Attaching Legs with Dowels/Dominos and Wood Movement

[TomInNC]

My (newbie) understanding is that if you were attaching a table top to an apron, you don't want to simply screw a solid wood top to the apron because that will constrain the movement of the top, potentially causing it to crack. I've seen a few examples of tables and benches without aprons where one of the legs is doweled or Dominoed into the table top (see 2 examples below). Doesn't using a dowel or domino to attach the leg create the same issue with constrained movements that a screw would cause? 

 

https://www.popularwoodworking.com/article/simple-live-edge-slab-table/ 

 

 

 

 

 

[BillyJack]

Attaching a leg to an apron isn't a problem. You could Domino , M&T, pocket hole, etc. Fastening the top is a different animal. You want to allow for movement..

 

We made two table using Dominos on a waterfall top. Both failed. Doing it over I would use a spline...

 

I rarely watch these videos. As small as the table is, I doubt you could cause much damage to it..

 

 

[G Ragatz]

Two different situations, I think, in the examples you offered.

• In the video, all of the grain is running the same direction in the top, the "waterfall" part and the leg - so expansion/contraction will be in the same direction in all three pieces, so I wouldn't expect any problem.
• In the popular woodworking article, the dowels are inserted into the top in a line, along the grain, so expansion/contraction of the top can occur on either side of the dowels, so there should be no problem with the top. What I'm not sure about is whether there would be enough expansion/contraction in that leg to cause it to crack, if those dowels are all glued in tight. Could probably address that by gluing one or two of the dowels near the center of the leg tight, and enlarging the holes (either in the top or the leg) for the other dowels to allow for some movement.

[BillyJack]

The piece is too small as I mentioned.  Large piece would be different. 

 

It's not the grain that's the problem or how it's running, if the it was a larger piece and you rocked it, the joint would fail...

[wtnhighlander]

A spindle leg that is doweled into the table top is free to move WITH the top as it expands and contracts, so long as it is not constrained by another component that does NOT move in the same direction. Aprons and stretchers are typically oriented such that they cause such a constraint, whereas 'staked' legs have none. 

The case of 'waterfall' style support members requires that the grain be oriented in the same direction as the top they are solidly joined with, so that their movement is equal. Be aware that stylistic designs, like the waterfall end with single opposing leg, while visually interesting, offer little to no resistance to cupping of the table top, or of the waterfall leg. From a purely practical standpoint, that design is less desireable. 

'Staked' designs, where the spindke / post legs are mortised directly through the table top or chair seat, offer simplicity of construction, but require heavier components for strength, and still provide no resistance to cupping. That is why the style is rarely used in large tables.

[TomInNC]

On 10/4/2022 at 8:00 PM, wtnhighlander said:

A spindle leg that is doweled into the table top is free to move WITH the top as it expands and contracts, so long as it is not constrained by another component that does NOT move in the same direction. Aprons and stretchers are typically oriented such that they cause such a constraint, whereas 'staked' legs have none. 

The case of 'waterfall' style support members requires that the grain be oriented in the same direction as the top they are solidly joined with, so that their movement is equal. Be aware that stylistic designs, like the waterfall end with single opposing leg, while visually interesting, offer little to no resistance to cupping of the table top, or of the waterfall leg. From a purely practical standpoint, that design is less desireable. 

'Staked' designs, where the spindke / post legs are mortised directly through the table top or chair seat, offer simplicity of construction, but require heavier components for strength, and still provide no resistance to cupping. That is why the style is rarely used in large tables.

Thanks. So in your example above, would this mean that if a leg was doweled into the top AND the legs were all constrained from moving with stretchers, we would expect the table top to eventually crack because it cannot move?

I think a related example is something like the table linked below. If I am following this correctly (big if), then a key component of the design is that the grain for the short stretchers runs vertically so that the movement of the through tenons follows the top. If the grain on those stretchers was not oriented vertically, would this prohibit the tenons from moving and cause the top to crack?

https://www.popularwoodworking.com/techniques/arts-crafts-through-tenons/

 

[TomInNC]

On 10/4/2022 at 4:03 PM, BillyJack said:

Attaching a leg to an apron isn't a problem. You could Domino , M&T, pocket hole, etc. Fastening the top is a different animal. You want to allow for movement..

