Codes referenced
Cylindrical specimen, constant head of 10 kg/cm², steady-state flow, coefficient of permeability in cm/s (Cl 8.1). A method, with no acceptance value of its own.
500 ± 50 kPa for 72 ± 2 h on a specimen at least 28 days old, then split and report the maximum depth of the water front in mm.
Design code for liquid-retaining work: low permeability through mix, curing, admixtures and crack control, plus site retention testing. No IS 3085 limit.
British ISAT — a surface absorption measurement on the face of the concrete, not the coefficient IS 3085 reports.
Darcy's-law methods for open-graded pervious concrete. A different material and a different test family — its numbers do not transfer to dense concrete.
Why permeability matters
Permeability is the master durability property: every long-term deterioration mechanism of concrete is driven by the ingress of water carrying dissolved species. Chloride ions, sulphates, CO₂, dissolved oxygen — all travel through the connected capillary pore network in the cover zone. A high-permeability concrete cannot be saved by chemistry at depth; corrosion, ASR, sulphate attack, freeze-thaw and carbonation all accelerate proportionally. The coefficient of permeability is the number that carries this into service-life calculation, and IS 3085 is the Indian method for measuring it — on a specimen cast in the laboratory or cut out as a core from an existing structure (Cl 1.1).
IS 3085 test setup — step by step
- Cast the specimen — cylindrical, with the height equal to the diameter. The standard size is 150 mm diameter and 150 mm high (Cl 5.1). Where the nominal aggregate does not exceed 20 mm, the diameter and height may be reduced to 100 mm; where the nominal aggregate exceeds 40 mm, they should be not less than about four times the nominal aggregate size.
- Compact and finish it properly (Cl 5.2) — cast in split moulds with a removable collar of about half the height, compact by hand rodding or vibration as it will be on site, then remove the collar and strike off level with a straight-edge using a sawing motion. No further floating or finishing, which would raise the fines to the surface.
- Cure for 28 days, unless the engineer-in-charge specifies otherwise (Cl 5.2).
- Prepare the test surface (Cl 7.8) — the surface has to be carefully prepared by sand blasting or chiselling, because even a thin highly impervious skin can result in considerable underestimation of the permeability.
- Set up the cell (Cl 3) — the permeability cell is a metal cylinder with a ledge at the bottom to retain the specimen, a flange at the top, a removable cover plate and a sheet-metal funnel, all bolted together with a rubber or neoprene gasket that makes the joint water-tight. The reservoir carries a graduated side-arm gauge glass, and the whole system is filled with de-aired water.
- Apply the pressure head (Cl 6.1) — 10 kg/cm² is the standard head. It may be reduced to 5 kg/cm² for relatively more permeable specimens, where steady state of flow is obtained in a reasonable time, and increased to 15 kg/cm² for relatively less permeable specimens, where sealing can be fully effective.
- Run the test (Cl 7.6) — record the initial gauge reading, place a clean weighed collection bottle under the cell to catch the percolate, and record the quantity of percolate and the gauge readings at intervals. Steady state of flow is the stage at which outflow and inflow first become equal.
- Hold it for about 100 hours (Cl 7.6) — the test continues for about 100 hours after steady state has been reached, and the outflow used is the average of all the outflows measured during that period.
- Calculate and report (Cl 8.1 and Cl 9.1) — the coefficient of permeability comes from K = (Q · H) / (A · T · L), and the report carries the mix, the age at commencement, the duration, the specimen size, the test pressure, the test temperature, the coefficient at test temperature and the coefficient corrected to the standard temperature.
Acceptance criteria — where the number actually comes from
Read IS 3085 from end to end and you will not find a single acceptance value, because there isn't one. It is a method, not a pass/fail test: Cl 8.1 gives you the formula for the coefficient of permeability and Cl 9.1 tells you what to put on the report. Where the limit comes from is somebody else's decision, and it is worth knowing which document that is before you sign off on a test certificate.
