Dining table project (Oct 2017)

Written by lcamtuf@coredump.cx

1. But why?

This was our old dining table, bought about six years ago. At first glance, it was still looking fine:

Well, except... not really. Here's a close-up:

I actually refinished it three years ago - but with three cats and three kids, it was a short-lived success. The table was a cheap veneer-on-particleboard contraption, so it probably would not have survived another sanding. In fact, the slightly darker spot in the bottom left corner is where I accidentally went through the veneer during the first refinishing. I fixed it the best I could, but it was a clear warning sign.

Short of redoing the entire veneer, no amount of wood necromancy could have made the table look whole again. The dark finish only made things worse: the table was close to a large window and every scratch showed up as a white line.

In the end, we decided that it's time for something new. And given that many of the nicer solid hardwood tables cost $2.5k or more, I figured it's a pretty good opportunity to earn back some of the money spent on workshop tools...

2. The plan

Of course, the first step is to come up with a rough action plan. Here's a sketch made in a CAD program:

The top is going to be made out of eight rows of relatively narrow maple boards, each about 1" thick (4/4), 4" wide, and 22" long. Working with small segments makes it easier to pull this off in my relatively cramped workshop - and eliminates most of the issues with warped or imperfect lumber.

To make the top look more hefty while keeping its weight (and cost) manageable, the whole thing is going to be framed with ~2" thick (8/4), 2" wide strips running along the edges. This trick also helps hide the end grain.

The base is supposed to be cheap 4x4 construction lumber. I want to finish the base in some dark color to provide a counterpoint to the light-colored maple on top. I think the contrast would go well with the rest of the room.

The overall size of the table will be ~6 ft by ~3 ft, basically the same as the outgoing one. The height is 29" or so.

3. The top

Well, first things first. Here's about $300 worth of S4S hard maple:

Maple seems like a pretty good choice for a dining table: it has a very light color, tends to have interesting grain patterns, is reasonably hard (about 3.5x harder than pine), and doesn't cost much. Only about three fifths of it should be needed for this project, but I figured it's better to be safe than sorry.

The next step is to cut the boards to the approximate lengths with a circular saw and a speed square. As mentioned earlier, I have a pretty small workshop, so this is being done in the backyard, to be followed by more precise cuts indoors with a bandsaw:

Here are the boards, fresh off the circular saw:

This is closer to what's actually needed for the table, although all of that wood is still gonna be cut lengthwise into more narrow strips. But first, I'm going to run one edge through a jointer to have a consistent point of reference when running the lumber through a saw:

With that out of the way, I'm making final cuts on a bandsaw to turn ~10" wide boards into 4" strips:

The flat edge created by the jointer is to the left, resting against the blue fence. The photo shows the second 4" board being cut; the rightmost 2" strip goes onto the recycle pile. I'm also cleaning up the ends of each board and trimming them to the expected length:

This is done with a miter gauge (yellow) to get the angles right.

I also decided to run the surface that will be eventually facing up through a jointer to correct warping. This was mostly a waste of time, because without a biscuit joiner, I did not get the boards to line up perfectly and still had to plane the assembled top later on:

With cutting out of the way, here's a mock layout to figure out a viable grain pattern:

Note that the thicker edge is going to be flush with the top surface when all is said and done - but that obviously doesn't happen if I just put it all on the floor.

While each board was nominally 4" by 22", you can see that I cut several of them into halves, then put one half at the bottom of the row, and the other at the top. This had the effect of breaking up the pattern, so that the short edges do not line up across the rows. I think it looks better this way. It's basically the trick that's used for most wooden floors. Makes the table more robust, too, because its integrity doesn't depend on the relatively weak end-grain joints.

The next step is gluing the first row of boards and the 2" edge, after flipping both upside down:

This way, the planed top surface of each board is resting against the workshop table - and then it's clamped down to be sure. As mentioned, this mostly worked out fine, but still needed planing down the road - a biscuit joiner would allow for more perfect alignment, but I don't have one... yet?

Some sheet plastic is taped underneath to keep the glue from sticking to the desk. And here's the second row:

I clearly need more clamps! I'm using Titebond 3, which has a pretty short clamping time - all in all, it's taking about 40 minutes per row:

At this point, I'm out of desk space (and mostly ouf of clamp reach), so I decide to glue the second half separately and then join them at the center as the final step of the assembly. Here's the first half of the table with clamps removed:

This is the underside, so I am not going to put any effort into finishing it, but I'm putting a coat of shellac to prevent warping due to humidity variations. Otherwise, the unfinished side would expand and shrink more than the finished top. Not a huge concern for indoor furniture in California, but hey - no harm.

And now, for the moment of truth... the first half of the table flipped over. Thankfully, nothing seems out of whack; some of the boards were 1/64" or so out of level, but that's not hard to plane down. There weren't any major gaps, although I spread a good amount of latex wood putty anywhere I could see a hairline gap or any other nicks. Better safe than sorry.

