One Degree
The clamps were on and the glue was already spread when I figured out something was wrong.
Seven boards of cherry, a little over an inch and a half thick, six feet long, forty five inches across when they were all together. This is the dining room table, the one I’ve mentioned here a few times without ever writing about. The top is solid cherry and I’d been looking forward to this glue-up for weeks.
The seams were tight. That’s what threw me. Every joint closed with no gap, no light, nothing to complain about. But the panel wouldn’t lie flat. As I brought the clamps up, sections of it rose and fell. Not much. A sixteenth in one place, a little more in another. Not a cup, not a bow, and not anything I had a name for while I was standing there with glue going off.
You don’t get to think it through at that point. I had extra thickness in the boards, so I committed. I broke the glue-up into two halves, twenty two inches and twenty two and a half, clamped them, and let them cure. Then I ran each half through the wide belt sander and took off whatever it took to get a flat face. It was a lot. Once both halves were flat I jointed the mating edge on each and glued the two together, then finished the seam with a hand plane and a card scraper.
I started at a little over an inch and a half. The finished top is a little under an inch and a quarter.
When the second glue-up was clamped and I finally had time to think, I walked over to the table saw and put a square on the blade. It was off about a degree, tilted to the left, which is not the direction that saw bevels.
Our table saw gets used by a lot of people. The bevel stop is deliberately set a little past zero, below it, so that dust packing up on the stop doesn’t keep you from reaching a true ninety. That’s the right way to set it. What it also means is that the blade can travel past vertical if somebody pushes it there, and somebody had. It was parked past the stop, leaning the wrong way, reading zero on the scale and sitting a degree off.
I know this about that saw. I carry a four inch square in my apron and I check the blade before I rip anything that matters. It takes about eight seconds. I have taught this to members. I have said it out loud in classes.
I ripped those boards on a Wednesday evening and glued them up Thursday. I did not check the blade. I don’t have a story about why. I wasn’t rushed, nobody was talking to me, I just didn’t do it.
A degree is nothing on one cut. Rip a single board with the blade off a degree and you get an edge that leans a degree off perpendicular, and you will never see it, feel it, or catch it with a straightedge across the face. The board is fine. Put two of those edges together and the joint still closes tight, because two mating surfaces will meet whether or not either one is square to anything. What you’ve built is a very shallow hinge.
Then you do it thirteen more times, all leaning the same direction, because the saw didn’t change its mind between cuts.
The panel wasn’t warped and it wasn’t cupped. It was faceted. Forty five inches of cherry that was really a slice of a very large polygon, tight everywhere, flat nowhere. Every cut I made that evening was a good cut. Straight, clean, no burn, no wander. I did the work correctly. The work was not the problem.
I’ve written before about being a lazy woodworker, by which I mean I’d rather solve a problem once in the setup than fight it fourteen times at the bench. This is the same argument coming at me from the other side. Care in the execution doesn’t rescue you from a bad setup. If the fence is wrong, careful cutting gives you careful errors. If the blade is off a degree, forty years of hands will produce fourteen beautiful joints that add up to a top you have to sand a quarter inch off of.
Five sixteenths of thickness across forty five inches of eight quarter cherry, gone into a dust collector, because I skipped eight seconds on a machine I already knew I couldn’t trust.
The table is fine. At an inch and a quarter it still reads heavy and nobody sitting at it will ever know it was supposed to be thicker. I know.
I can’t tell you I’ve fixed the underlying thing. I’ve checked the blade every time since, but I was checking it every time before, too, right up until the evening I didn’t. That’s what bothers me. The habit didn’t erode. It held and held and then one Wednesday I walked past the step without noticing. I don’t have a system that catches that, and the systems I can think of turn into checklists I’d stop reading by March.
Fourteen good cuts. One degree.