| Document | What it does | Sets a permeability limit? |
|---|---|---|
| IS 3085:1965 | Method: cylindrical specimen, constant head, coefficient of permeability in cm/s | No — method only |
| Project specification / ITP | States the value the designer will accept for that structure, and the method to be used | Yes — this is where the number belongs |
| IS 456 | Durability design: maximum water-cement ratio, minimum cement content and cover, by exposure class | No — it sets the inputs, not the test result |
| BS 8007 | Design code for concrete structures for the storage of liquids: low permeability through the mix, the curing, the admixtures and crack control, plus site retention testing | No IS 3085 limit |
BS 8007 is worth reading for water-retaining work, but for a different reason than a test limit. Its Cl 2.7.5 asks for low permeability and gets there through the recommendations on concrete mixes, aggregates, minimum cement content and strength, curing and admixtures, with complete compaction without segregation on site as the non-negotiable part. Section 9 then asks for the structure itself to be tested — a 7-day liquid retention test with the drop in level not exceeding 1/500th of the average water depth of the full tank, 10 mm or another specified amount, and a roof flooded to a minimum depth of 25 mm for 24 hours. The millimetre figures you will still meet quoted as “IS 3085 limits” on Indian test certificates are not in IS 3085 at all — treat any of them as a project-specific number and ask where it came from.
Comparison — IS 3085 vs DIN 1048 / EN 12390-8
Constant-head flow test. Cylindrical specimen 150 mm diameter and 150 mm high, cured 28 days, 10 kg/cm². Measures the water that comes through and reports a coefficient of permeability in cm/s.
Depth of penetration of water under pressure. Specimen at least 28 days old, 500 ± 50 kPa for 72 ± 2 h, then split and the maximum depth of the water front measured in mm.
The German national series that EN 12390-8 replaced. It survives in older specifications, so read the edition named in the contract rather than assuming it matches the EN.
The two look alike from a distance and are not comparable at all. IS 3085 answers “how fast does water move through this concrete?” in cm/s. EN 12390-8 answers “how far did the water get in 72 hours?” in mm. A depth in millimetres cannot be turned into a coefficient without the flow data behind it, and a split-face water front will never tell you the coefficient. When a contract names both — which happens on projects with European consultants — run both and report each against its own method reference; don't put the two numbers in the same column.
What controls the result — w/c, curing, SCMs
| Variable | Effect on the coefficient | What the standard says |
|---|---|---|
| Water-cement ratio | The single biggest lever. Less water means fewer capillary pores and a less connected path for the water to travel through | IS 456 caps the maximum water-cement ratio by exposure class |
| Curing | Decides whether the matrix reaches its design density at all. The pore structure keeps refining as long as there is water to hydrate with | Cl 5.2: 28 days, unless the engineer-in-charge specifies otherwise. IS 456 Cl 13.5.1: 7 days for ordinary concrete, 10 days with mineral admixtures or blended cements |
| Supplementary cementitious materials | Silica fume, GGBS and fly ash react with portlandite to form additional CSH gel that plugs the gateway pores — but they need the curing time to pay for it | The 28-day curing of Cl 5.2 is what gives them time to react |
| Surface preparation | A thin impermeable skin on the face makes the concrete look far less permeable than it is | Cl 7.8: prepare the surface by sand blasting or chiselling, or the result is an underestimation |
| Seal around the specimen | Water that leaks past the seal never goes through the specimen, so the measured flow is meaningless | Cl 7.8: the seal must be effective; leakage gives entirely misleading results |
| Dissolved air | Air locks in the pore network show up as an apparent reduction in permeability | Cl 7.8: air content of the water entering the specimen not to exceed about 0.2%; drain and refill with fresh de-aired water when it does |
| Evaporation from the bottle | Water lost from the collection bottle reads as flow that never happened | Cl 7.8: run a blank on a similar bottle containing water and apply the correction; the inflow measurement is the additional check |
| Test temperature | About 10% on the coefficient for every 5 °C above the standard temperature, and the same reduction below it | Cl 7.7: 27 ± 2 °C preferred; otherwise record the actual temperature and correct |
Each step is dominant in turn, but only the first three are about the concrete. The rest are about the test: they decide whether the number on the certificate describes your mix or describes the rig. We have seen perfectly good concrete come back as marginal because the specimen face was float-finished, and we have seen a leaking seal reported as a comfortable pass. Neither is a mix problem, and both are invisible in the result.
Pressure, temperature and the 100-hour run
Three procedural details decide whether the number means what it is supposed to mean:
- The head is a choice, not a constant — Cl 6.1 sets 10 kg/cm² as the standard, allows 5 kg/cm² for relatively more permeable specimens where steady state of flow is obtained in a reasonable time, and allows 15 kg/cm² for relatively less permeable specimens where the sealing can be fully effective. The pressure used belongs on the report (Cl 9.1), because the coefficient depends on it.
