Replies

1. 
   Westchester | Feb 16, 2013 05:50am | #1

   Welcome

   Not sure what design you're describing however long grain to long grain alternating the boards has traditionally proven to be the better choice.  Can you submit a drawing of your table top ?

   SA

2. 
   RalphBarker | Feb 16, 2013 12:31pm | #2

   Assumptions

   If I understand your description correctly, your "widthwise" approach would place shorter boards, still glued long-grain to long-grain, at right angles to the length of the 8 1/2' table. This also assumes that the width of the table would be something less than the length.

   The result of this approach would be a significant increase in the amount of seasonal expansion and contraction, since the majority of expansion/contraction occurs cross-grain. The amount of comparative increase would be a function of how wide the table is, but it might be a factor of 3x or so.

   Since the majority of expansion/contraction would be occuring along the length of the table, the design of the supporting structure, and how the top is attached to the support structure, would both need to accomodate that extreme difference, or face failure of the table's structure.

   How much expansion/contraction the table would face depends on where the table is located. But, seasonal differences range between 4% and 14% in the U.S. With the traditional lengthwise board orientation, the width of a 36" table might vary as much as a half inch across the width. With the widthwise design, that might be as much as 2 1/2" between seasons.

   This is a handy tool for calculating expansion/contraction:

   http://www.woodbin.com/calcs/shrinkulator.htm

1. 
   Walnut74 | Feb 16, 2013 06:52pm | #3

   I don't yet have a complete drawing of the table as this one issue was something I wanted to look into.  

   [ If I understand your description correctly, your "widthwise" approach would place shorter boards, still glued long-grain to long-grain, at right angles to the length of the 8 1/2' table. This also assumes that the width of the table would be something less than the length.]

   Yes this is exactly correct.  The table would roughly be 100" x 44".  I already assumed that the top would be floating, most likely with figure eight fasteners.  So from these replies I am not necessarily looking at compromising strength, just mutliplying the expantion/contraction amounts?  And that makes sense to me given the cross grain movement.  

   So as long as it is fastened in a way to accomodate this movement, will the majority of this movement be expressed in the in the ends of the table and the amount that they hang over the legs/aprons?

   Also wondered if it would be a dangerous game to try and put trim on the edges, if not impossible.  Or am I prob better off just emracing the look of end grain across the length?  Hey I really appreciate the replies and I will definately check out the exp/contr calculator.

1. 
   gdblake | Feb 16, 2013 09:04pm | #4

   Your table top may be substantially weaker

   It is not a good idea to run the grain side to side instead of the length of the table.  For proof, cut an 11" square of 3/4" pine.  Now karate chop it perpendicular to the grain.  Quickly rub your hand until it stops hurting and then try the same karate chop parallel with the grain.  The pine board will break easily this time.  The cell structure of wood makes it stronger across the grain than with the grain.  That said, if you have some method of support running the length of the table you can get away with this approach.  Just make sure you have adequate support in your design or come Thanksgiving the table may split in half just as you start to carve the turkey.

   gdblake

2. 
   RalphBarker | Feb 17, 2013 10:58am | #5

   Movement

   I like to think of wood movement as "starting everywhere" and ending "where you tell it". Naturally, it is "cumulative" across the expanse, every board being moved, and adding its share to the accumulation. It is also possible to "direct" it by fixing one line across the expansion. Anchor the center and movement is directed to both ends. Anchor one end and all the movement is directed to the other end.

   I don't think the typical figure-8 fasteners would allow for enough movement. Instead, I'd use cleats - small blocks with tabs that would travel along a dado in a support beam of some sort, hidden under the top. During the design work, consider adding an alternate method for lifting/moving the table, as well, so the weight isn't being supported by just the little cleat tabs.

   Also, I wouldn't attempt to cover the end grain with a breadboard end side. There is just too much cumulative movement to handle in that manner.