When floor flatness and levelness of elevated slabs
are measured using Fnumber technology, timing of measurements is
critical because of the shortterm deflection between shoring and
reshoring and longterm deflection after shores are removed. This
deflection decreases Fnumbers that then may not meet specified values.
The higher the initial F_{F} number, the greater the negative eF_{F}ect
of deflection. Because of this, Section 4.8.4.1 of ACI 11706 and ACI
11710, "Standard Tolerances for Concrete Construction and Materials,"
requires that, "Floor test surfaces shall be measured and reported
within 72 hours after completion of slab concrete finishing operations
and before removal of any supporting shores."
The required time limit applies to both Fnumber (F_{F} and
FL) and gapunderstraightedge measurements made in accordance with ACI
11706 and ACI 11710. But it wasn't always that way.
Requirements for timing of flatness measurements
The first ACI tolerance document, ACI 11781, stated, "Floor tolerance measurements should be made the day after a concrete floor is finished and before shoring is removed, in order to eliminate any effects of shrinkage, curling, and deflection."
The first ACI tolerance document, ACI 11781, stated, "Floor tolerance measurements should be made the day after a concrete floor is finished and before shoring is removed, in order to eliminate any effects of shrinkage, curling, and deflection."
When ACI 11790 was published, however, Section 4.5 included no time requirement for measuring F_{F}
floor flatness, and the Commentary gave the following rationale: "Since
neither deflection nor curling will significantly change a floor's F_{F} value, there is no time limit on the measurement of this characteristic."
The statement in the Commentary indicating that "neither deflection nor curling will significantly change a floor's F_{F} value" has since been shown to be incorrect. This is why ACI 11706 and ACI 11710 require that all measurements – F_{F}, F_{L},
and the gap under a straightedge – be made within 72 hours. However,
not all specifiers and contractors are aware of this change. The next
two sections explore the effects of deflection on Fnumbers and why the
change was made.
Measured Fnumbers on deflected elevated slabs
In an article titled "Floor Tolerances," published in Concrete International in June 1988, Eldon Tipping reported floor flatness and levelness measurements made on castinplace concrete structures using widemodule panjoist construction. The depth of the total framing system was 20.5 inches, which included the 4.5 in. floor slab thickness. The joists were 7 in. wide and spaced at 5 feet on center. Slab profiles were measured using the Fnumber system for shored and reshored conditions, and after the shores had been removed.
In an article titled "Floor Tolerances," published in Concrete International in June 1988, Eldon Tipping reported floor flatness and levelness measurements made on castinplace concrete structures using widemodule panjoist construction. The depth of the total framing system was 20.5 inches, which included the 4.5 in. floor slab thickness. The joists were 7 in. wide and spaced at 5 feet on center. Slab profiles were measured using the Fnumber system for shored and reshored conditions, and after the shores had been removed.
The composite flatness Fnumber values before and after shore removal were F_{F}
24.4 and 24.3, respectively. Thus, the composite flatness Fnumber
dropped less than 1 percent after shore removal. This led to the
statement that, "There was a negligible change in the flatness Fnumber
between the shored and shores removed condition."
A second article coauthored by Tipping and K.S. Rajagopalan in
Concrete International, titled "Flatness and Levelness of Elevated
Surfaces," reported floor flatness and levelness measurements made on a
similar castinplace concrete structures but with 6in.wide joists
spaced at 6 ft. Again, the shored and unshored F_{F} numbers
were nearly identical at 26.0 and 25.7, respectively. The authors'
conclusion was that, "The lack of change in these values from the
supported condition confirms the assumption of Fnumber theory that
deflections have a minimal impact on flatness Fnumbers."
In these two reported studies, the assumption that deflections have a
minimal impact on flatness Fnumbers was confirmed primarily because
the buildings were stiff. For one building, the initial camber at
midspan of the 44footlong beam was 0.625 in. and for the 38footlong
beam the camber was 0.25 in. The cambers for these spans represent
deadload deflections of L/845 and L/1824, indicating a very stiff floor
system.
No cambers were required on the other building, but deflections were
observed from the measured surface profiles. The engineer of record
calculated the total initial deflection of about 1/2 in. at the midpoint
of the larger bays. This agreed well with the measured deflection of
about 0.4 in. based on surface profiles. The largest bay on this project
was 39 ft. long. This corresponds to a deflection magnitude of L/1170,
which again indicates a very stiff framing system.