- Temperature moves the answer — Cl 7.7 asks for 27 ± 2 °C and gives the correction: each 5 °C above the standard temperature raises the coefficient by about 10%, and the same reduction applies below it. A test run at 32 °C that is not corrected reads about 10% high.
- Steady state has to be reached — Cl 7.8 is blunt about it: the flow must be permitted to attain steady state before the coefficient is calculated, and the inflow and outflow rate data, or graphs of them, are how you show that it did. Read the outflow while it is still climbing, and the coefficient comes out too low.
Source: IS 3085:1965 Cl 6.1 (pressure head), Cl 7.6 (running the test, steady state and the 100-hour period), Cl 7.7 (test temperature), Cl 7.8 (precautions).
Worked example — one 150 mm cylinder, 10 kg/cm², 100 hours
| Specimen | 150 mm diameter × 150 mm high, cured 28 days |
| Face area A | π × 7.5² = 176.7 cm² |
| Thickness L | 15 cm |
| Pressure head H | 10 kg/cm² = 1000 cm of water |
| Period T | 100 h of steady-state averaging = 360,000 s |
| Percolate Q | 10.0 ml collected in that period |
Cl 8.1 gives the calculation:
| K | = (10.0 × 1000) / (176.7 × 360,000 × 15) |
| = 10,000 / 954,180,000 | |
| K | = 1.05 × 10⁻⁸ cm/s |
A coefficient in the 10⁻⁸ cm/s range is what a well-compacted, well-cured concrete with a modest water-cement ratio returns. The ratio between two mixes is the cleanest internal check on a test campaign: same specimen size, same head, same 100-hour window, so the ratio of the percolates is the ratio of the coefficients. A second specimen that collected 95 ml over the same 100 hours would report 1.0 × 10⁻⁷ cm/s — ten times the value above, on the same rig, with nothing changed but the concrete.
Temperature is the correction you apply afterwards. If that test ran at 32 °C, Cl 7.7 puts the value at test temperature about 10% above the standard-temperature figure, so the coefficient at test temperature is 1.16 × 10⁻⁸ cm/s and the corrected value is 1.16 ÷ 1.10 = 1.05 × 10⁻⁸ cm/s. Both numbers belong on the report (Cl 9.1), along with the mix, the age at commencement, the duration, the specimen size and the test pressure — a certificate carrying only “K = 1.05 × 10⁻⁸ cm/s” cannot be checked by anybody.
Common deviations we see in permeability test reports
The formula needs the face area and the thickness of the specimen that was actually tested. A cube of 150 mm has a face area of 225 cm²; a 150 mm cylinder has 176.7 cm². Put the wrong one into the calculation and the coefficient moves by a factor of about 1.27. Cl 5.1 specifies cylinders with the height equal to the diameter.
Cl 7.8 is explicit: leakage through the seal gives entirely misleading results. No arithmetic recovers a test that leaked, because the water never went through the specimen.
Cl 7.8 requires the specimen surface to be carefully prepared by sand blasting or chiselling, because even a thin highly impervious skin can result in considerable underestimation of the permeability. A cast specimen that was floated and finished is not a permeability specimen.
Cl 7.8 puts a limit of about 0.2% on the air content of the water entering the specimen: excess dissolved air locks up the pore network and the permeability reads low. Draw periodic samples from the cell drain cock, and drain and refill with fresh de-aired water as soon as the limit is exceeded.
Cl 7.8: the outflow observation is liable to be affected by evaporation during collection. Run a blank observation on a similar bottle containing water and apply the correction — and use the inflow measurement as the additional check it is there to be.
Cl 7.6 requires the test to continue for about 100 hours after steady state has been reached, with the outflow taken as the average of all the outflows measured in that period. A single reading taken at the moment the flow steadied is not the result.
Field-checklist blocks
A short questionnaire, an inspector's checklist, and three worked examples that show how an IS 3085 test is set up, reported and checked, so that the number on the certificate describes your concrete rather than your rig.
Questionnaire — ask when the test is booked, before the specimen is cast
- Where does the acceptance value come from? IS 3085 sets none. Cl 8.1 gives the formula and Cl 9.1 the report contents; the limit belongs in the project specification or ITP. If the test request does not name that document, the result will arrive with nothing to judge it against.
