FF vs. FL: Why a Single “Flat and Level” Clause Isn’t a Specification

FF (floor flatness) measures local curvature over short intervals, typically 12 inches, while FL (floor levelness) measures broader elevation change over longer 10-foot spans. Both are calculated from elevation data collected under ASTM E1155/E1155M, and a spec needs both numbers, not a general phrase like “flat and level,” to be enforceable.

Architects write “provide a flat and level floor slab” more often than you’d expect on projects where the racking supplier, the robotics integrator, and the concrete contractor all have different opinions about what “flat” means. That single clause creates the dispute, because nobody defined the test method, the acceptance number, or who’s responsible for measuring it. When the floor fails to perform, three parties point at each other and none of them can point at a document.

FF and FL are not interchangeable, and they don’t measure the same thing. FF captures bumpiness, the kind of localized waviness that makes a forklift operator feel every joint. FL captures gradual tilt or dish across a bay, the kind of long-wave variation that causes water to pool or racking to lean slightly out of plumb over 40 feet. The Floor Covering Reference Manual’s AA4 guidance draws this same distinction and notes that FF/FL evaluation under ASTM E1155 is a separate methodology from ASTM E1486, which addresses waviness and wheel-path criteria for different applications. Mixing those methods, or substituting one for the other, produces numbers that don’t mean what the spec writer thinks they mean.

This is also where straightedge language gets people into trouble. A 1/4-inch-in-10-foot straightedge tolerance sounds precise, and it’s been used in concrete specs for decades. But a 2024 issue of ACI’s Concrete International demonstrated that a 1/4-inch variation under a 10-foot straightedge can correspond to a SOFF value ranging from roughly 24.0 to 45.9, spanning “conventional” all the way to “flat” classifications. A straightedge tolerance and an F-number are not the same acceptance criterion, and specifying both in the same clause doesn’t add precision. It creates a floor that can pass one test and fail the other.

The Numbers That Belong in Every FF/FL Spec: SOFF, SOFL, MLFF, MLFL

A defensible spec sets four numbers: SOFF and SOFL (specified overall values for the whole slab or test area) and MLFF and MLFL (minimum local values that no individual test section may fall below). Per ACI’s Floor Flatness Report, MLFF and MLFL default to three-fifths of the specified overall values unless the project documents state otherwise.

SOFF and SOFL describe the average condition of the slab, calculated from all the sample measurements taken across the test area under ASTM E1155/E1155M. They tell you whether the floor, on the whole, meets the classification you asked for. MLFF and MLFL exist because an average can hide a problem: a slab could post an excellent overall FF number while one corner sits dished or cratered enough to cause real trouble under a narrow-aisle lift or a sensitive piece of equipment. The minimum local value catches that outlier.

Leaving MLFF/MLFL out of a spec is one of the most common gaps IFTI sees in project manuals. It’s an easy fix, and it closes off the argument that a “passing average” is the same thing as a passing floor.

Matching the Classification to the Floor’s Actual Use (Not Its Marketing Label)

Floor classifications like “flat” or “super flat” correspond to specific SOFF/SOFL ranges, and the classification you specify should be driven by what the floor actually needs to support, not by a term a contractor or supplier suggests. ACI 302.1R connects specified flatness to slab classification, finishing method, and construction sequencing.

The table below reflects representative values published by the American Society of Concrete Contractors in ACI’s Concrete International (May 2022). These are industry reference points, not a substitute for project-specific engineering judgment, and the current edition of ACI 117 should be consulted directly and incorporated by reference for the governing tolerance provisions.

A distribution center with standard pallet racking on 42-inch aisles rarely needs “super flat” numbers, and specifying them anyway adds cost and construction risk without a corresponding benefit. Conversely, a manufacturing floor carrying a coordinate-measuring machine or a precision assembly line may need tighter local values than the “flat” classification provides. The classification should follow from a conversation with the equipment supplier, the structural engineer, and, where robotics or automated storage are involved, the integrator, not from whichever label sounds most impressive in the project manual.

Random-Traffic vs. Defined-Traffic Floors: Why Narrow-Aisle and Robotic Floors Need a Different Clause

Overall FF/FL numbers assume forklifts and foot traffic move randomly across the slab, averaging out local irregularities. Narrow-aisle, very-narrow-aisle (VNA), wire-guided, and robotic (AMR/AS-RS) floors don’t work that way: equipment repeats the same wheel path thousands of times, so a defined-traffic spec needs longitudinal and transverse tolerances measured along those specific paths, not a facility-wide average.