These two castinplace concrete buildings were very stiff as
compared with buildings incorporating concrete slabs on metal decking
supported by a structural steel frame (see the analysis to follow). But
the conclusion that F_{F} is not affected significantly by dead
load deflections has been applied to all building types. The problem
with this can be illustrated by using the analysis method described
below.
Effect of deflection on Fnumbers of elevated slabs
Section 4.5.5 of ACI 117R90 stated that, "neither deflection nor curling will significantly change a floor's F_{F} value." No references in support of this claim were cited, although the claim may have been based on the two studies described previously. To predict what would happen to F_{F} numbers if shores were removed and dead load deflection occurred before Fnumbers were measured, we used the mathematical analysis described below.
Section 4.5.5 of ACI 117R90 stated that, "neither deflection nor curling will significantly change a floor's F_{F} value." No references in support of this claim were cited, although the claim may have been based on the two studies described previously. To predict what would happen to F_{F} numbers if shores were removed and dead load deflection occurred before Fnumbers were measured, we used the mathematical analysis described below.
First, we simulated initial Fnumber profiles representing varying
floor quality, then superimposed structural deflection values on the
profiles. The deflection was assumed to vary with position along the
beam, with the initial deflection equal to L/360, L/480, and L/960
– deflection values typically used in building code requirements, where
L is the length of the span. The deflections were calculated at 1 ft.
increments along the beam and added to the simulated Fnumber readings
at the same increment. A 30foot span was assumed because it is a common
bay size for steel buildings. Table 1 shows the results of this
simulation.


Table 1 
The analysis shows that for a stiff structure and an elevated slab with an F_{F} 25 value, a deflection of L/960 (3/8 in. for a 30foot span) decreases F_{F} by only 4 percent. Even for an initial profile representing an F_{F} 30 floor, a deflection of L/960 affects the F_{F} value by only about 7 percent. But for an initial F_{F} value of 50, the L/960 deflection causes about a 24 percent decrease in F_{F}. Thus, the higher the initial F_{F} value, the greater the effect of dead load deflection. A composite overall flatness of F_{F}
35 is the maximum specified value typically used for suspended slabs
(ACI 302.1R04). Based on the analysis and at this specified value, a
deflection of L/960 – which indicates a stiff building – will probably result in a reduction in F_{F} no greater than about 10 percent. Unfortunately, the same cannot be said for deflection values of L/480 and L/360,
which are common for structural steel framing systems with concrete
slabs placed on metal decking. Because these slabs deflect much more
than floors in reinforced concrete frame buildings, the effect on F_{F} can also be expected to be greater.
Observing the time limit on Fnumber measurement is essential
To ensure that Fnumber measurements reflect only the floor flatness and levelness resulting from the concrete contractor's efforts, these measurements must be made within 72 hours after completion of slab concrete finishing operations and before removal of any supporting shores. Otherwise, low Fnumbers are likely to be the result of deadload slab deflection. For slabs placed on unshored metal decking, deflection occurs while the slab is being placed, and may make it diF_{F}icult to achieve flatness values above F_{F} 25 without the use of specialized equipment and elevation monitoring during slab placement. In such slabs, flatness is improved by adding more concrete to compensate for the deflection.
To ensure that Fnumber measurements reflect only the floor flatness and levelness resulting from the concrete contractor's efforts, these measurements must be made within 72 hours after completion of slab concrete finishing operations and before removal of any supporting shores. Otherwise, low Fnumbers are likely to be the result of deadload slab deflection. For slabs placed on unshored metal decking, deflection occurs while the slab is being placed, and may make it diF_{F}icult to achieve flatness values above F_{F} 25 without the use of specialized equipment and elevation monitoring during slab placement. In such slabs, flatness is improved by adding more concrete to compensate for the deflection.
If floor flatness is measured months or even years after the building
has been in service, Fnumbers will be even more adversely affected
because of longterm deflection resulting from creep.
More information about deflection analyses is contained in
Tolerances for CastinPlace Concrete Buildings, a book written by the
authors of this article and published by the American Society of
Concrete Contractors (www.ascconline.org).