 

We made two table using Dominos on a waterfall top. Both failed. Doing it over I would use a spline...

 

I rarely watch these videos. As small as the table is, I doubt you could cause much damage to it..

 

 

So what caused the failure when you used the domino for a waterfall joint? As I mentioned in another thread, I'm in the middle of a waterfall table build now. Because of the epoxy in the waterfall, however, a spline isn't an option as it would likely be visible through the epoxy.

[wtnhighlander]

16 minutes ago, TomInNC said:

 

I think a related example is something like the table linked below. If I am following this correctly (big if), then a key component of the design is that the grain for the short stretchers runs vertically so that the movement of the through tenons follows the top. If the grain on those stretchers was not oriented vertically, would this prohibit the tenons from moving and cause the top to crack?

https://www.popularwoodworking.com/techniques/arts-crafts-through-tenons/

 

Yes, you have the concept exactly.

To visualise wood movement in relation to its grain structure, imagine each 'growth ring' of the tree as a bundle of tubes, arranged in concentric circles. When the tubes are wet, they swell noticably in diameter, but hardly at all in length. As they dry, the diameter of each tube shrinks. As the tubes forming each 'ring' dry out, the ring as a whole shrinks as well, but because there are other rings inside, there is a greater proportional  reduction in circumferance of the ring, than in its diameter, and we see a split form in the log.

A board is sliced longitudinally from the log, with some orientation to the rings. If the wider dimension (face) of the board is oriented along a radius of the log, the board's width will shrink a bit less, and more uniformly through the thickness, than if it were cut oriented from a geometric cord of the circular rings. The radial cut is called 'quarter-sawn', and is considered the most stable. The 'cord' cut is called flat-sawn, and is most likely to cup, but produces the often desirable cathedral grain appearance. It is also the cheapest to produce. 'Rift sawn' means the board is oriented half way between these extremes, producing relatively straight grain lines on all for sides of the board. This is frequently chosen for table or chair legs, providing the best compromise between uniform appearance and stability.

[TomInNC]

On 10/4/2022 at 10:40 PM, wtnhighlander said:

Yes, you have the concept exactly.

To visualise wood movement in relation to its grain structure, imagine each 'growth ring' of the tree as a bundle of tubes, arranged in concentric circles. When the tubes are wet, they swell noticably in diameter, but hardly at all in length. As they dry, the diameter of each tube shrinks. As the tubes forming each 'ring' dry out, the ring as a whole shrinks as well, but because there are other rings inside, there is a greater proportional  reduction in circumferance of the ring, than in its diameter, and we see a split form in the log.

A board is sliced longitudinally from the log, with some orientation to the rings. If the wider dimension (face) of the board is oriented along a radius of the log, the board's width will shrink a bit less, and more uniformly through the thickness, than if it were cut oriented from a geometric cord of the circular rings. The radial cut is called 'quarter-sawn', and is considered the most stable. The 'cord' cut is called flat-sawn, and is most likely to cup, but produces the often desirable cathedral grain appearance. It is also the cheapest to produce. 'Rift sawn' means the board is oriented half way between these extremes, producing relatively straight grain lines on all for sides of the board. This is frequently chosen for table or chair legs, providing the best compromise between uniform appearance and stability.

Thanks. And so in the case of the through-tenon example in the PW link, the cross grain nature of the long aprons isn't an issue because this this apron isn't attached to the table top? Is that the right way to think about this?

Are you aware of any books that focus on wood movement issues for joinery? I gone through quite a few furniture building books at this point, and most seem to just say : (1) wood swells across its width; (2) cross-grain causes problems; (3) ignore movement at your own peril. I think I would understand how to approach these types of issues much better if I could find a book or collection of articles that walk through movement issues in a case-study setting like the PW article. Either that, or I should just build a bunch of furniture that breaks. Haha.

