Success: Failure Accomplished

May 31, 2016 — 8 Comments

Back in April of this year, my friend Chris Schwarz wrote a post about a staked table he built. I’ll confess – my first reaction to the photo was “WTF?” – but there was something about the table’s aesthetics that intrigued me and I couldn’t quite put my finger on it. It was as if the photo of the table begged me to stare at it longer, goading my brain to make sense of it. I should have responded with an anyeurism emoji.

I know from personal experience with some of the staked pieces in The Anarchist’s Design Book (and other similar forms) that photographs don’t express the (potential) spatial elegance of these forms very well, so I was willing to look past the picture. And as I did (I studied that photo three or four times the day I received it), the more I wanted to “riff” on Chris’ table. So I started sketching.

{"focusMode":1,"deviceTilt":0.03907515481114388,"whiteBalanceProgram":0,"macroEnabled":false,"qualityMode":3}It became clear that I had to build the thing I kind of-sort of had in my head in order to move on with life. But even after sketching, it was more of a conceptual puzzle that needed working out rather than a design. Whatever it ended up being, it apparently had the following criteria:

  1. Three legs. I thought four would resolve “too well” or too easily.
  2. Round. I wanted something to place next to my “big chair” which has a Scandinavian aesthetic that lends itself to curves. Also I thought that the roundness would help alleviate some of the M. C. Escher problems that occur when your brain has corners against which it can “register” what it’s perceiving.
  3. Funky geometry – square or round, everyone knows what an end table with equally ordered legs looks like. I wanted the legs to look like they were wrapping around a column that wasn’t there.

I would normally spend a few hours in Sketchup trying to work this out into something that would then come together in the shop. But I’ve spent the last year doing home improvement projects in plywood, particleboard and veneer and wanted to get to work immediately. So after taking some measurements of my chair, I decided to make good on my commitment to clear out my lumber rack and go straight to prototyping, knowing that the best outcome would be a failure to learn from (followed by a funeral pyre for the prototype).

{"focusMode":0,"deviceTilt":-0.07968433300522371,"whiteBalanceProgram":0,"macroEnabled":false,"qualityMode":3}As suggested in The Anarchist’s Design Book, I first made a quick model by turning a small disc and using coat hanger segments to play with angles. My goal was to get a very rough idea of the issues at play, not to faithfully represent the piece (because I had no real clue on where it was going yet). The model proved to me that the concept wasn’t completely bonkers, but also that it needed more careful consideration than banging a bunch of sticks into the bottom of a circle and calling it a day. It was clear that much of the overall design would be “derived” from a number of elements:

  • The radius between each leg and the center of the tabletop (I’ll call this the leg radius)
  • The angle between the legs and the bottom of the tabletop (I’ll call this the splay)
  • The angle of the legs from the center (I’ll call this the rotational angle)
  • The length of the legs, which would determine the proportional distance at which the legs would appear to intersect (I’ll call this the fleemkoopen stropfheimer)

Though I love incorporating curves into the things I build, I don’t work on round furniture very often, so while many of these design considerations are also present in rectangular furniture, it took my brain some time to reorient itself to working on radii off a circle’s center rather than more Cartesianally-oriented distances from edges.

I had 11” wide stock in the lumber rack so the table was going to be some multiple of that–I decided 22″. I decided to work first on a small square piece of 8/4 poplar. This would allow me several attempts at finding a good set of angles for the mortises and experiment with different ways of marking them out and drilling them. My plan was to drill the mortises and use dowels to evaluate the angles–basically a larger version of the disc-and-coat hanger model.

{"focusMode":0,"deviceTilt":-0.006697782780975103,"whiteBalanceProgram":0,"macroEnabled":false,"qualityMode":3}I failed over and over, each failure more exciting than the last. Some failures were cognitive failures (e.g. forgetting which layout line was the rotational angle) and some were construction failures (sloppy brace-and-bit handling). But after a few attempts on a couple of boards I landed at something that was close enough from a design perspective. I had also streamlined techniques for markup and drilling with a brace and auger bit.

