Replies

1. 
   Scooter1 | Jun 19, 2003 02:02am | #1

   Kroozer:

   Yep, I know exactly what happened. I see it all the time. Your tile setter got lazy and installed what is sometimes called a "scratch coat" installation.

   But before I get to details, I re-read your post and make the following assumptions in the order of the materials:

   3/4 ply; radiant heat; mesh; thinset; and tile.

   Thinset is not an appropriate medium for setting tile on plywood. The problem you see is that plywood, being wood, moves at one rate (wide variations with temperature and humidity) and thinset and tile, being a masonry product, moves at a different rate. The thinset stuck to the moving plywood but was sheared off from the tile, which couldn't move as much.

   When installing radiant heat I install a mud bed over it; in fact, I install a mud bed over all floors. It is the old fashioned way of doing things, and it just happens to the best way. No new product can replicate what a mud bed does for the flatness, strength, and sticking ability of a setting bed.

   Younger guys are using SLC's over radiant heat now, and I have tried that as well, but it is not as flat. Not bad, a tad easier, but not as flat.

   Some even use Gyp-Crete, but if it ever gets wet, you are in trouble. I won't use that stuff.

   If you are not installing radiant heat and tiling directly over plywood, you need a "de-coupler", a fancy term for a membrane which separates the two products which expand at different rates. That "de-coupler" can be a sheet membrane like NobleTS or a PVC product, a liquid membrane like Laticrete Blue92; or a real new and nifty product, Schulter's Ditra. I like Ditra a lot. It is very thin and Schulter guaranties it over plywood. I don't know if Ditra is designed for installation over a scratch coat with radiant heat. Call up the technical rep and ask them. They are in NY.

   The scratch coat method(s) I have seen all fail. I have seen them over fiberglass mesh and they fail; over 2.4 and 3.5 diamond lathe, and they fail, and over 2x2 wire, and the fail. Its just that the thinset has no structural strength and is gluing together two products which expand much differently. It is a recipe for a disaster. It is like making a cabinet out of very wide wood and trying to glue end grain (doesn't expand very much) to side grain (which does expand). The cabinet or box generally cracks, right?

   I would pull up the tile and float a mud bed, like the old foggies (like me) do. Next best choice might be an SLC or Ditra. Sorry.

   Regards,

   Boris

   "Sir, I may be drunk, but you're crazy, and I'll be sober tomorrow" -- WC Fields, "Its a Gift" 1934

1. 
   DaveGobis | Jun 19, 2003 03:17am | #2

   Any movement accomodation joints at the wall line? Versa Bond is a good thinset. However, it is rated as good by the manufacturer, CFlex, better, and FlexBond best. Most manufacturers use a three tier system. Performance of VersaBond is only 2/3 that of Flexbond.

2. 
   Mooney | Jun 19, 2003 07:13am | #3

   Here Here ! Very well done !

   Tim Mooney

3. 
   hdcruiser | Jun 20, 2003 12:15am | #7

   Boris,

   Thanks for the indepth assesment. However, here's the rest of the story: The radiant heat manufacturer recommended the aforementioned installation technique. Second, mud job would not work, as this is an "open kitchen" layout, where there is no wall between the kitchen and the dining room / living area. The mud base would raise the level of the tile floor way above the adjoining wood floor. Third, we have used the tile on plywood method on numerous jobs before (no radiant heat, though), with great results. I even did my own kitchen the same way and it still holds strong and solid after ten years! I am really starting to think that it was a bad batch of Versabond (from Home Depot, which should explain it). Now, i need to pick up the tiles, but the clincher is that there are kitchen cabinets sitting on them! I cannot remove these, as there are granite countertops installed on top! How's that for being f#$%@'d???

   Kroozer

1. 
   Scooter1 | Jun 20, 2003 01:12am | #8

   Kroozer:

   You are lucky that it worked in the past.

   I can tell you that both my direct experience (over 20 years) and texts (2 books and various handouts) and classes at the Tile Council all tell the students that direct tiling over plywood is a no-no, except as I explained. If you are doing your own tile, I might suggest such a class. You can reach Dave Gobis at http://www.tileschool.org/ He is an old tilesetter and has seen it all and is in charge of education at the TCA. He sometimes even posts here.

   If this was the work of a sub, then I would change subcontractors.

   If this method was honestly recommended by a manufacturer of the radiant heat product, then I would look to them to honor their warranty. One of the things I have learned is that while I might think a certain method is the "right" (read: ONLY) way to do a job, if a manufacturer of a tile product specifies a different method and is willing to back it up with a warranty, then you go for it, and you truly find out if the mfgr is full of s h i t or not. One recent example that is coming up is tiling on OSB. For 20 years, everyone would tell you that it is a no-no, and one has to membrane it (the thinset causes it to swell) . But suddenly Schulter says it now OK. So, even though it this not "right", if Schulter says it is OK, then I guess it is OK.

   So I would file a claim against the manufacturer of the heating product. If they are willing to sactify a "wrong" method, then your installation is on their nickle.

   Regards,

   Boris

   "Sir, I may be drunk, but you're crazy, and I'll be sober tomorrow" -- WC Fields, "Its a Gift" 1934

2. 
   db68 | Jun 19, 2003 08:23am | #4

   sounds like a bad batch of thinset, or incompatible thinset for porcelin tile.  if the thinset to plywood is sound you could install the porcelin with new thinset over it and risk the plywood movement issue.  however and isolation membrane is the better choice.  plywood and tile do not mix.   Ditra is a nice product to use and goes down quick.  I have only been using it for about two months and don't know how it will fare twenty years from now.  good luck.

3. 
   kai230 | Jun 19, 2003 06:17pm | #5

   I personally saw the tile installer butter the back of the tiles before installing them,

   A sound tiling surface usu precludes the need for back-buttering, unless the tiles aren't flat. It takes a skilled eye or experienced tradesman to know when buttering is needed. Just because they were buttered doesn't mean they were buttered properly, or installed and tamped down properly.

   I have no idea re the radiant heat thing, but I'd be surprised if it wasn't a contributing factor.

4. 
   djohan | Jun 19, 2003 08:24pm | #6

   I've read most of the posts and pretty much agree with the idea that the thin set was not sufficient or that there needs to be a release between the tile and the plywood.  Personally, I think the problem is that there is too much flex in the plywood to make a complete base for the tile.  If I read your post correctly, for all intents and purposes, there isn't anything on top of the plywood that adds additional strengh to the job.  Although additional thickness would slow down the reactive time of the heating, I think you need an additional half inch of wonderboard over the plywood to stop the flex that is going on which is popping the tiles off.  I've seen many jobs done directly over plywood with a bond breaker and many of them fail because the plywood just isn't capable of providing a solid base for the tile by itself.

5. 
   NOTRIX1 | Jun 20, 2003 04:02am | #9

   I think it's the thinset on plywood that's the culprit as Boris points out.

   I work as a carpenter for a tile contractor and he ALWAYS lays down something between ply and tile. The current job is Denseild, thinset and screwed down to the ply subfloor.

   I don't know much about the tile work but always get "the look" when I ask if we're gonna set the tile on top of the plywood.

   Sorry for your troubles. I hope you find an easy fix.

   N