The Sounds of the Unsound - Though it can't be seen with a cursory glance, an expert ear can tell that the hollow sounds plaguing this tile installation were caused by application of setting materials with an improperly sized trowel.
Suppliers, installers and consumers of ceramic tiles tend to believe that there is little mystery in the installation of ceramic tile and grout. However, there are a number of complaint situations that seem to defy explanation. Others exhibit more obvious reasons for failure.

Some bond problems are obvious when an installation is performed over an unsuitable or questionable substrate, or over concrete that has been treated with curing compounds. Form-release compounds used in lift slab construction will also create bond problems. Shot blasting is generally accepted as the best method of removing these compounds.

Testing for the presence of these problem-causing substances - as well as for contaminants such as oil, etc. - is quick and simple. All you need do is thoroughly sweep clean the suspect area of the slab and drip beads of water on its surface. If the beads are not absorbed by the concrete within a few minutes, a potential bonding problem is indicated.

The stains in this tile floor were caused by use of an unsuitable substrate.
Certain substrates are classified by the industry as "unsuitable" or "questionable." In fact, the National Tile Contractors Association (NTCA) has compiled a list of such substrates. A few of them stand out in particular. On the unsuitable list, I've seen problems with fire-resistant plywood, particle board, lauan plywood, felt paper, and scribing felt.

The questionable list contains 21 different substrates. Any of these questionable substrates may be suitable provided the setting material manufacturer has designed its product to accommodate these situations. Therefore, it is important to consult with the manufactures for guidance if you are dealing with a questionable substrate.

The questionable category also lists substrates that have had applications of old cutback or other adhesives. Removal of almost all the cutback is vital before proceeding with an installation because, even though the old adhesive may appear dry, the application of another setting material (and the tile itself) may create a vapor barrier that causes the cutback to dissolve when moisture is present. This could cause bond failure and/or staining. A painted substrate could re-emulsify for the same reason.

The cracks in this tile were the result of point-loading over an uneven substrate.
The questionable list also names gypsum concrete. Although problems may occur, methods have been devised to allow for satisfactory ceramic tile installations. You may use a sheet membrane for isolating your tile, or use a water-free urethane adhesive that will not affect the gypsum concrete. Just be sure you choose products approved by the manufacturer for use over gypsum concrete.

Cork is also listed as questionable. However its use is becoming more accepted due to the need for crack-isolation systems and sound-control installations. Again, it is very important that you only use cork that has independent laboratory data to back it up. The thickness and density is important, as you may experience corner cracking in the title due to heavy point loading.

But there are more disconcerting facts to consider. Loss of bond can be viewed as an adhesive failure if your setting material releases from the substrate or the tile. If, in the loss of bond, your material has bonded to the substrate and the tile but has separated within itself, you've experienced a phenomenon called "cohesive failure." In this case, the problem resides with the setting material.

There are a variety of reasons for cohesive failure of bonding materials. Often, mixing water evaporates too quickly for the mortar to develop sufficient strength. This can be caused by tiling in hot, dry weather and/or using warm to hot water from garden hoses. Over-watering or re-tempering the mix with water, and mixing the material at too high a speed or for too long, can create air bubbles. The exterior forces - such as traffic or floor vibration - that are applied to the mortar before it has cured and developed its specified strength can also be at the root of the problem. Old, out-of-date mortar may also fail cohesively.

Adhesive bonding failures may occur for various reasons. Concrete curing compounds, form-release compounds, painted or otherwise contaminated substrates, improperly applied setting materials, use of incorrect trowel sizes, mortar that's been allowed to skim over, deflection of floors and walls, mold release on the back of the tile, and dusty backs (a common problem with Mexican pavers) can all contribute to bonding problems.

And, of course, you may discover a cohesive failure (resulting in powdery mortar) which may require analysis by a testing laboratory. In one instance, it was discovered that the water-retarding chemical in the material was insufficient and the mortar underwent incomplete hydration as a result. This in turn created weakness and, ultimately, cohesive failure.