Nine Hundred Millimetres, Already Spent
The grid on this one was settled two years before anybody drew a slab. Parking below wants 8.4 m centres in both directions, the consent fixes 3.6 m floor to floor over eight storeys, and the occupier's brief asks for 2.7 m clear beneath the ceiling. That leaves 900 mm between one finished floor and the underside of the one above it, and the 900 is already spoken for: 150 for the access floor, 400 for a 300 mm duct with its hangers and a luminaire under it, and whatever the slab needs. The slab is being asked to fit inside 350 mm before a single load has been applied to it.
A reinforced flat slab on that grid ends up near 300 mm once punching shear at the columns and long-term deflection have both had a say. It fits, technically, with nothing left over for a downstand, a transfer, a tolerance or a change of heart about the duct route. Stress the same floor and the depth moves toward 210 mm. Ninety millimetres does not sound like a structural decision until it is multiplied by eight storeys, at which point it is 720 mm of building height, or one more coordination meeting that never has to happen.
So there are two questions on the table, and they are asked in this order. What is the shallowest depth that can be defended without committing to an explicit deflection calculation this early — and if that number is too deep for the zone, is this span long enough that post-tensioning still leaves anything behind after the tendon gives back what it gives back the moment the wedges bite? The second question has an unforgiving answer on short spans, and it is routinely skipped until the specialist's scheme comes back with more strand than anyone budgeted.
What a stressed floor slab is made of
- Finish and levelling layer — adds dead load the slab was sized for and takes up the camber a stressed floor arrives with, so it is thicker at mid-span than the section drawing implies Self-Leveling Underlayment (SLU) Volume Calculator
- Top reinforcement mat — carries hogging over the supports and holds the crack widths there; in a stressed floor it is also the bonded steel the code asks for regardless of prestress Concrete Slab + Rebar Combined Mass Calculator
- Tendons on their drape — high over the columns, low at mid-span, so the curvature pushes upward against gravity — the drape height is what the balanced load is proportional to Post-Tensioned Anchor Seating Loss Calculator
- Bottom reinforcement mat — sagging steel at mid-span, and the layer whose cover is measured off the soffit rather than off anything the tendon chairs are set to Rebar Calculator
- Slab concrete — the depth every other decision on this page is competing for, and the one dimension that is free to change until the services layout is signed One-Way Slab Minimum Thickness Calculator (ACI 318)
- Column head — where two-way shear concentrates and where a flat slab is most often forced deeper than bending alone would have asked for Reinforced Concrete Tied Column Axial Capacity Calculator (ACI 318)
The Ratio That Excuses You From Calculating
ACI 318 Table 7.3.1.1 is not a design method. It is a permission slip: hold a solid one-way slab at or above the thickness in that table and the code lets you skip the explicit deflection calculation altogether, on the argument that a member of those proportions has never been the one that sags visibly. Everything about it — the fact that load appears nowhere in it, the coarseness of four divisors for the entire universe of one-way slabs — follows from that being its job. It is a screening tool that happens to be enforceable.
The divisors are span over 20 simply supported, span over 24 with one end continuous, span over 28 with both ends continuous, and span over 10 for a cantilever. Span there is the clear span, face to face of supports, not the grid dimension. On a 400 mm wide band beam the difference between the two is 400 mm of span, which on an 8.4 m grid is a 14 mm change in the answer — small, but it is the sort of small that gets absorbed into somebody's rounding and then quietly re-appears as a fabrication query.
Continuity is the cheapest lever on the sheet, and it is a lever that exists at scheme stage and nowhere later. Going from simply supported to continuous at both ends moves the divisor from 20 to 28, which is a 29 per cent reduction in required depth for the identical span — put the other way round, a simply supported slab is 40 per cent deeper than the continuous one beside it — because a continuous slab hogs over its supports and the sagging moment at mid-span drops accordingly. That is why an edge bay, continuous at one end only, is the bay that sets the floor thickness on most buildings — and why moving a column line inboard by half a metre to make an edge bay behave like an internal one is worth more than any amount of arguing about concrete grade.
The cantilever row catches people because the length being divided is not the length they are thinking of. It is the projection from the face of the support to the free edge, and a 2 m balcony at span over 10 asks for 200 mm — against the 286 mm the 8.4 m internal bay it grows out of needs, on under a quarter of the span — and it overtakes that bay the moment the internal clear span drops below 5.6 m. Nothing has gone wrong; a cantilever has no far support to share the work with, and the table is being blunt about it.
