Avoiding concrete tolerance traps

October 2009 » Features » MATERIAL MATTERS
Beware of mismatched or unachievable tolerances
Bruce A. Suprenant, Ph.D., P.E.

There is often a gap between specified and achievable tolerances for cast-in-place concrete. Floor-to-ceiling interior partitions may be manufactured so they can be shimmed 1/4 inch to fit and then caulked, thus allowing concrete contractors only a 1/4 inch combined tolerance for the floor surface and soffit of the floor above. There are similar tolerances for exterior building finishes. Such tolerances, however, don’t match the tolerances for concrete construction given in documents of the American Concrete Institute (ACI). Further, some tolerances in the ACI documents are difficult, if not impossible to meet. The resulting disputes and corrective work can delay construction and increase costs, with no commensurate increase in billable hours for the design professional or adjusted prices for the contractor. Structural engineers should be aware of the following examples of such construction issues.

Interior partitions that don’t fit
When constructing a multistory building, the focus is on elevation control for the elevated slab top and bottom surfaces. Who is responsible for elevations deviating from the plan elevation, and what are the applicable tolerances? This depends on when the elevations are measured.

Section 4.4.1 of ACI’s Specifications for Tolerances for Concrete Construction and Materials Commentary (ACI 117-06), allows a ± 3/4-inch tolerance for the top surface of formed suspended slabs and Section 4.4.2 allows ± 3/4 inch for formed bottom surfaces before removal of supporting shores. For example, a designed floor-to-soffit height of 10 feet, built within allowable tolerance, can range from 9 feet 10-1/2 inches to 10 feet 1-1/2 inches. If the partitions can’t accommodate these dimensions, high repair costs can result in disputes regarding responsibility for repairs.

Some may claim that the ACI tolerance is too loose, but, if anything, it’s too tight. As-built measurements from several sources (see Table 1) indicate that the standard deviation for top-of-slab elevation is about 0.4 inches. Assuming a normal distribution, 99.7 percent of all measurements would be expected to fall within a ±1.2-inch envelope (± three standard deviations) instead of a ± 0.75-inch envelope. Thus, the ACI tolerance (about ± two standard deviations) accommodates about 95 percent of the field data. Based on as-built data, the ACI tolerance should be increased to ± 1-1/4 inches to more accurately represent what is being achieved on typical construction projects. But let’s accept the current ACI 117-06 tolerance and explore another possible cause of elevation tolerance disputes — deflection.

Surface elevation measurements may be made when the partition fit problem is first noticed. Note, however, that surface elevation measurements for a tolerance check must be made before removing supporting shores. The window during which these measurements can be made has closed long before partitions are installed. Requiring the contractor to maintain a ± 3/4-inch elevation envelope for slabs after shoring removal does not recognize slab deflection resulting from self weight (unfactored dead load).

Structural design in accordance with ACI’s Building Code Requirements for Structural Concrete (ACI 318-08) requires the engineer to limit deflections that adversely affect structural serviceability by using a minimum overall thickness, or by setting maximum permissible values for computed deflections. Unfortunately, maximum permissible computed deflections are defined in Table 9.5(b) of ACI 318-08 as “that part of the total deflection occurring after attachment of nonstructural elements (sum of the long-term deflection due to all sustained loads and the immediate deflection due to any additional live load).” A footnote tells how long-term deflection must be determined, but also notes that it may be reduced by the amount of deflection calculated to occur before attachment of nonstructural elements. Thus designers using slab thicknesses less than the minimum overall thickness may claim to be responsible only for long-term deflection, which does not include initial deflection due to the self weight of the slab. This is a loophole in the ACI 318 requirements because designers aren’t responsible for self-weight deflections and neither are contractors that meet the project specifications.

