Loose sliding dovetails in winter

[Winston Chang]

I made a table where the legs are attached to a batten, and each batten is attached to the top using a sliding dovetail.

420997

The battens are 22" long and tapered, so that one end is 3" wide and the other end is 1/16" smaller. When I assembled it, it was during a humid part of the summer and the fit was very tight. I had to hit it with a mallet to get to move the last inch or so. The table was rock solid, even though I didn't use glue on the battens.

Now it's winter and the air is dry, and the battens are loose -- I can slide them out by hand with no resistance, and the table rocks back and forth when it's bumped.

I asked in these forums a while back about seasonal movement and was advised to not worry about it, but it looks like I should have worried about it a bit more. :(

Any advice about how to deal with the loose battens?

And in the future, if I build something like this again, how should I deal with seasonal movement? I obviously don't want it to be loose in the winter, but I also don't want it to be so tight in the summer that it would crack. I wonder if it would have helped if, in the battens, if I had oriented the wood so that the expansion across the dovetail was radial with respect to the growth rings.

[Andrew Hughes]

My best guess is your sliding Dt that you cut in the top is too close to the edge. On your next one move it in and do change the grain on the male part quarter or rift.
What wood am I seeing looks like Alder or maple.

[Greg Funk]

I think your dovetail is probably just too wide. If you want it that wide it's better to orient the grain vertically in your diagram as you surmised. An alternative would be to make the dovetail narrower. Something like the diagram (not to scale) below.

421007

[mreza Salav]

I have used sliding dovetails but in much smaller cross-grain applications and with glue. These joints are not forgiving at all and slightest change in the dimensions (as you have noticed) will cause in a loose joint. I am not sure why you chose to use that type of connection of batten to top. For table tops I always use mechanical fasteners (screws).

[Andrew Seemann]

Well, your sliding dovetail did exactly what it was supposed to do, allow wood movement with the change of the seasons. And what you experienced is also why it usually is not used in that situation, Typically for a table top, you would use table top hardware like figure 8s or Z clips, buttons or cleats in a groove, screws in elongated holes, or any of the other standard ways to attach a table top to the frame/legs. Tables have been around a long time; most of the details have been worked out. If you don't see tables constructed in a particular way, there is a good chance that it may not have stood the test of time (and the seasons).

You don't have your location listed, but depending on where you live, there may have been no way to make that joint work. It would not fare well where I live, with the extremes between the humid summer and the bone dry winter. As far as advice gleaned on this site, there are places in North America where it probably would have worked just fine; you just aren't in one of those places it seems.

You could permanently fix the top to the cross piece in the center with some screws or dowels, That would keep the top from sliding off, but still allow the top to move with the seasons. I usually do something like that when I use screws to attach a table top. The center gets a normal hole, but the ones not in the center are elongated. It is a pretty table and worth trying to save.

Just chalk it up to learning. I have a pretty significant body of work from about 25 years ago that proved it took me a long time to get the hang of wood movement:)

[Dan Friedrichs]

Winston, that table looks great - I love both the design and execution of it. Very nice work!

Maybe I'm missing something, but why can't you just add a plugged screw (or such) at the center of the batten? The top will still be free to expand and contract width-wise.

[Brian Holcombe]

A sliding dovetail works best in conjunction with other joinery. As you found they’re not good as the only joints holding a table up.

Use it for what it does best (lets the top move) but make the supports in a way I which they are self sufficient.

Now about sliding dovetails;

Width helps structurally bit as you found the joint loosens with seasonal change. This can be minimized if the joint compresses slightly as it is assembled. If they slip together without much compression they will be loose pretty much forever.

[Brian Holcombe]

Here is a table I made over the summer. The sliding dovetails went in tight, they were about 40" wide and that was some work!

The tenons go into the table top, the sides of the tenons are made with clearance for seasonal movement but the large faces which bear against the end grain in the mortise are made tight (compression fit).