The conditions attached to the table are the part that gets read once and then forgotten. It applies to solid one-way slabs of normalweight concrete reinforced with Grade 60 bar, that are not supporting or attached to partitions or other construction likely to be damaged by large deflections. Change any of those and the number moves: higher-grade reinforcement and lightweight concrete both carry modification factors in ACI 318, and the partitions exclusion is not a footnote — a full-height blockwork line landing mid-panel on a slab sized straight off the table is precisely the case the table declines to cover.
Outside the United States the same permission exists by a different route and produces a different number. EN 1992-1-1, Eurocode 2, controls deflection through a limiting span-to-depth ratio that moves with the reinforcement ratio and the concrete class rather than through four fixed divisors, with the values completed by the National Annex; CSA A23.3 Design of Concrete Structures and AS 3600 Concrete Structures each carry their own minimum-thickness provisions. They were never written to agree, so a depth taken from the divisors above is an ACI depth and is worth labelling as one on any drawing that crosses a border.
Put the clear span and the real support condition in and read the depth the code will accept without argument — then check it against the zone you actually have, because that comparison is the whole scheme-stage decision and it takes ten seconds.
The slab's clear span length.
How the slab is supported at each end.
Minimum slab thickness
0.696 ft
This is the code's deflection-exempt minimum thickness for normal-weight concrete with Grade 60 rebar — it doesn't replace a strength (moment/shear) design, and lightweight concrete or other rebar grades require an adjustment factor per ACI 318.
- Equivalent in inches
- 8.36 in
- Span/depth divisor used
- 28
They open the calculator with your figures already in it
One-Way Slab Minimum Thickness Calculator (ACI 318): 0.6964 ft — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
What this calculation does not cover
- This is a deflection-control ratio, not a design. It says nothing about whether the slab can carry its load: bending and shear capacity, reinforcement area, bar spacing and cover are separate calculations, and on a heavily loaded floor it is strength that ends up setting the depth, not this ratio. Fire-resistance rating and the depth needed to bury conduit or drainage can also demand more than the figure here.
- The ACI table behind this applies only to slabs that do not support, and are not attached to, partitions or other construction likely to be damaged by large deflections. A full-height block wall landing mid-panel, a rigid partition head, bonded stone or glazing on this slab puts you outside the table, and the explicit deflection check it exists to excuse becomes mandatory again.
- The divisors are written for solid, nonprestressed one-way slabs in normal-weight concrete with Grade 60 reinforcement, and nothing is adjusted when you fall outside that. Lightweight concrete and other bar grades carry modification factors in ACI 318 that this calculator does not apply, and post-tensioned slabs, two-way slabs and flat plates, ribbed and waffle floors, and composite metal-deck slabs are not covered by the table at all.
- It sizes one span with one support condition. On a continuous floor the deepest bay sets the whole slab, and this will not find that bay for you — the edge bay, continuous at one end only, is usually the one that governs. The answer is a bare minimum rather than a pour dimension, so round it up to a thickness you would actually build.
- This is ACI 318's number. Eurocode 2, CSA A23.3 and AS 3600 each control slab deflection through their own provisions and will not produce the same depth, so a thickness taken from here is an ACI figure and is worth labelling as one on any drawing that crosses a border.
Four Divisors, and Everything They Do Not Know
The table is silent on load, and that silence is deliberate rather than an omission. A 300 mm slab carrying a plant room and a 300 mm slab carrying an open office satisfy Table 7.3.1.1 identically, and only one of them will pass a strength check. Depth from the ratio is a starting point for the drawing, not a design; the moment and shear capacity at the chosen depth still has to be demonstrated, and on heavily loaded floors it is strength, not deflection, that ends up governing.
The table is also silent about time. Concrete creeps under sustained load and shrinks whether it is loaded or not, and the deflection a floor eventually settles at is a multiple of the one it shows at striking — the sustained-load multipliers in ACI 318 and the guidance in Concrete Society Technical Report TR58, Deflections in Concrete Slabs and Beams, both exist because the elastic answer is not the answer anyone ever complains about. Where the finish is brittle — bonded stone, a rigid partition head, a lift landing threshold — it is the long-term increment after the finish went on that does the damage, not the total, and the limits it is judged against sit in IBC Table 1604.3 Deflection Limits rather than in any thickness table.