This is the single most consequential distinction in the whole article, and it’s the one architects miss most often. The Concrete Society’s guidance on floor surface regularity is explicit that surface-regularity requirements depend on the intended application, and for industrial floors carrying defined-traffic equipment, it points directly to Technical Report 34 (Concrete Industrial Ground Floors). TR34 addresses the tolerances that matter to a guided vehicle or a mast that travels the same six-inch-wide corridor thousands of times a day: elevation change along the direction of travel, elevation change across the aisle width, and joint conditions at specific stations, rather than a global SOFF/SOFL average across the whole warehouse floor.

A “flat” or even “super flat” overall classification can still contain a localized dip precisely where a wire-guided vehicle’s wheel tracks, because that dip may not move the facility-wide average enough to fail the SOFF number. Meanwhile, the vehicle itself may bind, drift off guidance, or trigger a fault at that exact spot. If your project has any narrow-aisle racking, robotic picking, AS/RS, or fixed-guideway equipment, the spec needs a TR34-based defined-traffic provision for those specific aisles or zones, layered on top of (not instead of) the random-traffic FF/FL clause for the rest of the slab.

When and How Testing Must Happen, and Who’s Responsible

ASTM E1155/E1155M governs both the timing and the method of measurement: testing must occur on the newly finished slab before it’s opened to construction traffic, following the test-area layout, sample-line spacing, and elevation-measurement procedure the standard specifies. The spec should name who schedules and pays for that testing, and who’s on the hook for corrective action if a test section fails.

Vague timing language is a common source of disputes. If the spec doesn’t say testing happens within a defined window after final finishing and before the floor is loaded, trafficked, or covered, a contractor can argue the results don’t reflect the “as-delivered” slab, and an owner can argue the opposite. The spec should also state explicitly who selects the test areas, whether that’s every placement, a percentage sample, or the full floor, and whether the general contractor, the concrete subcontractor, the owner, or an independent third party is responsible for retaining a qualified testing agency.

Independent testing matters here specifically because the party that placed the concrete has an obvious interest in the results. A test report generated by the placing contractor’s own crew, using uncalibrated or unfamiliar equipment, doesn’t carry the same weight as a report from a third party working strictly to ASTM E1155/E1155M with documented equipment calibration. Per ACI’s Floor Flatness Report, a defensible test report should include, at minimum: project and slab identification; date and time of testing relative to placement; the boundaries and layout of the test area and sample lines; individual and overall FF/FL results; identification of any sections falling below MLFF/MLFL; equipment type and calibration status; and any conditions that limited or affected the measurement.

Related to timing: floors with moisture-sensitive flooring systems (resinous coatings, adhered vinyl, certain coatings over slab-on-grade) often need concrete moisture testing under ASTM F2170 or ASTM F1869 on a similar schedule to flatness testing, since both affect when the floor can safely be finished. Coordinating both tests with the same independent party avoids scheduling gaps between when the slab is flat enough to test and when it’s dry enough to finish.

Sample Specification Language: Random-Traffic Floor (Adaptable Block)

The following is adaptable spec language for a standard random-traffic slab. Every numeric value must be revised to match the actual classification, equipment, and use of your specific floor; nothing here should be copied verbatim without engineering review.

Sample Specification Language: Defined-Traffic Floor Areas (Adaptable Block)

Use this block for narrow-aisle, VNA, wire-guided, or robotic zones, in addition to (not in place of) a random-traffic clause for the rest of the slab. Values, wheel-path locations, and tolerances must be coordinated with the racking or robotics supplier and revised for your project before use.

Common Specification Errors That Lead to Disputes and Re-Work

Most FF/FL disputes trace back to a small set of recurring gaps in the specification itself, not to unusually difficult concrete work. Watch for these when writing or reviewing a spec:

• Specifying “flat and level” or a straightedge tolerance without naming ASTM E1155/E1155M and numeric SOFF/SOFL values at all.
• Setting SOFF/SOFL without setting MLFF/MLFL, leaving no protection against a passing average that hides a failing local area.
• Combining a straightedge tolerance and an F-number requirement in the same clause, as if they measure the same thing.
• Choosing a classification (like “super flat”) because it sounds thorough, without confirming the equipment or use actually requires it.
• Omitting test timing, so testing happens after the floor has already been trafficked, loaded, or covered.
• Failing to name who selects test areas, who pays for testing, and who’s responsible for corrective work if a section fails.
• Applying only a random-traffic FF/FL clause to narrow-aisle, VNA, or robotic zones, with no TR34-based wheel-path provision.
• Citing ACI 117 or ASTM E1155 by name without incorporating the current edition by reference, leaving the acceptance criteria open to interpretation.

Fixing these gaps at the spec-writing stage costs nothing but attention. Fixing them after the slab is placed usually means grinding, shimming, or in the worst cases, partial re-pour, along with a schedule delay nobody budgeted for.