To counter claims of responsibility for excessive deflections observed after shoring has been removed, some contractors take elevation readings on the slabs while the shoring is still in place. If elevations are measured after shoring removal, they then have data showing that the slab elevations were within tolerance after concrete placement and finishing. Even then, there are sometimes disputes about application of the ± 3/4-inch elevation tolerance envelope. The envelope is centered on the design elevation, so the surface of an undeflected slab could still be 3/4 inch lower than the design elevation. Deflection measurements must be related to the as-built initial slab elevation — not to the design elevation. If no measurements of the initial slab elevation were made, elevation readings of the slab surface at the columns should be taken as the baseline for additional slab surface measurements.

Stucco finish tolerances don’t match concrete frame tolerances
When a building consists of a concrete frame with concrete block infill walls and stucco exterior finish, the plastering contractors expect the columns, edges of floors, and block walls to be within 1/4 inch of the same plane. They base this on Section 5.2 of the American Society for Testing and Materials’ Standard Specification for Application of Portland Cement Plaster (ASTM C 926), which requires the concrete surface to be “straight and true to within 1/4 inch in 10 feet.”

This is a case of dueling specifications because ACI 117-06 sets less restrictive tolerances for the concrete frame. Section 4.1.1 allows a surface deviation from plumb of 0.3 percent of the height above the foundation. For a 10-foot-high wall or column, this is roughly 3/8 inch in 10 feet. Section 4.2.1 of ACI 117-06 allows the edge of a floor to deviate from the plan location by ± 1 inch as compared with the 1/4 inch allowed by ASTM C 926. If the outer side of the infill masonry wall is built flush with the edge of the floor, this would partially solve the problem, but the difference in plan location of outer edges of floors and columns can still miss “straight and true” by much more than 1/4 inch.

In post-tensioned concrete frames, movement of perimeter slab edges and exterior columns can exacerbate the problem. Some shortening of the slab occurs after the tendons are initially stressed, but time-dependent drying shrinkage, creep, and possibly contraction due to cooler temperatures can result in a final shortening of 3/4 inch or more. Thus, if there is already a difference in edge location of the columns and slabs (within tolerance), it can be accentuated, primarily by the combined effects of tendon tensioning plus concrete shrinkage and creep.

Thin-set stucco applications have increased the stucco contractors’ demand for tighter tolerances. ASTM C 926 indicates the nominal plaster thicknesses for two-coat and three-coat work on concrete are to be 3/8 inch and 5/8 inch, respectively. For three-coat work with a metal plaster base the nominal thickness is 7/8 inch, which still doesn’t match the concrete tolerances. However, Section 5.2.2 of ASTM C 926 states that if the concrete substrate is smooth or nonabsorbent it shall be prepared to receive portland cement plaster methods that include application of a dash-bond coat applied forcefully against the surface, left untroweled, undisturbed, and moist cured for at least 24 hours (Section 5.2.22). Section 5.2.3 then states that where bond cannot be obtained over the entire surface to receive plaster by the dash coat or other alternatives, or where total plaster thickness will exceed the total thickness mentioned earlier (3/8, 5/8, or 7/8 inch) for concrete bases, furred or self-furring metal plaster base shall be installed. Using the metal base and thicker plaster does increase cost, and also may cause durability problems in areas near salt water because the metal base may corrode. But it does come closer to matching the concrete tolerances.

As one solution for this problem, the owner could provide a bid allowance to the stucco contractor for installation of any needed furred or self-furring metal plaster base and extra plaster thickness. Providing an allowance enables the owner to compare plastering bids on an equal basis. If all the allowance isn’t used, the money is returned to the owner.

One of the impossible tolerances
Section 7.5.2.1 of ACI 318-08 states that, “Tolerances for d [effective depth] and for concrete cover in flexural members, walls, and compression members shall be [per Table 2, page 35] except that tolerance for the clear distance to formed soffits shall be minus 1/4 inch. In addition, tolerance for cover shall also not exceed minus 1/3 the concrete cover specified in the design drawings and project specifications.”

The same tolerances for effective depth and concrete cover over reinforcement were also included in previous code requirements. But the absolute tolerance of 1/4 inch for the clear distance to formed soffits and the 1/3 limit on cover tolerance are often at odds with the effective depth tolerances.

