I have a butternut shelf screwed between the legs of a table with pocket
screws. Since that is not particularly secure, I put a support under it,
also between the legs. To avoid problems with wood movement, I made the
grain run the same as the shelf; actually I just used cutoffs from trimming
the shelf. So, while it is better, it still isn't too great.
I had a brilliant idea; replace the butternut support with a walnut support
with grain running the other direction (across the shelf). Because of the
improved grain direction and the improved material, it will be many times
I figure I can do this because, according to my chart, walnut moves
grainwise almost as much as butternut does cross grain. And if anyone
notices the different wood, it will simply look like a design accent.
Is this idea sound?
You probably wonder how I ever get anything built, working like this.
Fortunately I have a lot of time.
BTW, you need a new chart. Might I suggest
http://www.fpl.fs.fed.us/documnts/fplgtr/fplgtr113/ch03.pdf They'll let you
know that wood in general moves 0.1% overall along the grain while your
butternut moves ~ .2% per EMC percentage point tangentially (6.4/30%). EMC
summer of 15%, EMC winter of 6% means more than an an eighth per foot.
and walnut moves 5.5% radially, for a difference of 0.9%.
That is closer than the radial differences of most woods, and everyone says
to ignore movement radially. (Of course, everyone could be wrong.)
We must be reading different charts, because table 3-5 says butternut is
very stable, while your figures show it is very unstable. What table are
your figures from?
Once again, having difficulty interpreting your question. I see a long
piece with grain running E-W prospectively screwed to a flat piece with the
grain running N-S. I interpret cross-grain as tangential, long grain as
along the piece.
The figures from 3-5 represent the average from ~30% MC - the fiber
saturation point - where wood begins to shrink, to zero MC - oven dry.
Thus you take the percentages, divide by 30 to get the % for a 1% change,
multiply for total change.
Look at 3-1, where A is influenced mostly by radial, B by tangential
shrinkage. The virtually shrinkless dimension is along both, not across.
If you put a long-grain rail across a tangential, or even radial grain
board, you'll want to allow for the movement as indicated.
In woodworking, success allows for the movement by fixing the wood where we
want the relationship to be unchanging, and "floating" the other attachment
You are correct, I was confused. Thank you for pointing it out to me. My
chart isn't wrong; I just didn't understand it. In fact, my chart says 1/8"
over a foot, which is what you got also.
But, the text with my chart ("Wood" by Fine Woodworking) says that the
movement is reduced by half when varnished; so that is 1/16" over the 12"
The problem is that both the cross grain butternut shelf (fortunately most
of the grain is radial) and the ripped walnut are pocket screwed to the
legs. The shelf can't expand its 1/16" without putting stress on the
The screws have 10 threads per inch. If I back each screw off a half turn,
that will be putting 1/10" of slack into the supports; so when the shelves
swell, they will not stress the support.
Does THIS make sense?
The movement is not reduced by anything, the wood is just slower on the
uptake and release over the seasons. Wood fiber swells to the water
available. Varnish just limits how much is available how fast. Lesson there
is to use the same finish in the same thickness in and out or suffer the
consequences. Even varnished, if you put 'em in a basement over the summer
with high humidity they'll swell like hell anyway. Not to mention what'll
happen here where the dewpoint is at minus 32. Four humans, two dogs and
dinner can't make 35% RH inside. Time to dry a load of clothes....
Is this because you're still having stability problems with the supports in
there, or because you are exploring a bit and just want to play with the
artistic side of this? The potential for wood movement does not, by itself
necessarily make for instability. In what way did you attach the supports
you mentioned, to the table legs?
table legs by pocket screws.
Between the legs, both the shelf and the support are crossgrain, and
butternut is very weak cross grain. A more rigid attachment would have
helped; had I planned better, I would have glued the support to the shelf,
and then glued the assembly to the legs with biscuits. But I assembled the
legs to the upper case body first, so gluing the shelf in afterwards wasn't
going to happen.
So, I have to try to clean up afterwards and make a strong rigid assembly.
I don't have any figures, but I bet a lengthwise walnut support is 5X as
strong as a crossgrain butternut support, as well as being substantially
I did some tests and found that two pocket screws in butternut will hole my
full weight, as long as the structure is rigid. As soon as it moves a
little, it breaks apart.
And yes it is a bit artistic. I generally only get artistic when forced
into it by practical matters.
If I understand your post correctly, your main concern revolves around the
'cross grain' situation betwen the shelf and its support(s)?
From your description, the way you attach the supports to the _legs_, if you
are indeed doing so, should be irrelevant for all practical purposes.
If there is no mechanical attachment between the shelf and the new supports,
IOW, no screws, nails, brads, pegs, etc.or glue, and the shelf is just
sitting on the support(s), then wood movement should not be a problem
between the two.
However, most any time you create a "cross grain" situation between two
pieces of wood that are _fastened_ together in some manner, you will need to
address the cross grain situation. There are a number of ways to do this.
But, you can't really get a definitive answer until you specify the method,
if any, by which you are attaching the new support(s) to the under side to
The shelf is pocket screwed to the legs, and the support is pocket screwed
to the legs.
