Troubleshooting a Moving Table Top

Another good discussion on a typical learning experience: a table top is restrained on one face from moving in service, and instead the wood cups. December 8, 2012

I have a 3/4 poplar, wood table top, 42" square that has cupped over two years due to the finish being uneven on the top and bottom. The top was out 1 1/4" over the 42" span. I removed most of the finish on the top, down to the primer. Now the finish is fairly even on both sides. Over the last three-four weeks, the top has straightened to where it is only out 5/8". Will it take another three or four weeks to go flat again? I keep it in the shop where we have higher moisture than in the heated and cooled areas. I thought this is best since the customer rarely uses their heat or a/c.

Forum Responses
(Furniture Making Forum)
From contributor L:
My first reaction is that is a huge bow for that length and would lead me to mistrust it in the future. I suspect the wood was unstable to begin with rather than it being just the finish difference. I wouldn't expect it to go back to flat as you've probably attained most of what you'll get. The cupped side has undergone some compression. Although there are some tricks that sometimes work, given that itís poplar, wouldn't it be safer, easier, and less expensive to rebuild the top with stable, flat wood?

From contributor R:
Is the top off the table? How was it fastened? You would need a fair sized slot with poplar for seasonal movement.

From the original questioner:
I have already replaced the original with an MDF top, but I like knowing what I have done wrong and hoping to get ten dollars of education from a several hundred dollar mistake. I was also thinking of adding two cleats (or runners) across the bottom to help with stability. I failed to mention that there is already an additional set of 3 /4" x 1" strips around the under outer perimeter of the table. I did this to help with stability and also to make the top look 1 1/4" thick. I really had not questioned the stability or quality of the poplar, the stock was kiln dried, and has showed no signs of cracking or checking, but I will now. I was actually surprised the wood had not split on the bottom due to being forced into such a tight curve. I did check moisture content just now, showing 8% on top and bottom, wished I had checked when it came back in.

From contributor O:
It sounds as if you created a cross grain restriction by adding the perimeter strips to increase thickness and stability. If the 42" wide top is not allowed to move and it is flatsawn poplar, the top will expand 3/8" with a 3% decrease in MC. The width must be allowed to float (to change width) with normal seasonal variations, despite the finish. The wood will move, no matter what restrictions you try to add.

From the original questioner:
The wood is flatsawn. I believe the wood itself is still in decent condition, there is no warping or twisting at all. Only the cupping across the width, so I can see where the strips underneath can be a problem. I did this table at the customerís request, and they actually would prefer getting this top back again. The customers are 75 plus years old. The man believes solid wood is best and he doesn't care for plywood or MDF, etc. Not a point I want to argue, but he would prefer the original top back, even with a slight cup. If I remove the strips and can get the top back to being fairly flat, would you guess that it would be more stable this time? This has nothing to do with money any longer, just trying to make an old man happy.

From contributor E:
Are your glue up pieces 3-4" wide, urned end for end? Are all the growth rings turned the same way? If they are wide this may be the issue. If so, could you rip the top and re-glue?

From contributor O:
There is no reason that a solid poplar top, flatsawn or otherwise would be anything other than flat if it was made flat and proper. Do you understand the problems with cross grain construction? Is that what you have made with the build-up strips?

I would consider starting over with some nice 6/4 rough poplar. Rip the widths, face, and edge on the joiner, then let it sit for two-three days. Face and edge again and glue for width, then widebelt. Attach to the frame with mounts that can tolerate 3/8" movement either way. Do not restrict its movement or you will be back where you started from.

From contributor B:
This is what I call chasing a problem. You will spend more time trying to repair this top and the finish it than it would take to glue up a new top and paint it to look like the old one- the customer will not know the difference. Use correct construction techniques, as have been discussed here, and you will get this done quickly, learn a good lesson, and make your customer happy.

From the original questioner:
Thanks everyone for all the good responses. I am not too concerned with wasting time on this top. I spend only a few minutes here or there, sort of a break from my other routine. I find these situations to be "good teachers" about what not to do in the future. I will soon reconstruct another top using thicker material as suggested. I will leave off the bottom edge strips, and I will continue flipping my boards to alternate the growth rings. I will finish the top and bottom evenly, just in case.

