Cracked Tile

[ltlspf]

I have a contractor that would like us to review cracked tile at an new residence that has just been built. The tile is something like travertine, softer than some with 16" square tiles, placed diagonal to the wood I-joists below. There is also 1 1/2" of lightweight gyp-crete between the top of ply and bottom of tile bed. The wood joists are at 16"o.c. with 3/4" T&G plywood perpendicular. The contractor stated that the cracks seem to be at 4 ft o.c. located near the plywood joints and that they occur randomly along the joist spans. The total deflection of the joists is around L/700 for a 17'-0" span.

Not sure if this is more of a tile installation problem or one due to the specifics of the supporting structure? Maybe some sort of differential shrinkage between the gyp-crete and the mortar bed of the tile?
Any ideas?

Thanks for any input!

 

[spats]

I've run across this before, but I can't remember all the specifics. It's basically a matter of the tile needing to be totally unbounded from the substrate/structure by use of a slip sheet, and detailing so that the two can move independently.

 

[ishvaaag]

The tiles seem to follow the plywood deformation, that seems be not stiff enough. We use to accept service loads total deflections L/500 as acceptable for non structural loads but this is an oversimplification. We should be looking as in reinforced concrete for live deflection on what existing. We should be looking more at distortion than at deflection, or one of these indexes based on panel distortions published over there. Even so, secant distortion of 1 cm in 500, is 1 centimeter in 250 cm. Distortion around the supports, tangent instead of secant, is over 1.5 times that, or 1.5 cm in 250. I remember the old texts of AISC limiting distortion by drift (between beam and columns) to 1/400. If our distortion at point of maximum slope had to compare to this, it would had 1 cm in 166 distortion (not pears with pears but get the idea). Now, when following subsidence in built areas by tunnel construction, care is taken on not producing above 1/400 distortion between column supports, more or less equal to the old AISC criteria for drift. This is taken a value for which the problems caused by subsidence are still able to be repaired by ordinary means, not a warranty of there not being problems. But ... the bad news is that masonry of brick work of well cooked terracotta breaks before 1/1700 deflection. Stone may be less fragile, but not always.

Hence look at the deflection of the plywood. The contractor is signalling the joists as hardpoints, hence the problem could be just in the plywood, and maybe could be reinforced to lesser deflections, something to be studied.

 

[ltlspf]

Thanks for your input.
Spats, I did find some more info online that stated the same. They call it a crack isolation membrane or anti-fracture membrane that needs to be installed between the gyp-crete and the tile bed to separate any minor cracks that may occur in the gyp-crete layer from extending up into the tile. I will have to check with the contractor and see if this may be the issue.

 

[Ron]

If you have l/700, I wouldn't look at the structural support as being the issue. More likely shrinkage cracking reflecting through. Your thinset bond to the gypsum must be great!

A crack isolation membrane will correct this issue for the most part.