Thursday, February 28, 2013

Start at the bottom

My son's birthday is coming up soon.  He is in law school and I decided it would be nice to give him a Cross pen and pencil set I purchased as a young professional some forty years ago.  That led naturally enough to the idea of making him a box to hold desk items:



It is made from scrap pieces of walnut and cherry I had.  I really like the contrast between the two, particularly the way it accentuates the dovetails.  My dovetails continue to improve and I was quite pleased with the way these turned out.  I also like the way the sliding lid looks and works, an idea I got from Paul Sellers; to me, it makes the box look sleek and clean.  In addition, I think there are a number of applications where being able to remove the lid as opposed to having an open, hinged lid is preferable.  Since I have a small plow plane, it was a breeze to make.

There is something quite unusual to me about the lid, which is bookmatched.  You will notice that the back half is darker.  The odd thing is that if you look at the box from the other side, it's still true--the back half looks darker.  I can guess at the explanation but I have never been aware of this before.  Wood is a fascinating material.

To me, the most interesting aspect of making this small box was what to do about the bottom when using hand tools exclusively.  Paul Sellers just glues on the bottoms of his boxes, which I have done in the past.  I guess I have been indoctrinated, but it's just not something I am comfortable with.  I don't want the bottom to split in the next forty years, and it seems like seasonal expansion and contraction could take quite a toll.  I also don't think that a bottom sticking out from the sides would look all that good with a sliding top, but that's purely a matter of taste.

I made an effort to estimate the maximum amount of movement that might occur.  Making the worst case assumption of a maximum change in moisture content of 6%, I got a change of 1/16" over a 4" bottom using this shrinkage calculator.  That is more than I expected and seems like it might be enough to split the bottom.  My assumptions may not be realistic but this calculation was enough to convince me that I don't want to glue solid bottoms onto boxes.

It's funny how when you actually begin making an item, you run into issues that woodworkers have been addressing for years but which have never occurred to you before.  The alternative I chose is one that is generally recommended against for hand tool work because it is, in the vernacular, a PITA, but I didn't find it all that onerous.  I plowed grooves in the sides and made stopped grooves in the ends into which I fitted a bottom panel with 1/16" room for movement.



It's the stopped grooves that take time with hand tools.  I made mine by chopping a small "mortise" in each end and then using a router plane to plow out a groove between them.  I think the reason that it went relatively fast for me is that I have the optional fence for my Lee Valley router plane.  I have been surprised on several occasions at how well it works.  A not very skilled user like me can plow very straight grooves quite easily.  The only thing that slows the process down is that you have to keep adjusting the blade as you deepen the groove, but if you do both sides at once it really doesn't take all that long.  If you were making a number of boxes at once, you would only have to make these adjustments once, so that would cut the time per box even further.  I think the result looks nice and matches the look of the top.  You could also make this stopped groove with knife walls and a chisel and, given the short length, it wouldn't take all that long.

There's obviously many ways to skin this cat and here are a few.  I found a post by Bob Rozaieski in which he suggests plowing a groove in a small molding and then fitting the molding to the inside of the box.  Seems like a good option to me.  One small disadvantage, though, is that the molding would stick out from the sides of the box on the inside.

Mike Siemsen contributes several other approaches.  One, which he says has been demonstrated on episodes of The Woodwright Shop, is to miter a portion of the corners where the bottom will go and then plow grooves on all four sides.  I understand this in theory, but I think it might be hard for me to get my dovetails and the miter to close up simultaneously.  Still, it's a creative idea if you can pull it off.

Another idea Mike says was used historically is to create a cross grain rebate on two sides of the box, cut your dovetails normally and then plow the groove on the thicker portion of these sides.  Quite an elegant solution if you have the means and skill to cut accurate cross-grain rebates.

In part you should choose a method based on the tools you own.  If I had a nice rebate plane, I think I would try that method but, since I don't and I do have a router plane with a fence, I am happy with stopped grooves.



4 comments:

  1. An excellent box! I think I read some where that some plow a groove all the way and then use plugs in the gaps.

    I like your idea of the router plane with the fence. Must go try that .. if I ever finish my silly desk.

    ReplyDelete
  2. "Another idea Mike says was used historically is to create a cross grain rebate on two sides of the box, cut your dovetails normally and then plow the groove on the thicker portion of these sides. Quite an elegant solution if you have the means and skill to cut accurate cross-grain rebates."

    I was trying to visual this one and could quit visualize it. could you explain a bit more?

    ReplyDelete
    Replies
    1. As I understand what he is suggesting, you create a rebate on each end of the pin boards, which means that the pins would be only half the thickness of the boards. Then when you plow the groove it will be behind the pin and not show. Wish I knew how to use Sketchup.

      Delete
  3. A beautiful box Andy. I like the way the bookmatched grain in the top is pointing the way home and I think you made the right decision with the base - built to last.

    ReplyDelete