And the table stops applying entirely the moment the floor is prestressed. ACI 318's minimum-thickness tables are written for nonprestressed members; a post-tensioned slab has no code shortcut and owes an explicit serviceability calculation from the outset. This is the trap in the height budget above: the ratio was used to reject 300 mm, the answer 'go PT' was reached, and the number 210 mm that replaced it did not come from any table at all. It came from what PT specialists habitually work to on that grid, which is a scheme assumption to be tested rather than a value to be relied on.
| Floor type | Where the first depth comes from | What usually ends up governing it |
|---|---|---|
| Solid one-way slab, on walls or beams | ACI 318 Table 7.3.1.1 — span/20, /24, /28 or /10 | Nothing, if the table was applied to the clear span and its conditions hold |
| One-way slab on wide band beams | The slab spans beam to beam; the band takes the long direction | Band beam depth against the services zone, not the slab |
| Flat plate, no drops, no column heads | ACI 318's minimum-thickness provisions for two-way slabs without interior beams | Two-way shear at the column faces, then long-term deflection |
| Flat slab with drop panels or column capitals | As above, with the thickened zone credited at the columns | The panel dimensions and depth the shear check demands |
| Post-tensioned flat plate | A specialist's scheme ratio — no code table applies to a prestressed slab | Explicit deflection and stress checks, plus punching at the columns |
| Post-tensioned band beam and one-way slab | Band ratio and slab ratio taken separately, tendons banded one way | Tendon layout, anchorage zone geometry and stressing access at the edge |
When a One-Way Slab Stops Being One
A slab supported on four sides distributes its load in both directions, and the proportion going each way follows the stiffness, which follows span to the fourth power. Once one side is roughly twice the other, so little of the load takes the long route that treating the panel as spanning one way is both true enough and conservative — which is where the customary 2:1 aspect threshold comes from. On the 8.4 by 8.4 grid at the top of this page nothing about it is one-way, and applying a one-way divisor there is not a conservative simplification; it is the wrong table.
The distinction is worth holding onto because it decides which calculator is honest. A one-way divisor answers a real question about a real slab type — a slab on band beams, a slab between walls, a stair landing, a canopy — and it answers it in seconds. On a flat plate the same arithmetic gives a plausible number for the wrong mechanism, and the check that actually sets the depth is two-way shear on a perimeter a short distance out from the column face, where a 250 mm plate on a 400 mm column is being asked to carry an entire bay's load through a band of concrete a few hundred millimetres wide. That check has no ratio and no shortcut, and this site does not publish one, because the perimeter geometry, the moment transferred into the column and the presence or absence of shear reinforcement all change the answer.
What the Strand Is Actually Buying
Post-tensioning a floor is load balancing before it is anything else. The tendon is pulled to a high force and held on a drape — high over the columns, low at mid-span — and that curvature turns the tendon force into a distributed upward pressure on the concrete. Set the drape and the force so the upward pressure cancels most of the permanent load, and the slab spends its service life carrying only what is left, which is why a stressed floor can be shallow without being lively. The design lever is the drape height, and drape height is limited by slab depth and cover, so the shallower the slab the harder each strand has to work.
The second thing it buys is a compressed section. Axial precompression closes flexural cracks that a reinforced slab simply lives with, which matters for water tightness, for stiffness under service load, and for the effective section used in any deflection calculation. It is also the reason a stressed floor is a restrained floor: precompression only exists if the slab is allowed to shorten. Stiff cores, long floor plates and shear walls at both ends of a bay all take prestress out of the concrete and put it into the walls, and the fix — pour strips left open for weeks, delayed closure details, releasing one end of a stiff element — is a programme decision made at scheme stage or not at all.
The third thing is depth, and depth is why anyone on a tight height budget starts the conversation. That saving compounds: a shallower floor means a shorter column, a lighter building, smaller foundations and less concrete per square metre, which the embodied carbon assessment notices. Set against it is a specialist package, an anchorage detail at every stressing edge, a stressing sequence the programme must accommodate, and a slab that can never afterwards be cored, chased or drilled without a tendon survey.
The thing it does not buy is proportional benefit on a short span. Prestress is delivered by strain, the strand has to be stretched to be useful, and there are fixed lengths of stretch that come straight back off the top irrespective of how long the tendon is. On a long tendon those fixed losses disappear into a large elongation; on a short one they are a large fraction of everything applied. That is the mechanism behind the trade instinct that PT stops paying below a certain span, and it deserves better than an instinct.
Six Millimetres of Wedge Travel
Watch a monostrand anchorage as the jack releases. The strand is at full jacking force, the wedges are sitting loose in the tapered casting, and as the load transfers the wedges are dragged into the taper until they lock. That travel is a few millimetres, the strand follows it, and the tendon shortens by exactly that distance. Anchor set, wedge draw-in, seating loss: three names for the same few millimetres. Typical values run from about a quarter of an inch to three eighths — roughly 6 to 10 mm — and the number belongs to the anchorage system, not to the job, so it comes off the manufacturer's data sheet and nowhere else.