This table is rock solid, myself and my client pushing against the side could not get it to budge. The majority of that strength is in the tenons which enter the top, the table was merely strong prior. The dovetails do help but they're not relied upon to prevent racking movements.

https://brianholcombewoodworker.com/wp-content/uploads/2019/12/img_3853.jpg

[David Utterback]

Can you glue an inch or two the snug end and let the remainder of the joint move?

[Brian Holcombe]

You can, but then the joint can't be easily taken apart, better to just use a peg or screw in the center.

[Jim Becker]

The simple solution is to drive a pin or screw up from the bottom that simply keeps them from sliding...in the center as noted so that things stay even. The reason you're getting the issue is that the grain direction on the top of the leg is cross grain to the table top. The wood contracts and expands across the grain. Therefore, the male side of dovetail is getting narrower while the female side in the bottom of the table isn't changing size at all in width due to grain direction.

[Winston Chang]

I am not sure why you chose to use that type of connection of batten to top. For table tops I always use mechanical fasteners (screws).

The design was inspired by the staked desk from The Anarchist's Design Book by Christopher Schwarz. In his desk, though, the dovetails aren't tapered, and used he glue on one end of the batten.

[Winston Chang]

Winston, that table looks great - I love both the design and execution of it. Very nice work!

Maybe I'm missing something, but why can't you just add a plugged screw (or such) at the center of the batten? The top will still be free to expand and contract width-wise.


Thanks! I may end up doing what you suggest, though I'm a bit concerned that if I use a screw, the racking motion will cause it to chew up the wood and loosen up over time.

[Winston Chang]

Here is a table I made over the summer. The sliding dovetails went in tight, they were about 40" wide and that was some work!

The tenons go into the table top, the sides of the tenons are made with clearance for seasonal movement but the large faces which bear against the end grain in the mortise are made tight (compression fit).

This table is rock solid, myself and my client pushing against the side could not get it to budge. The majority of that strength is in the tenons which enter the top, the table was merely strong prior. The dovetails do help but they're not relied upon to prevent racking movements.


Fantastic work! Was it assembled by sliding the battens into the top, and the putting that assembly on the legs?

[Andrew Seemann]

Thanks! I may end up doing what you suggest, though I'm a bit concerned that if I use a screw, the racking motion will cause it to chew up the wood and loosen up over time.

You can put a few screws in at the center of the batten. Just keep them within a space within about 3 or 4 inches at the center. There won't be enough expansion/contraction to matter over that distance. You are mostly trying to center the top on the batten and keep it from sliding out. Also/alternatively since the top is removable, you could take the top off, route a groove most of the length of the top and add table cleats on the inside side. You might even be able to put figure 8s in as an option.

Hey, just noticed you are from MN as well. Yep, that design will be problematic here. Just wait until January when the temp is -20F and there is no moisture in the house:) As you build furniture, you'll need to keep in the back of your mind that summer to winter expansion here can be double what is in some other parts of the country, and certain things just don't work like they do other places. Most folks don't understand how decks can make gunshot noises in winter.

[Winston Chang]

While I was working on the table, I found this article in FWW by Andrew Hunter about attaching tabletops this way.
https://www.finewoodworking.com/2010/06/03/better-way-to-attach-tabletops (https://www.finewoodworking.com/2010/06/03/better-way-to-attach-tabletops)

In it, he writes:

This taper eliminates the precision needed to fit a straightcleat and the need for glue. With no glue, you can remove thebase from the top if needed, and if the fit of the cleat loosens due to wood movement, you can tap it back home or remove and shim it if necessary. I’ve never had to do either because I start off with very dry wood and use quartersawn lumber for the cleat.


Now that I look at it again, I realize that the quartersawn part is important -- the wood that I used is cherry, which apparently moves twice as much in tangential direction compared to the radial direction.

For now, I may shim it with wood shavings or paper and see how it holds up after another year of changing humidity.

[Andrew Seemann]

I dug out that copy and took a look at the article. I don't know that I would call it a "better" way to attach a table top, maybe a different way, and subject to some limitations. FWW did some Cosmo-style headlines back then. That was about the point I canceled my subscription.

