>
> alan ferrency[SMTP:alan@l...] said:
>
> (a bunch of stuff about round mortise and tenon joints being inherently
> unstable snipped)
>
> >The stretchers
> >in a Windsor chair are compression members, not tension members (in
> >his book he makes the stretchers 1/4" longer than they should be, to
> >make sure they're pushing the legs out).  Since they work to push the
> >legs apart, it doesn't help much to construct them to resist a
> >pulling-out force.
>
> Okay, here's where I'm really confused.  I've heard this theory before
> (probably from Mike Dunbar's newsletter), but I don't yet understand it.
> Could someone explain to me how the stretchers of a Windsor chair
> are under compression?   I don't deny that it's true, I just don't understand
.
> My more-or-less intuitive analysis indicates just the opposite.  Perhaps
> I should have paid more attention to vectors in high school Physics class.
>

Ok here is a pitiful attemp at ascii art

                   ___________________
                  |     /   /
                  |    /   / thick chair seat
                  |   /   /
                  ---/   /------------
                    /   /
                   tapered leg

The picture is incorrect in that the leg is conical and going through
an equally conical hole.  When you assemble the legs and put them
into the seat, the lengthwise and crosswise stretchers are oversized
by a quarter inch (you ream the conical holes in the seat and fit the
legs, then pick off the exact dimensions and add 1/4 inch).  The holes
in the seat all point away from the center, so you get all the (glue
smeared) legs started and pound them home. The thickness of the seat
forces the stretchers to compress, the further in you pound the legs
the more you are pushing the tops of the legs apart, which pulls
the bottoms of the legs together.  The force of your weight on the
chair is equivalent to pounding the legs in further.  The compression
is the bottoms of the legs pulling together compressing the stretchers
(which are more than half way down the length of the leg so pulling
in).

In my old "windsor style" dining room chairs the seats are thin and
the legs don't go all the way through.  The weight of a person sitting
on the chair pushes all the legs out and the stretchers are the only
thing holding it together (until the glue breaks _again_).  The trick
in the original style that Mike got early although he hadn't figured
out the physics yet when he wrote the book, is the 1/4 oversize on the
stretchers, the _thick_ seat, and the tapered leg holes going _through_
the seat.

I think John Alexander's chairs are more reliable than commercial,
he is taking advantage of the different shape available between wet
and dry (I think he is also the author of "the incredible duckbill
joint" in FWW on hand tools that is this joint with a spoon bit
and very careful grain orientation).  The woven seats also add a lot
of pull at the top of the legs to stay together.  I would be
uncomfortable with the longevity of an Alexander chair with a
solid seat though...

The simplist way to convince yourself is to make a stool with a
Dunbar type undercarriage and a round top.  You could drawknife
the turned parts, the only requirement is a thick top where the
legs go through, conical leg tops and seat holes (conceptually
like how Morse tapers work) and egg shaped tenon ends in the
stretchers.  Follow the book to pick off the final stretcher
dimensions.  Once you have done it it makes perfect sense.

Esther eoh@k...