The arithmetic that follows is the whole point. A fixed shortening spread over a tendon length is a strain, and strain times the elastic modulus of the strand is a stress. Seven-wire prestressing strand carries a modulus of about 28,500,000 psi in the Post-Tensioning Institute's Post-Tensioning Manual — near enough 196.5 GPa — so a 6.35 mm set on a 30 m tendon costs roughly 42 MPa, while the same 6.35 mm on a 6 m tendon costs about 208 MPa. Nothing changed except the denominator. That is why short tendons are inefficient, and it is a far more concrete statement than 'PT does not work on short spans'.
Put those losses against what was applied. ACI 318 caps the stress in the strand at jacking and again at the anchorage immediately after transfer, both as fractions of the specified tensile strength; for the Grade 270 strand of ASTM A416 that is a jacking stress in the region of 1,488 MPa. A 42 MPa seating loss is under three per cent of it and disappears into the rest of the loss budget. A 208 MPa seating loss is fourteen per cent, taken off the top before friction, elastic shortening, creep, shrinkage and relaxation have taken anything at all, and a scheme costed on nominal strand force will be short by roughly that much.
One caveat has to travel with the arithmetic, because the simple ratio is only true in the case it was built for. Friction between strand and sheathing resists the pull-back, so on a long tendon the seating effect dies out within a setback length and the rest of the tendon never feels it — the loss is concentrated near the live anchorage and the simple full-length figure both overstates how much of the tendon is affected and understates the local drop at the end. The method behind the calculator below is the short-tendon case, where the setback length exceeds the tendon and the whole thing shortens together, and that is exactly the case a designer worrying about short spans is in. Beyond roughly thirty metres, or wherever the specialist's friction calculation puts the setback, treat the number as a trend rather than a design value.
The scheme-stage answer this points to is not 'lengthen the span'. It is 'lengthen the tendon'. A tendon run continuous across three 8.4 m bays is a 25 m tendon, not an 8.4 m one, and the seating loss falls by two thirds because the denominator is the tendon, not the span. Continuity across bays, stressing from one end with the dead end placed where the loss matters least, an anchorage system with a smaller published set, and — where none of that is enough — power seating or shims: these are the levers, and every one of them is decided by how the tendon layout is drawn rather than by how much strand is bought.
| Tendon length | Stress lost to seating | Share of jacking stress | Force lost per 12.7 mm strand |
|---|---|---|---|
| 6 m | 208 MPa (30,200 psi) | about 14% | 20.5 kN |
| 8 m | 156 MPa (22,600 psi) | about 10% | 15.4 kN |
| 10 m | 125 MPa (18,100 psi) | about 8% | 12.3 kN |
| 15 m | 83 MPa (12,100 psi) | about 6% | 8.2 kN |
| 20 m | 62 MPa (9,000 psi) | about 4% | 6.2 kN |
| 30 m | 42 MPa (6,000 psi) | about 3% | 4.1 kN |
Run your own tendon length and your anchorage's published set against the strand you are assuming, then set the loss it returns against a jacking stress near 1,488 MPa — that fraction, not the span, is what tells you whether this floor is worth stressing.
The length of the post-tensioned tendon.
The seating loss distance for your specific anchorage system.
The nominal cross-sectional area of one strand.
Stress loss from anchor seating
12,100 psi
This simplified short-tendon method assumes the seating loss propagates the full tendon length — a licensed PT designer must verify this assumption and account for friction losses along the tendon for a final design.
- Stress loss (imperial)
- 12,117.35 psi
- Force loss per strand
- 1.85 kips
- Force loss per strand (imperial)
- 1,853.78 lbs
They open the calculator with your figures already in it
Post-Tensioned Anchor Seating Loss Calculator: 12,117 psi — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 12,100 psi — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- Friction is not in this model. It spreads the wedge draw-in uniformly over the entire tendon, ignoring the strand-to-duct or strand-to-sheathing friction that in a real tendon confines seating loss to a set-influence length near the stressing end. Where friction matters, the actual loss at the anchorage is higher than this figure and falls away to nothing further along the tendon.
- This is the seating loss alone, not the effective prestress. Elastic shortening of the concrete, creep, shrinkage, steel relaxation, and wobble and curvature friction along the tendon profile are all excluded, as is any set at the dead end.
- The strand modulus is fixed at the standard 7-wire value and cannot be changed. The result does not apply to threadbar or other bar systems, or to strand of a different grade or modulus. The force figure is per single strand and is not multiplied up for the strands in a tendon or the tendons in a slab.
- Nothing checks that the computed loss is physically achievable. The calculator never compares it against your jacking stress or the strand's breaking strength, so a very short tendon length returns a loss larger than the prestress the tendon could ever have held, with no warning on the page.
