Over the years, we have spent a lot of time discussing various means of bonding tile, but often we do not give enough consideration to the actual underlayment material which we ultimately bond to. Just what is an underlayment anyway? Underlayments can do many things, but their purpose is to provide a base of support for a tile installation.
While some types of underlayments have great strengths, such as concrete, the actual support must always be provided by the structure. In slab-on-grade applications, this is rarely a problem. However, in suspended slabs (slab-on-deck), many variables can affect their suitability. High-strength underlayments such as backer boards, self-leveling compounds and even mortar beds will not compensate for structural deficiencies -- in particular lack of support. Some products or installation methods may tend to mask the lack of structural support in the short term, but the structural inadequacy often surfaces at a later date. Retailers, contractors and installers with the help of marketing statements, tend to get too caught up in the “strength” of an underlayment. Without question, each type of underlayment product needs to have attributes that provide a solid base of support for the tile installation. That is why an underlayment product having a compressive strength of 30 PSI can perform as well -- or in some cases much better -- than one with 7,000 PSI.
Many methods listed in the Tile Council of North America (TCNA) Handbook utilize or mention membrane systems. Most traditional underlayment products have standards listed under the American National Standards (ANSI) A118 Material Specifications. The tile world revolves around both of these documents. They are not intended as market documents. As a matter of fact, you cannot find a single manufacturer listed in them. To get an underlayment method in the handbook requires substantial independent testing and documentation, plus an approval process. The requirements for a product to become an American National Standard are even more rigorous. Does this mean products that are not in TCNA methods or not covered under ANSI are inferior? Not necessarily. It takes years of proven performance to get into either document. To give an example: Wonderboard was the original cement board patented in 1968; however, it did not become a method until competitive products were introduced to the marketplace, and basic testing criteria were established. The first year you find the title CBU, Cementious Backer Unit (Wonderboard) in the TCNA handbook is 1985. On the flip side of the coin, roofing felt has been used for a tile underlayment for many years. It has been tested for basic performance as an underlayment many times in all manner of conditions and consistently fails the basic performance tests for an underlayment. There is also a substantial history of failure as a tile underlayment. We will confine ourselves to proven products and methods.
Self-leveling compounds are seeing increasing use as underlayments. This can be a sore spot for many installers, as getting paid appropriately for using these products can be very challenging. The term “self-leveling” is a bit of a misnomer. A more accurate description would be floor flattening underlayment, but there is nothing catchy about that term. It is indeed possible that a floor may be made perfectly level by their use. Before you and your customer commit to using these types of products, make sure everyone understands what the term “flat” or “level” means, and establish your goal accordingly. Growth in this product category is driven by big tile, which demands flat floors -- flatter than possible to achieve within the skill level or equipment of many concrete contractors. But, it is possible to achieve the desired floor flatness with concretes. You only have to visit your local Big Box or Sam’s Club to see what “super-flat” floors look like. In new construction, it is possible to specify flatter floors than the typical 1/4 inch in 10 feet no more than a 1/16-inch variation in 12 inches called for under the recommendations published by the concrete industry. However, when doing 18- or 24-inch tile, the cost of placing a floor with the 1/8 inch in 10-foot tolerance needed for large-module tile would double the cost of the concrete work at a minimum. When it comes to wood structures using dimensional lumber floor joist, the saying is “crown up,” but any crown is a potential flatness problem for tile floors. Due to the increasing preference for ever-larger tile sizes, we are seeing an increasing usage of both mortar beds and self-leveling underlayments in both residential and commercial applications. Complete details for either would have to be the subject of their own articles.