Gluing both halves together:

Both pieces lined up pretty much perfectly and needed just a gentle touch with a hand plane. That was a blessing because I only have two 50" clamps - not nearly enough to overcome any serious misalignment. As can be seen, I'm using several smaller clamps and some scrap lumber to prevent the thin joint from bowing up when clamped, too.

With the both halves glued, it's time for some finishing touches! I'm trimming off the two shorter edges of the table (which just have coarsely aligned board ends sticking out) to attach the remaining 8/4 strips ("breadboard") and complete the outer frame. I'm using a fancy Emerson Tool clamp-on guide for this, although a straight piece of wood would also do:

Next up, attaching the strips with a risque franken-clamp contraption while using a planer to clean up the top surface:

The breadboard stuff is something that I really should have tested before applying glue, it worked a lot better in my head than in reality - it was difficult to align it just right. Again, a biscuit joiner or some other mortise-and-tenon approach would have made things simpler.

As an aside, breadboards are a tricky topic, because when humidity changes, they can expand and contract at a different rate than the rest of the table. There is some mildly conflicting advice on how to handle this; some authors suggest free-floating breadboards, which have the drawback of possibly looking unsightly when they get out of alignment with the rest of the table. A smaller proportion of authors goes for a ton of dowels and hopes for the best (the risk here being that the table, if prevented from moving, may develop cracks).

In this case, given that the indoor climate in Northern California is extremely boring (basically 70° F and 60% RH all year), I'm gonna hope for the best. It might end up being a humbling lesson, though.

With the breadboards done, we're almost there - it's time for sanding. I'm using 60 grit with a detail sander; this takes a while, but I really like this tool - it's well-balanced, so it doesn't mess up my wrists the way that an hour with a cheap random orbit sander would. I follow with 100 grit, a quick coat of dewaxed shellac (as a sanding sealer), then 220 grit and 400 grit:

At this point, I move back into the garage. I wipe the table clean with a tack cloth, then apply the first coat of satin polyurethane:

About ten more layers to go. I'm using wipe-on gel poly, since it's a lot less demanding than brush-on formulations - basically impossible to mess up, no drips, no dust specks, no fish eyes, doesn't require sanding between coats, etc. The trade-off is having to do more layers, especially for high-wear surfaces.

Here's a close-up of the surface, five coats in. This is about the sheen I was aiming for, but I will keep adding several more coats for durability:

Well, the hard part of the project seems to be done!

4. The base

The next weekend, I'm back in the backyard with some janky 4x4 Douglas fir from Lowes. Lots of knots and dings, plus a couple mystery stains - but at least it's reasonably straight. I trim it with a circular saw making four pieces that are about 15" long, four pieces around 28", and one piece about 50" long:

This is a small saw, so it can't really tackle a 4x4 in a single pass - I have to make multiple aligned cuts while rotating the lumber. I had a larger miter saw some time ago, but ended up selling it because it was just collecting dust... I'm not smart.

The surface of the lumber is a bit too rough for my taste - but because it has a ton of knots, I am not brave enough to run it through a jointer. Instead, I decide to use a bandsaw to trim off about 1/16" on every side:

This is the wood as it came off the bandsaw. Better than expected, I'm not even gonna sand it. My wife requested something that's "rustic, but not too rustic" - and I think this fits the bill:

Bonus: a hefty pile of deluxe premium ultra-low-grade Douglas fir veneer:

The next task is to assemble the main part of the legs for the table. The desisn is pretty self-explanatory: two 28" pieces and two 15" pieces glued together. One 15" piece is lined up with the top of the legs, the other is about 7" off the ground:

And for the final touch, a 50" cross-bar connecting the legs:

Since the legs are pretty hefty, this is more for appearance than stability. I plan to attach the top with some steel angle brackets, so the whole thing should be pretty hard to wreck.

I was sort of unsure about the finish. I wanted the bottom to be dark to provide contrast for the maple top and to better match the chairs and the walls. But I was torn between opaque paint (dark gray or black) and staining the base to preserve the look of wood. The risk here is that fir doesn't take stains particularly gracefully, so this could end up being a blotchy mess... oh well, I can always paint later, right?

This is Minwax "red chestnut" penetrating stain. Actually turned out better than expected, although the contrast between early and late wood is pretty sharp. To tone down the rings, I spray the whole thing with a light coat of tinted polyurethane (Polyshades "Bombay mahogany"). It ends up looking pretty sharp:

Several more coats of satin clear poly, some felt pads on the bottom, and I think I'm done with the base.

5. Putting it all together

Not much to discuss here; basically, just a bunch of angle brackets to keep things together while allowing for disassembly:

The brackets don't allow for a whole lot of movement, but given the simple design of the base with a fair amount of give, I don't expect this to become a big deal. In any case, not hard to monitor and revisit later on.

And here's the moment of truth - the whole shebang:

Turned out pretty good, I think! The total cost of materials is about $250. Took about 15 hours of work, give or take.

6. Closing words

Well, that's all folks! If you have any questions, feedback, or critiques, you can reach me at lcamtuf@coredump.cx.

Your lucky number is: 20861566