I believe the end result will be a well pleased customer who will have the solid wood top he desired, at no extra expense. I will have kept a valuable customer and friend, gained some worthwhile experience. I only joined WOODWEB after I built and delivered this table top, and I was able to determine problems would soon be arising from the information you guys were sharing. It sort of helped me to be ready, and have some really good explanations for the customer when he ask me what was going on. Thank you again everyone. WOODWEB is a big help to those of us in rural areas with limited resources.

From contributor D:
Using quartersawn material and sealing your product well will help the most for your next exterior tabletop. Have a look at fig. 4-3 on page 4-5 of Wood Handbook, Wood as an Engineering Material. It is available to download free on this site. It pretty much says it all. Alternating growth rings is good practice, as you can see wood does not just shrink or swell in width, but distorts in shape. Grain orientation means everything in a glue-up. I suspect those who think it is a myth are biased because they've been extremely lucky, their products have been isolated from moisture changes, or they just haven't done enough volume of work to detect it yet. Wood swells and shrinks across its grain lines the most, and sooner or later you have to accept it.

From contributor L:
In addition to what Contributor D says, I would add that the most critical part getting a stable flat top is in the preparation before glue up.

1. If the wood is significantly cupped or twisted at the start it has tension in it and will continue to cup or twist with moisture changes. This is not the same thing as expansion or contraction of a stable board. Choose different wood.

2. Get the moisture content close to final. "Kiln dried" doesn't mean much if it was stored in an unheated area. This can take several weeks.

3. Rough mill a little oversize, taking equal amount off each side if possible. Let it equilibrate stickered as long as feasible before final milling and glue up.

From Gene Wengert, forum technical advisor:
Alternating grain is indeed a myth. A top warps because the moisture is changing. The cleats or other restriction inhibit expansion or contraction on the bottom. The correct way to fasten cleats is to use a slotted hole so the screws can move as the top changes its MC seasonally.

A vapor resistant finish, top and bottom, will slow the changes, but not stop them. The MC will change seasonally, so a flat table will change MC and may warp then. One key is to make sure that the MC when you make the table is the same as the customer will experience. A fast change in MC from your shop to the customer's home is worse than the slow seasonal change.

From the original questioner:

This is an interior table, although the customers tend to not run the a/c much. That was probably my first mistake. We have an air conditioned shop, and the wood had been in there for several months during the summer while we ran the a/c consistently. Right now the weather is in-between most days, meaning we are not running heating or cooling, same for the customer. This means I probably have the most favorable matching conditions for the next four-five weeks in regards to relative humidity in both our shop and the home. I better get started.

From contributor W:
To the original questioner: Your top would have been fine without the cross grain pieces added as has been said but you did not say how you fastened it to the apron or frame? This is where you need to allow for the seasonal movements. The top should be allowed to move while staying flat . A slot running in the apron or frame mated with small blocks screwed to the underside of the top is one way.

From Gene Wengert, forum technical advisor:
We know that wood moves about 1% in size for each 4% MC change, which is about the change that occurs from summertime to wintertime in the interior of a building. This movement is across the grain, although quarter sawn grain moves about half of this amount. However, lengthwise, move is very stable (with a few exceptions) so movement is zero. Hence, any frame that has the grain of a cleat or support at 90 degrees to the table top (or a metal frame) and is securely fastened will stop movement of the top on one side and this will create seasonal warping. It is critical to allow the top to move on both sides seasonally. A finish does not stop moisture change.

From the original questioner:
I actually glued and screwed the strips to the underside of the table, all the way around, all four sides. I also understand that the strips (3/4" x 1 1/2") will not expand with the same rate of expansion, that will be seen in the wood in the top, as it expands across the width. I have attached a picture showing the cup (which is now 3/4") and some tape showing the orientation of the boards. It seems to me that if the strips are restricting movement the table top would be showing a crown instead of the cupping. Isn't Gene saying the bottom side is not expanding as much as the top side, due to the bottom strips being fixed in place?

Click here for higher quality, full size image

Click here for higher quality, full size image

From Gene Wengert, forum technical advisor:
Usually the wood is too wet for the environment, so the top dries and shrinks. It is rare, but can happen in the summer, that the wood is too dry.

From contributor Y:
A better solution for the thicker look is to make the top over length then cut a couple of inches off each end and "fold them under and glue. The grain is all the same direction the outside two edges are just parallel rips. Problem solved. Finish both sides the same and put a couple of extra coats on the end grain. You will almost always loose if you cross grain.