- This is not a post-tensioning design and does not replace one. Anchor set must come from the anchorage supplier's data for the system actually being installed, and the stressing sequence, tendon profile, jacking force and final effective prestress are a licensed PT designer's work.
What Has to Be True Before a Jack Turns Up
Everything above is a scheme decision, and scheme decisions are cashed on site by people who were not in the room. Two of them decide whether the floor gets the prestress the calculation assumed. The first is concrete strength at stressing: transferring a large anchorage force into young concrete splits it, so the specification names a compressive strength that must be demonstrated before any jack is connected — a requirement most project documents inherit from ACI 301 Specifications for Structural Concrete — verified on cylinders to ASTM C39 Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens, or on cubes to EN 12390-3, cured alongside the pour rather than in a laboratory. Field-cured specimens are the honest ones here, because the question is what the slab has reached, not what the mix is capable of.
The second is the drape. The upward pressure the whole design rests on is proportional to the height between the tendon's high and low points, and on a 210 mm slab that height is a hundred-odd millimetres between two covers. Lose 15 mm at the low point because a chair settled or a bar crew walked the tendon flat, and a meaningful fraction of the balanced load walks out with it. ACI 117 Specification for Tolerances for Concrete Construction and Materials sets the placing tolerances; ACI 423.7 Specification for Unbonded Single-Strand Tendon Materials governs what the tendon itself has to be. Surveying tendon heights before the pour is the cheapest quality assurance on a stressed floor and the first thing dropped when the programme slips.
Striking and backpropping are their own decision, and a stressed slab behaves unlike a reinforced one at that moment. It is designed to be supported by its own prestress rather than by props, so premature stressing and premature striking are two different mistakes with two different consequences, and the shoring and reshoring sequence in ACI 347 Guide to Formwork for Concrete is written on the assumption that somebody planned the load path through the floors beneath. A stressed floor also lifts. Camber at mid-span is real, it is intended, and it has to be allowed for in the levelling layer above and in every threshold, façade bracket and lift guide that meets the edge of it.
- Fix the clear span, face to face of supports, and the real continuity condition at each end before quoting any depth.
- Take the ratio depth first, and compare it against the zone left after the services and floor build-up have taken theirs.
- If the ratio depth does not fit, price the tendon before assuming post-tensioning will: length, not span, sets the seating loss.
- Draw the tendon layout continuous across as many bays as the pour sequence and stressing access allow, and count the length that produces.
- Get the anchorage system's published anchor set in writing, and rerun the loss with that number rather than a typical one.
- Hand the specialist the geometry, the restraint conditions, the strength required at stressing and the deflection limits the finishes impose — and let them own the losses you have only estimated.
The Depth Is Never Only the Slab's Business
A slab depth agreed at scheme stage is a commitment made on behalf of five other packages. The façade contractor sets bracket positions off the slab edge and the edge thickness; the mechanical layout depends on the void the depth left behind; the core walls take the restraint a stressed floor imposes on them; the foundations carry the mass; and the programme absorbs whatever the striking and stressing sequence costs. None of those parties get consulted when the depth is picked, and all of them get a bill if it changes later.
That asymmetry is the argument for doing the arithmetic on the day the grid arrives rather than the week the package goes out. The two numbers on this page take a minute between them: the code depth for a one-way slab of that span and continuity, and the seating loss for the tendon length that layout would actually produce. Neither is a design. Both are enough to tell you which of the two schemes is worth developing, which is the only question anybody is really asking at that stage — and both are cheap enough that there is no defence for guessing instead.
Settle these before the depth goes on a drawing
Five things that are free while the grid is still a sketch and expensive once the façade and the services have been set out against them.
- Clear span and continuity at each end — Face to face of supports, not the grid line. Continuous at both ends is 29 per cent shallower than simply supported over the identical span.
- The zone genuinely left for structure — Floor to floor, less the clear height, less the access floor and the deepest service run with its hangers. That remainder is the brief.
- Whether the panel is one-way at all — Roughly 2:1 or longer and a one-way divisor is honest. On a square bay it answers the wrong mechanism, and punching at the columns governs.
- Tendon length, counted along the layout — Continuity across three bays makes a 25 m tendon out of an 8.4 m span, and the seating loss falls with the denominator.
- The anchorage's published anchor set — Six to ten millimetres depending on the system, taken from the manufacturer's data sheet rather than from a typical value.
Opens the calculators above on one screen with the dimensions from this article already filled in. Quantities only — this site publishes no price list, because local prices vary too much to publish honestly.