 

[BillyJack]

Like I said in the beginning, the small table won't be much of a problem. The tables I made were 12' long with36" or 42" legs. Can't remember as ot was maybe 8 years ago. It had a center partition for support for each leg and was intended for bar stools on both sides at a restaurant.  I believe it was 36" wide. Looked good and under carts it moved fine. Made it through painting and fit out. On the truck...... They had to ifted and put on the box truck, but as soon as one touch the floor and got dragged it strapped a joint. 

Sent deposit back and that was the end. 

I may have a picture somewhere.

[TomInNC]

On 10/5/2022 at 10:46 AM, BillyJack said:

Like I said in the beginning, the small table won't be much of a problem. The tables I made were 12' long with36" or 42" legs. Can't remember as ot was maybe 8 years ago. It had a center partition for support for each leg and was intended for bar stools on both sides at a restaurant.  I believe it was 36" wide. Looked good and under carts it moved fine. Made it through painting and fit out. On the truck...... They had to ifted and put on the box truck, but as soon as one touch the floor and got dragged it strapped a joint. 

Sent deposit back and that was the end. 

I may have a picture somewhere.

Ah, ok. I was asking because I am about to put a waterfall joint on a table. In my case the table is 15 inches wide by about 4 feet long, so much smaller. 

[BillyJack]

It's not something you want to drag. We stopped making them. 

 

[Dave H]

To decrease the chance of failure of the waterfall leg, I would ad a glue block to the under side of the waterfall leg not the full length but long enough to not be seen from the side of the table.

[TomInNC]

To keep with the theme here, does anyone know how Nakashima attached the slab legs to the table top? Given what people are paying for his work, I assume that the tables have held up well over time. There are a few examples at the link below where a table (e.g., the Conoid) has at least one very wide leg that appears to be joined directly to the table. I wasn't having any luck finding pictures of exactly how the legs were attached.

 

https://www.architecturaldigest.com/story/a-look-at-george-nakashimas-instinctual-woodworking

[wtnhighlander]

I can't speak to Nakashima's method, but others I have seen used a heavy cleat under the slab top. The cleat runs across the grain, but is recessed into a stopped dado with alternating sections cut like a sliding dovetail. The cleat has a tenon formed in a mating shape. To install it, the cleat is shifted to align the tenon's 'tails' with the straight sections of dado. Insert, shift, and glue a filler block into the open end of the dado that leaves room for the slab to expand, but not enough for the cleat to fall out. This is a type of 'keyhole' dovetail joint.

The cleat is typically bulky enough to add significant strength against racking the joint. The wide leg requires joining to the cleat to allow for expansion as well.

And it appears the Nakashima examples incorporate a length-wise member at or near floor level, which lies the wide leg and thin leg together, providing strength against the dragging problem mentioned by @BillyJack. Some tables place that structure at the top, which also works, but the member needs to be a bit wider to achieve the same strength.

I would suggest searching for literature on general architectural principals, and on the properties of wood. These should inform you how materials of known properties should be used in construction, and what aspects of various wood species make them more or less suitable for certain situations.

If you just want to know specifics about a particular case, asking here is a good choice! 

[BonPacific]

On 10/5/2022 at 2:10 PM, TomInNC said:

To keep with the theme here, does anyone know how Nakashima attached the slab legs to the table top? Given what people are paying for his work, I assume that the tables have held up well over time. There are a few examples at the link below where a table (e.g., the Conoid) has at least one very wide leg that appears to be joined directly to the table. I wasn't having any luck finding pictures of exactly how the legs were attached.

 

https://www.architecturaldigest.com/story/a-look-at-george-nakashimas-instinctual-woodworking

There's some shots of the underside of older tables on antique resellers.

https://www.prices4antiques.com/Furniture-Suite-Dining-Modernist-Nakashima-George-Table-Side-Chairs-5-Pieces-D9669968.html

https://modernism.com/items/39/occasional-tables/682-610-george-nakashima-origins-mid-century-modern-side-table

https://www.capsuleauctions.com/auction/lots/george_nakashima_minguren_ii_coffee_table_spw-2021_8738

[Chestnut]

15 hours ago, TomInNC said:

To keep with the theme here, does anyone know how Nakashima attached the slab legs to the table top? Given what people are paying for his work, I assume that the tables have held up well over time. There are a few examples at the link below where a table (e.g., the Conoid) has at least one very wide leg that appears to be joined directly to the table. I wasn't having any luck finding pictures of exactly how the legs were attached.

