dovetailing a carcass

[Dan Hahr]

It seems to me that the glue (like the hide glue used years ago) might eventually fail. If the tails were on the top or bottom of the carcass, lifting it by the top to move it might be disasterous. I like to think my furniture will be around long enough for the glue to fail one day...
Dan

[Frank Drew]

Dan,

Glue certainly weakens over time (a long time, to be sure), but I consider that a very good argument for placing the tails on the top and bottom if you want your carcase to hold together and the doors and drawers to fit properly even when it does.

[Adam Cherubini]

You guys are right, I'm wrong. I'm thinking of very narrow subset of 18th c chests of drawers. When I think of rules, how one makes drawers for example, there DO seem to be similarities regionally and even nationally for the furniture I look at. I have no idea about what Chris Becksvoort does.

You won't find great differences in anglo-american 18th c furniture in the Met, PMoA, Winterthur, or the Wallace Dewitt. The general practice is the tails are on the sides. Drawer dividers are dt'ed into the sides in front. Backers are nailed to rabbets in the sides in the back.

I think Jeff Greene's book, "American Furniture of the 18th c" is a good look at basic joinery. I recommend this book to all woodworkers for it's appendices alone.

Adam

[Frank Drew]

As long as we're citing authors and authorities, this from Bruce Hoadley's Understanding Wood:

"Nothing is more symbolic of the woodwork tradition than the dovetail joint.... commonly used in carcase construction. The joint consists of interlocking tails and pins, giving it strength in tension along the tail member but not along the pin. It should therefore be oriented to resist tension against the tails [the strength comes from the wedging action of the tails against the pin faces]. In a drawer, for example, the tails should be in the drawer sides, the pins in the drawer front. In case construction, the pins should be in the sides and the tails in the top [and bottom] to prevent the sides from spreading."

This surely won't be the last word on the subject, just another county heard from. But he does explain my thinking as well.

[Sean Hughto]

What stresses are you imagining that make the sides spreading any more likely than the top and bottom coming off?

It's just like a six board chest in that there is no obvious right direction to use the mechical strength of the joint (assuming away the super strong long grain connections at the key glue surfaces). We don't much worry about long grain edge table top glue ups having failures 100 years from now, why the great concern with carcase dts?

Finally, I think Hoadley's argument is a bit too simple in that it imagines teh carcase in a vacuum. Internal dividers (with sliding dt's as Adam mentions); tops, legs, stretcher's; aprons, etc. all come into play to restrain "side spreading" in various designs. For example, look at my piece above. The sides are trapped by the legs, top and stretchers. They ain't movin.

[Randy Klein]

Just curious. How many people have been sitting in their chair and all of sudden the bottom of a case falls off and the top is subsequently flung off? Honestly, has that ever happened? Most damage done to furniture is going to happen when those moving company goons get their hands on it. And at that point, the DT orientation won't matter. With that in mind, do whatever looks best to you.

[Bob Smalser]

What stresses are you imagining that make the sides spreading any more likely than the top and bottom coming off?

A carcass experiences significant racking stress every time it is pulled or pushed from its place to vacuum or dust beneath, which can be several times a month in some households.

The taller and heavier the carcass and the more physically weak the mover, the greater the stress, although mortised legs and shelves can help distribute the stress from the corners. Given enough time, most carcass corner joints eventually break their glue bonds because of this, and repairable glues are a thoughtful touch for the next guy who works on the piece.

Accordingly, for freestanding furniture I'd be very reluctant to consider any orientation other than tails on tops and bottoms.

[Frank Drew]

What stresses are you imagining that make the sides spreading any more likely than the top and bottom coming off?

Sean,

A case might have a side 16, 18 inches wide, whatever; there's a tendency for wide boards to cup (once the glue holding them in place weakens). That's why a mechanical joint, independent of glue, is important.

But other than a reversal of the laws of gravity, what, in normal use (and excluding abuse), is going to make the top or bottom of a case to come off? And even if a dovetailed top board cups a bit, it should still hold the sides in place as long as a little bit of the dovetails still hold; however, if a side board cups or otherwise moves even a little bit your drawers or doors won't fit as they're supposed to.

I don't think the placement of the dovetails is interchangeable, according to whim; I believe that there should be a good cabinetmaking reason behind our joinery decisions.

But people are allowed to make furniture however they want; I don't pretend to be the furniture police.

[Sean Hughto]

I don't think the placement of the dovetails is interchangeable, according to whim; I believe that there should be a good cabinetmaking reason behind our joinery decisions.

If I said something to make you believe I think that dovetails are always interchangeable according to whim, I apologize, as I do not believe this. I do think that the orientation of the dovetails can vary in a carcase depending upon a lot of factors about the individual piece in question (e.g., to be hung on a wall or sit on legs etc. etc.). Hence generalizations or rules are not particularly useful. One should think through the cabinetmaking reasoning, as you put it, for every piece.

I also think that there are times when the answer is that the answer to the tail/pin orientation question is indeed, six of one and half a dozeon of the other - there is no clear reason to favor one orientation over the other, and a dovetail is still the best joint that can be imagined for a rigid durable carcase. I suppose in such situations, one might wedge the pins to create mechanical locks in both directions, but wedges too, of course, rely upon glue ultimately, albeit, again the best kind of glue joint: long grain and plenty of relative surface area.

As far as what is going to make the top and bottom come off - well Bob makes a great case for the lifting and moving stresses on the top. Obviously if the case is loaded with heavy items, it is likely to have some significant stresses on the bottom too.

Also, I don't think properly prepared sides (just like a properly prepared table top) should cup very much, if at all. And internal structures like dividers and runners, etc. should help restrain such cupping just as aprons and such restrain table tops somewhat.

[Frank Drew]

Also, I don't think properly prepared sides (just like a properly prepared table top) should cup very much, if at all.

In general I agree with this and it's one reason I came to appreciate wood that's been air dried for a long time -- it seems more "relaxed" and less prone to funny business. But, still, wood can be a bear sometimes and it's best to take reasonable precautions.

[Bob Smalser]

To clarify for those who have asked, this is what "racking" or "shear" stress looks like. It occurs every time the carcass is moved, and remains the principle reason that tails are usually cut in the tops and bottoms of freestanding pieces.

http://en.wikipedia.org/wiki/Shear_stress



If your sides are cupping to any significant degree, you have more basic problems (wood selection and grain orientation during layup) than the orientation of your dovetails.