moisture levels in the wood, distortion caused through clamping joints, flex in wood and then too a range of other reasons. When we make frames and boxes our aim is always to start with straight lengths, a virtual impossibility in long lengths, but even in shorter lengths like the sides of boxes the most minor discrepancy in fractions of a millimetre will throw the square exaggeratingly off.
There are ways of checking for square without using squares and this is much needed when checking projects such as door frames, doors, window sashes and frames, boxes and so on. In fact, there is barely any project that can simply be checked with just a square alone. I would never check any project with a square whether the square be large or small. Rarely will it give a good reading and rarely could you ever rely on it. As it is the rarity and not the norm, I understand the question mark . . . it’s automatically questionable. The reason is that wood generally changes from straightness to slightly curving and though it may not be clearly visible, it keeps changing throughout its life, whether in the raw as lumber or in a finely finished piece. Our task is to minimise that liability and then too work ‘with‘ the wood to either mildly constrain it or totally allow it. A door, the lid of a box, a window sash in its frame all hang freely at two or three points by hinges. They are free to expand and contract according to the surrounding relative humidity. In fixed frames and boxes, we always strive for squareness but if the stiles or frame heads and sills are even slightly bowed we cannot rely on a square to test out for squareness; the registration face for registering the square against is compromised. Using a square in these circumstances will not work for us at all. Make a window frame a metre long and three-quarters of a metre high and it is most likely that the 12″ square in the inside corner will not tell us much of anything and certainly not the accuracy of 90º.
So what’s the answer? The only way to really check for square is to measure the two internal diagonals from inside corner to inside corner. If the two distances are equal then the frame is square. If there is a discrepancy then the difference between the two measurements must be halved and this split number will be how much out of square the frame or box is. I generally measure for a quick reference but I also use what we call a squaring rod where I can mark the two diagonals and then s[lit the difference using pencil markings. By marking the two diagonals onto the rod I can then split the difference to show where the frame can be pushed to to get it dead square.
For frames of different types, pushing the project to evenness and bracing one of the corners will usually keep the frame square, allow installation and fixing, and the frame will remain square. On boxes, we can maintain squareness by fixing the bottom to the sides or using grooves to house the bottom and keep it square that way. This is mostly for larger boxes as cabinets. Small boxes remain square after the glue goes off as long as you clamp them and they are square beforehand.
Fortunately, wood loses a percentage of its elasticity over the years and stabilises. Our aim is to get the frames and boxes as square as possible. The shoulder lines are the absolute first step to this. Working accurately at this micro level, the shoulders of our tenons and dovetails, things around there, usually translate into good levels of extended squareness throughout a project. Exactness in distance between shoulders is another critical point too. Any variation there will cause out-of-squareness and parallelity too. When this happens, we either need to correct or whack it out to even out the discrepancy. In big frames, a millimetre or two makes little difference, on a small box it can be more critical.
Sometimes it is not possible to go internal corner to internal corner because of some obstruction. In such cases, you can mark off a distance to each side from one end. Now measure diagonally from the inside corners to the marks. If they are equal in diagonal distance then the frame is square. If not, any discrepancy halved will give you the amount that it is out of square.
Though you can simply use a tape to measure internal corner distances, it is easier to use something more rigid like a length of wood with a point to fit neatly into the corners . . .
. . . but sometimes a small headless nail to the side of a stick at one end will work just as well.
As a final thought, remember that placing clamps out of parallel when glueing up can and will pull the box or frame out square. All clamps should always be parallel to the sides or frame and checked for square as discussed. If the project shows out of square still when checked than you can run the clamps out of parallel to correct the discrepancy too. Very hand.