 

https://www.architecturaldigest.com/story/a-look-at-george-nakashimas-instinctual-woodworking

I was going to say figure 8 style fasteners then I saw Bon's post above. He used various methods but figure 8 style fasteners was one of them. If you pay attention most of his base designs don't have the large slab leg with out some structure attached to the leg to provide additional support. I feel like Nakashima used half laps a LOT as well.

[TomInNC]

Thanks, BonPacific. Appreciate it. These are nice examples.

So looking at how the legs are attached, in this case  it looks like there are figure 8 fasteners as noted by Chestnut. Are the fasteners in this case necessary because the grain from the leg is not running in the same direction as the grain on the table top? That is, the fasteners are necessary because you have a cross-grain situation?

Now on this table, the grain of the leg is running in the same direction as the grain on the table. Is this a case where it would be safe to glue the wide leg directly into the table top using dowels, dominos, etc.? To me it seems like the attachment of the leg is going to be really similar to just constructing a waterfall joint. 

From everything I have read and watched, my understanding is that you are safe to glue joints when the grain runs in the same direction (e.g., a case that is dovetailed together) because the parts of the case will expand and contract at approximately the same rate. Since expansion and contraction depends on the width of the piece, would you expect to run into issues with doweling a leg into the base if there was a significant different in the width of the top and the leg?

 

 

 

 

[wtnhighlander]

Case#1, yes. Use figure-8s because the grain is crossed.

Case#2, where the grain is aligned, a simple mortise&tenon, dominoes, or dowels will do. I suspect the short chunk in the center of the leg on that particular example gives it extra strength against twisting force.

[BillyJack]

This is one of those projects you have to challenge yourself. I hate live edge and avoid it when possible. ..

[TomInNC]

Ya, this is definitely something I'm doing just for the learning process. 

So for the guild build for the waterfall table project, Cremona drills oversized holes into the upper portion of the trapezoid legs and attaches bolts using the WoodWhisperer thread taps. He makes a comment about how screws would also work here so long as screws were in oversized holes. He doesn't go into any detail, however, about how to properly use screws here, perhaps because it should be obvious...

To use screws in lieu of threaded inserts, would the proper process be to drill a hole through the leg that is wider than the diameter of the shaft of the screw. Then, drill a standard pilot hole in the table. Lastly, attach the screw to the table using a washer? 

My intuition here is that if I do take this approach, I wouldn't want to ensure that the screw wasn't over tightened against the washer so that the screw could still move within the drilled hole.

Just thinking out loud here. Thanks again for the help. 

 

[BillyJack]

Shop projects are learning.

[wtnhighlander]

@TomInNC, it sounds like you have the screw situation all worked out. 

Now for the hard part - get to the shop and start building!   Sometimes analyzing your plan will delay your progress more than diving in!

[TomInNC]

On 10/7/2022 at 2:33 PM, wtnhighlander said:

@TomInNC, it sounds like you have the screw situation all worked out. 

Now for the hard part - get to the shop and start building!   Sometimes analyzing your plan will delay your progress more than diving in!

Haha. I agree that analysis paralysis can be an issue. At least this time I had a good excuse: I had to wait for the epoxy to finish curing. I did manage to make some scrap wood coasters. Not sure if that's considered productive. 

[TomInNC]

I was working on building the trapezoid leg for the waterfall guild table this weekend, and I came across a bunch of figure 8 fasteners I purchased a while back. Above I said that I was thinking of using screws with an oversized hole to attach the trapezoid leg to the table top. As I discussed in another thread, I don't have a drill press, so the potential fudge factor using screws is not trivial. Do you think a few figure 8 fasteners would be sufficient for this type of application?