In the end, on an 11“ square piece of poplar, I landed on:

  • A leg radius of 3–7/8”
  • A splay of 124° or 56° depending on which way you splay. (I think in conventional terms this would be 34°, measured as the acute angle between the leg and a line perpendicular to the bottom of the tabletop, but my brain won’t accept that. In this case, I want to capture the angle to which I set my bevel gauge and to also reflect that the legs point into the table, not out)
  • A rotational angle of 17°
  • A fleemkoopen stropfheimer of 18–19“ for a total table height of 21–22”

Knowing that it was going to be way easier to work on an 11″ square piece than on a large round piece, I drilled the holes with a brace and auger bit (leaving the tapered mortise for later), laminated the block to the bottom of a 22” wide, 8/4 poplar panel. I marked and cut out a rough circle, affixed a face plate and went to the lathe. While at the lathe I realized that the curve I was shooting for needed more wood than I had laminated onto the bottom of the tabletop, so I made some design compromises. Turning the top was otherwise pretty straightforward (and lots of fun!), though by the time it was all said and done I would end up taking the tabletop back to the lathe three times to refine the shape.

As for the legs, I wanted something round-ish. I’ve had an 8′ long piece of oak stair rail sitting on my back porch for over a year, and by using my Jedi powers to check off the “mount new stair rail” item in my household to do list (i.e. I convinced myself I didn’t want a new stair rail), I decided to use that for prototyping.\

{"focusMode":1,"deviceTilt":0.6824432015419006,"whiteBalanceProgram":0,"macroEnabled":false,"qualityMode":3}It was convenient in that it was mostly round and made from wood, but was a piece of crap in all other respects (namely that it was laminated in both thickness and in length). But it got used and is no longer sitting on my back porch. In any case, I started with a pretty chunky ovular design by using offset turning, but unhappy with that I put them back on the lathe and turned them to be more svelte. I then planed two sides into each leg and did some rough shaping with some spokeshaves to get them to communicate “not round, not flat”, which is what I was going for.

I dry-fit the legs into the table and decided that the crappy grain from the stair rail was drawing too much attention to the legs, making it hard to evaluate the overall form. So I charred them and that helped immensely, as did reducing the contrast of the top by giving it a quick coat of stain.

{"focusMode":1,"deviceTilt":0.2292278250071664,"whiteBalanceProgram":0,"macroEnabled":false,"qualityMode":3}After staining the top and leveling the legs, I called this first prototype done. I almost tossed it right after inserting the legs, but I’m glad I took it through a rough “coloring” process because that changed my impression of the piece significantly.




It’s very much built as a prototype–I concentrated on stuff that I wanted to resolve in the design rather than fit-and-finish or engineering. I will probably give this to a friend or burn it, but I’ll keep it around a bit and ponder my next moves.

I’ll leave my more specific opinions on the piece itself this out of this post except to say that as a prototype, it was successful. It came together quickly (maybe 5–6 hours actual build time across a few days) and allowed me to experiment and refine both form and process. Most importantly: through the process of repeatedly failing, it’s very clear to me what I want to change as I go forward.

And that is what the second prototype is for …

8 responses to Success: Failure Accomplished

  1. Hey Narayan, when you say “charred” did you actually do the Seth Gould light-it-on-fire thing? How did that go?

    Great post.

  2. Am I right in that the legs (starting as oak handrail) had laminations that remained in the finished piece? if so was there any glue impact with the heat from charring?

    • Maybe–I didn’t check to be honest, though you can see a finger joint in one of the pics and maybe that’s the glue (maybe not). Nothing came apart.

      But for prototyping it was fine.

      • This is intriguing. I’m really interested in the charred finish (esp on oak) but the way it fails at glue lines is a serious design challenge (e.g. wide panels) Thanks for this data point. It has renewed my interest in experimenting. There may be some potential after going a bit down the rabbit hole of industrial finger jointed mill-work adhesives, though I’m not sure a 55-gallon drum-o’-glue will be convenient for my shop.

Trackbacks and Pingbacks:

  1. Success: Failure Accomplished – Woodworking Unplugged - June 1, 2016

    […] (crossposted from […]

Leave a Reply