Replies

1. 
   deadnuts | Jul 29, 2015 08:47pm | #1

   structural cracks?

   zzzagman wrote:

   The deck was structurally built to support the current tile, but due to the flexing in the wood, there are numerous cracks.

   This sentence contains two statements that contradict one another. If the latter is true, then the former apparently is not. If so, I recommend correcting the former before proceeding with your plan. I know this may sound strange coming from me, but your ultimate solution might involve a professional engineer.

1. 
   zzzagman | Jul 30, 2015 12:43pm | #3

   About structural soundness

   Thanks for the answer. To clarify, the deck rests on glulam beams that span about 20' and is surrounded by solid structures on all 4 sides. The pavers would terminate into these firm borders.

   As far as the comment about structural integrity... I am a mechanical engineer and --although I am not a specialist in building structures-- I understand the concepts. I can assure you that the architect and the structural engineer that designed the house *did* size it for the tile finish described. (I may be wrong, but I believe it was 24# dead load for the backer, thinset, tile and grout.) When the cracks appeared we had the calcs verified and the structural sizing *is* correct for the load.

   What is *not* well designed is the combination of a flexible structure covered with a rigid facing, such as tile. When large live loads occur, like 25-30 people parties, the glulam deflects within normal ranges, but the rigid tile cannot withstand that amount of flexing. We are trying to address that problem by replacing the rigid tile finish with a paver system, which is more flexible and would allow that amount of movement. (We are partial to a tile-like finish for aesthetic reasons and would strongly prefer it over the exposed wood decking look.)

   The question is for suggestions on how to approach the installation of pavers over a waterproofing membrane, without damaging it. Instinctively I am thinking of a nonwoven geosythetic, of a certain thickness and firmness. I envision the tiles resting on it, in lew of the sand bed. We could use sand "in between" the pavers, to provide the friction lock.

   I would appreciate comments and/or suggestions on this issue.

   Thank you.

1. 
   deadnuts | Jul 30, 2015 06:43pm | #4

   'firm' is a relatively loose term

   So what you're saying is that your deck was "correctly" designed so that the tile would crack under certain live load conditions. Yeah, okay; I doubt it.

   At any rate, if:

   a.. your deck is designed for the appropriate amount of dead and live loads  (building materials and people)

   b.. your waterproofing membrane system is sound

   c. your edges are designed to restrain any potential lateral thrust of floating paver system

   then you should be able to set and level your pavers on a bed of smooth sand with same loose material between pavers. Grains of sand are like millions of ball bearings allowing the paver load to only transmit axially to the supporting structure.

   I would still install a flexible control joint at the edges (strips of low density, closed cell foam similar to sill sealer) to allow for thermal expansion and contraction of the masonry finish so that your paving system does not exert excessove lateral loads on your restrained edges. If you're concerned about frictional wear across your membrane, perhaps you could use Schluter's Ditra mat as an additional decoupling, waterproofing, and wear layer. I think it would be good, cost effective insurance. Still, I can't imagine that friction would be a major membrane longevity issue unless you're in a heavy seismic activity area . If that is the case, then it may be a signal to move.

2. 
   mark122 | Jul 29, 2015 10:35pm | #2

   if your grout cracked before then you have to much movement regardless of how sturdy you may think your subfloor is. 

   how long are your floor joist, and what depth? a step up from 8's to 10', or 12's vs 10's would cut down alot of your deflection. A second layer of subfloor would also help but not nearly as much as the larger joist.

   i would opt for a sand base but how are you going to keep it from washing out? how are the pavers terminating at the exposed edges?

   mastic and weather are a bad idea.

3. 
   Mongo | Jul 30, 2015 09:47pm | #5

   I would have...

   ...recommended you use a true outdoor tile-on membrane, I've been using NobelDeck.

   You have a wood framed structure. Sure it might move a bit. But you have floating cement board. That'll move. You had tile spanning cement board joints. Those will crack for sure, even without a party going on.

   If you do pavers over what you have, would a polymeric sand type of filler work? Just offering an idea. Polymeric sand really isn't in my building vocabulary, as I'm not a paver type of guy.

4. 
   Mike_Mahan | Jul 30, 2015 11:03pm | #6

   Ballasted EDPM

   I'd be surprised if saltillo tile over unsecured backer board didn't crack. I'd guarantee it would regardless of the structure.

   I think what your want is a ballasted EDPM roof system with the pavers serving as the ballast. Replace your vinyl shower pan material with EDPM. Literature should be available from the manufacturer on how to use pavers for ballast.

5. 
   wmheinz | Jul 31, 2015 06:38pm | #7

   Paver pedestals are used on sloping decks to give a flat surface to install pavers.  Typically the pavers are 2' square x 2" thick hydro-pressed concrete pavers intended for this use.  They are supported at each corner by a plactic "pedestal" that also supports other surrounding paver corners.  One that I have used several times is Appian Way Systems (http://appianwaysystem.com/pedestal-systems/) which uses a cap, a base and a piece of PVC pipe cut to lenght required to make the system level.  There are several other companies...just google 'paver pedestal system'.  This system allows water to drain though the joints between the pavers, but keeps debris out. 

   These pavers weigh in at about 20 pounds per square foot - don't know if your roof system was designed for this much weight or not...becuse there is nothing actually attached to the structure or grout joints, the roof structure can flex some and affect the installation. 

   However, before stepping off into this type of design, I would really want to have a handle on why the old system failed.  I'm not surprised it did...just the flexing of the unattached concrete board would have been enough to destroy the grout/tile bond.

   
1. 
   Mongo | Jul 31, 2015 09:31pm | #8

   hah!

   I read "pavers" and I was thinking of the little sidewalk pavers. Told you they weren't in my vocabulary! lol

   Your recommendation? Perfect.

   He could even do EPDM with small EDPM sacrificial squares in place of your leveling feet.