The climate thing plays into this again; he works out of Gardiner NY, and while not coastal, definitely milder than MN. Quartered wood would have helped, although it may or may not have helped enough. Trying paper shims and seeing what it does over the next year isn't a bad idea.

[Brian Holcombe]

Fantastic work! Was it assembled by sliding the battens into the top, and the putting that assembly on the legs?

Thank you, yes that is exactly how it functions.

I don't taper my battens anymore, I just size them to compress slightly, wax and fit. I agree with using quarter sawn but my issue with tapered battens is that they push the wood movement to one side and so half the batten is always loose (if the entire batten isn't already loose).


I dug out that copy and took a look at the article. I don't know that I would call it a "better" way to attach a table top, maybe a different way, and subject to some limitations. FWW did some Cosmo-style headlines back then. That was about the point I canceled my subscription.

The climate thing plays into this again; he works out of Gardiner NY, and while not coastal, definitely milder than MN. Quartered wood would have helped, although it may or may not have helped enough. Trying paper shims and seeing what it does over the next year isn't a bad idea.

I can't speak with Andrew specifically but here in NJ we deal with plenty of that issue. Only saving grace is that old houses sometimes leak enough and used radiators. Newer houses with forced air heat are brutal on furniture in these climates. He's not far from me relatively speaking and I'm certain that his environment also deals with these same issues.

----

It's my opinion/experience that a sliding dovetail batten is not enough to hold up a table and keep it from racking. Heck, if you lean really heavily on one side of that table it may lever the leg enough to simply pop the wood right off the bottom between the top of the dovetail and the end of the board. It's not that strong.

The goal of dovetail battens, in my opinion, should be to help keep the table flat and allow it to move, that is all, it should be minimally structural.

Single dovetails, which is what a sliding dovetail batten is, are lousy against racking forces, they're basically a short splitting wedge attached to a huge handle.

[Derek Cohen]

Very interesting comments, Brian. This topic is of particular relevance for me as my current build plans to use tapered sliding dovetails to anchor the base for angled staked legs. I would not be overly concerned about this if the piece lived in my home, since Perth does not have a humid climate. However, it is going to live in Sydney, where no doubt there will be wide temperature swings and high humidity in summer.

What do you and others suggest: non-tapered sliding dovetail, ensure it is quarter sawn, and additionally secured with screws (with elongated slots for movement)?

Oh, the added base for the staked legs is necessary since they cannot go directly into just a case. The underside of the case needs to be built up in thickness. The sliding dovetail kills two birds with one stone - adds to the thickness and provides stiffness. It has to move as it is across the grain of the case, otherwise one could simply glue and screw a brace.

Regards from Perth

Derek

[Brian Holcombe]

Yes, that’ll help and I’d be tempted to add a stretcher, something that looked proper with staked legs.

[Tom Bain]

Brian — Do you happen to have a photo with the top on the table? Would be great to see the finished piece.


Here is a table I made over the summer. The sliding dovetails went in tight, they were about 40" wide and that was some work!

The tenons go into the table top, the sides of the tenons are made with clearance for seasonal movement but the large faces which bear against the end grain in the mortise are made tight (compression fit).

This table is rock solid, myself and my client pushing against the side could not get it to budge. The majority of that strength is in the tenons which enter the top, the table was merely strong prior. The dovetails do help but they're not relied upon to prevent racking movements.

https://brianholcombewoodworker.com/wp-content/uploads/2019/12/img_3853.jpg

[Brian Holcombe]

Thanks, Tom. Not yet, next time I head out there to that client’s location I will bring photo equipment.

[Lee Schierer]

I would guess that you didn't know the exact moisture content of the maple when you made the table. If your moisture was not at equilibrium with your environment then your problem could be due to normal shrinkage. Since the table is new and this is the first winter, I would wait and see what happens in the coming summer. If it is still loose, then the simple solution is to add a shim to tighten up the dovetail. If it tightens back up in the summer then it is strictly seasonal wood movement and you may have to look at a different solution. It might be worth while to determine how much further you would need to slide it together for it to be tight and seasonally adjust the table or add a smaller shim so you can keep the legs "centered" more closely with the seasons