The one-third reduction requirement was added in the 1971 revision to ACI 318 because the minimum allowable cover for cast-in-place shells and for the ties, stirrups, or spirals in precast concrete was 3/8 inch. Section 7.3.2.1 of ACI 318-71 also allowed a placing tolerance of ±3/8 inch. Applying that tolerance could thus reduce an allowable 3/8 inch cover to zero.

Allowing a tolerance of only 1/4 inch on clear distance to formed soffits, regardless of the specified cover, was first required in Section 7.3.2.1 of the 1974 supplement to ACI 318-71. The rationale for this requirement later appeared in Section 7.5.2 of the Commentary for ACI 318-83: “More restrictive tolerances have been placed on minimum clear distance to formed soffits because of its importance for durability and fire protection, and because bars are usually supported in such a manner that the specified tolerance is practical.”

The last portion of this statement is debatable. Are tighter tolerances justified because bars are usually supported in such a manner that the tighter tolerance is practical? Data for in-place concrete published in 1979 show that, on average, bottom bars in slabs were 0.31 inch below the specified location (Mirza and MacGregor, 1979). This indicates that in 1979, more than 50 percent of the bottom bars in slabs would not have been within the –1/4 inch tolerance to the soffit. Later data by another investigator supports this. Designers may recognize that many bottom bars will be out of tolerance and specify a concrete cover of 1-1/2 inches. Even though the one-third rule then allows a tolerance of 1/2 inch, the minus 1/4 inch tolerance is an absolute value that controls. And if contractors increase the height of the chairs or bolsters by 1/2 inch, they risk being out of tolerance on the plus side for d. Although a less restrictive tolerance on minimum clear distance to formed soffits is needed, changing the tolerance in ACI 117 won’t be possible unless Section 7.5.2.1 of ACI 318-08 is first modified during one of the code revision cycles. Until then, contractors are faced with a tolerance that is impossible to meet.

Solutions for tolerance problems
In many cases, incompatible tolerances at the interface of cast-in-place concrete and other structural systems can be reduced by using adjustable connections that can accommodate the concrete tolerances. This is usually a less expensive solution than requiring tighter concrete tolerances. When concrete tolerances are not generally achievable with current construction practices, the tolerances should be changed. As-built data have shown that there is a need for changing many such tolerances (Suprenant and Malisch 2009). However, changing the clear cover tolerance mentioned in this article requires a revision of ACI 318, which will take years for approval. And changes in ACI 117 are also years away. Although the time needed is an impediment, the necessity for change is apparent.

More tolerance information
Bruce Suprenant, P.E., and Ward malisch, P.E., are co-authors of tolerances
for cast-in-Place concrete Buildings, published by the american Society of concrete contractors in 2009. the book contains a summary of thousands of as-built measurements that indicate what tolerances are being achieved in buildings that perform well. it also includes descriptions of frequently occurring tolerance disputes and offers solutions for many of these. the book can be purchased by calling 866-788-2722 or visiting the aScc web site at www.ascconline.org

  • ACI 117-06, Specifications for Tolerances for Concrete Construction and Materials Commentary, American Concrete Institute, Farmington Hills, Mich., 2006, pp. 18, 33, 35.
  • ACI 318-08, Building Code Requirements for Structural Concrete, 2008, p. 89.
  • ASTM 926-04, Standard Specification for Application of Portland Cement Plaster, ASTM, West Conshohocken, Pa., 2006, 10 pp.
  • Mirza, S. and MacGregor, J., Variations in Dimensions of Reinforced Concrete Members, Journal of the Structural Division, ASCE, 1979, April, p. 57.
  • Suprenant, B. and Malisch, W., Tolerances for Cast-in-Place Concrete Buildings, American Society of Concrete Contractors, 2009, 143 pp.

Bruce A. Suprenant, P.E., is president of Concrete Engineering Specialists in Boulder, Colo. He can be reached at 303-499-0264 or suprenant@comcast.net. Ward R. Malisch, P.E., is technical director of the American Society of Concrete Contractors, St. Louis. He can be reached at 248-449-4444 or wmalisch@ascconline.org.

Upcoming Events

See All Upcoming Events