The butternut support is also screwed to the shelf. I had planned on doing
the same with the replacement walnut support. As a former engineer, I know
that stacked beams are substantially more rigid when fastened together.
However, my belief is that grainwise walnut and crossgrain butternut are
close enough in movement that it is not an issue. If I am wrong, I can
sacrifice the screw between the two; the walnut support will still be much
stronger than the butternut.
Don't count on any cross grain situation not causing some trouble over time
unless you fasten the two correctly. It does not have to be involved. As has
been suggested, if you are going to use screws, just slot the screw holes in
the walnut supports in the direction of anticipated movement of the
butternut shelf and you should be fine. If you make the screwholes oversize,
consider using a washer under the head of the screw if it doesn't show.
You are absolutely correct, any wood span should be stronger with the grain
my website to show the person I am building it for.
You can see the area I am concerned about in the lower left. The shelf is
butternut, the grain goes the length of the table. The shelf support is
walnut, grain goes perpendicular to the shelf.
The grain in the side panel is the same direction as the shelf and the top.
(The top is on temporarily, just to show proportions. It still has to have
a finished edge put on it.)
float the top, the side panels and the shelf. put solid stringers
between the drawers, at the top and bottom of the side panels to match
the ones supporting the shelf. and use a better joint than pocket
Hey!!! I was asking questions so we could get to this point. Stand in line
dambit! Now you went and ruined the punchline.
More seriously, I'm not sure I would have floated the top and the shelf, but
I'm not saying I disagree with that. I've not had problems with securing a
top or a shelf on this type of piece, so I never encountered the need to
Thanks for posting that link toller. It's much easier to consider these
things with the visual. I would not worry about the cross grain condition
presented by the shelf supports on the ends of the table. Remember that the
movement numbers you see in the charts are for unconstrained wood and you
have wood that is constrained by the assembly. Look at the table in your
kitchen for example - it has this very same construction. Look at styles
and rails in a typical face frame construction - same thing. Cross grain
construction does not make an absolute problem condition.
There - having said all of that... this thread has gone on long enough that
I need to ask - were you originally posing your questions because you are
concerned for cross grain construction as a principle, or because you are
dealing with a problem? If memory serves, you did not state a real problem,
but more of a concern. But then again, I've often had to admit that my
memory was the second thing to go...
The movement numbers are for expansion and contraction of wood fiber with
changes in RH. If they compress against frame members and blow them apart -
what could happen if the lower shelf isn't allowed a relief at the ends, and
always happens with mitered solid stock, it's shaky table. If restrained at
either end and seeking relief from lower moisture, the board splits and
Since it's easily prevented, it seems almost ridiculous not to anticipate
and allow. As I said earlier, that's what woodworking is about - the wood.
only a tenth of an inch, unconstrained, but it could still be a problem if
it actually pulls the screw out a hair and ruins the threading in the wood.
I backed off the screw tension to put a tenth of an inch of slop into the
supports. That "ought" to protect it. I will tighten them up in August.
Of course, they ought to be tight in August; right.
I was a little concerned about the straight of butternut, so I used oak
everywhere that wouldn't be visible; going as far as making the pieces
showing in the front out of oak with an outside piece of butternut glued on.
I didn't float the top because all grain on the cabinet section runs the
same as the top. I did glue the sides together, and when they are
glued/screwed to the top, it should be pretty solid. In fact, with the top
off, it is surprisingly rigid.
Well, let's look at it from this standpoint toller. A typical raised panel
door is very similar construction to your support and leg pieces. In fact,
so is the upper portion of your piece. There is a piece of wood floating in
the raised panel door and that relieves stresses which could occur, as long
as the panel does not swell and become a very tight fit in the groove. Most
do swell to become a very tight fit though, and in fact they really don't
float very freely at time of fabrication - or else you'd have a sloppy
raised panel door. In this door, you have a very good glue joint, but
that's all you have is a glue joint. How many have you seen that have
racked apart due to the cross grain construction? Likewise, do you expect
your upper part of your project to suffer expansion/contraction problems?
You do have cross grain construction there as well.
I've seen a lot of pocket screwed pieces and they sure seem to work, though
I problably would not have used that approach. Not that it means anything.
A lot of times I do things out of habit and not because it's necessarily a
better way. Again, they hold. I'm not a believer in backing off screws.
Maybe I'm wrong with this but I believe that if you give room, more room
will be created over time.
No, actually I don't. Every bit of grain on the upper part runs parallel to
the top. The only thing that could be considered cross grain would be
drawer slides, and where they are won't be affected by a 1/10" movement.
Well, thats not quite true, the front is crossgrain to the sides, and I will
have to make the drawer fronts a 1/16" larger than I might like to
accomodate summer swelling. I saw that as a tradeoff for not having to have
a floating top.
would have pocket screwed the sides, but now they are biscuited.
So far all of my pocket screwed joints have held up.
No, I am not thrilled with the back off method; had I planned better I would
have build the shelf differently.
I have a 25 yo cherry table and a 15 yo oak table. Both are pocket screwed
to the aprons, and both are fine. The longest crossgrain joint is about 20
inches; so apparently they can absorb the stress without detereorating.
HomeOwnersHub.com is a website for homeowners and building and maintenance pros. It is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.