- Is the specimen a cylinder? Cl 5.1 gives 150 mm diameter and 150 mm high, reducible to 100 mm where the nominal aggregate does not exceed 20 mm, and not less than about four times the nominal aggregate size where it exceeds 40 mm. The height has to equal the diameter.
- Is the specimen cured for 28 days, unless the engineer-in-charge specifies otherwise? And has it been cast, compacted and struck off with a sawing motion, with no further floating or finishing?
- Has the face been prepared by sand blasting or chiselling? Cl 7.8 is explicit that even a thin impervious skin results in considerable underestimation of the permeability. A float-finished specimen is not a permeability specimen.
- What head will be used, and will it be on the report? Cl 6.1 sets 10 kg/cm² as the standard, allows 5 kg/cm² for relatively more permeable specimens where steady state is reached in a reasonable time, and 15 kg/cm² for relatively less permeable ones where the sealing can be fully effective. The coefficient depends on the pressure used.
- At what temperature will the test run, and will the correction be applied? Cl 7.7 asks for 27 ± 2 °C, and gives about 10% on the coefficient for every 5 °C above the standard temperature, with the same reduction below it.
- Is the full 100 hours after steady state going to be run? Cl 7.6 requires the test to continue for about 100 hours after steady state is reached, with the outflow taken as the average over that period.
- If the contract names EN 12390-8 or DIN 1048-5 as well, will each result be reported against its own method reference? A depth in millimetres cannot be turned into a coefficient, so the two numbers do not belong in the same column.
Inspector's checklist — what to verify on site
- Specimen identification: diameter, height, age at commencement, mix reference and the curing record, against Cl 5.1 and Cl 5.2.
- Surface preparation record: sand blasted or chiselled, with no float finishing (Cl 7.8).
- Cell setup record: gasket, effective seal, de-aired water, and the dissolved air kept to about 0.2% with periodic samples drawn from the drain cock and the cell drained and refilled when it is exceeded (Cl 7.8).
- Blank observation on a similar water-filled bottle, with the evaporation correction applied to the outflow (Cl 7.8).
- Inflow and outflow rate data, or graphs of them, showing that steady state was actually reached before the coefficient was calculated (Cl 7.8).
- Head and duration on the report: the test pressure in kg/cm² and the total run, including the about 100 hours after steady state (Cl 6.1, Cl 7.6, Cl 9.1).
- Test temperature recorded, with both the coefficient at test temperature and the coefficient corrected to the standard temperature (Cl 7.7, Cl 9.1).
- Full report contents per Cl 9.1: mix, age at commencement, duration, specimen size, test pressure, test temperature and both coefficients, with the document the acceptance limit came from named on the certificate.
What happens if… three worked examples
What happens. A report that says only "K = 1.05 × 10⁻⁸ cm/s" cannot be checked by anybody, because Cl 9.1 requires the mix, the age at commencement, the duration, the specimen size, the test pressure, the test temperature, the coefficient at test temperature and the coefficient corrected to the standard temperature. The head is not a formality either: Cl 6.1 allows 5, 10 or 15 kg/cm² depending on the specimen, and the coefficient depends on the pressure used, so a result from a 15 kg/cm² run is not interchangeable with one from a 10 kg/cm² run.
Likely outcome. The number is compared against a limit that was written for a different method, and there is no way to establish whether it passes. Ask for the Cl 9.1 report, and for the document the limit came from. A millimetre figure quoted as an "IS 3085 limit" is a project-specific number, not a code value, and it is worth asking where it came from.
What happens. Cl 7.7 asks for 27 ± 2 °C, and the test ran 5 °C above the standard temperature, which is one correction step of about 10%. The coefficient at test temperature comes out at 1.16 × 10⁻⁸ cm/s, and the corrected value is 1.16 ÷ 1.10 = 1.05 × 10⁻⁸ cm/s. Left uncorrected, the certificate reads about 10% high, and a high coefficient means a more permeable concrete.
Likely outcome. A mix reported as more permeable than it is, so a limit it should clear becomes a marginal pass or a fail, and the error is invisible because the number still sits in a plausible range. Both figures belong on the report, along with the actual test temperature. The correction is trivial arithmetic, and skipping it makes the whole 100-hour run worth less than it was.
What happens. The formula needs the face area and the thickness of the specimen that was actually tested. A 150 mm cube has a face area of 225 cm²; the 150 mm cylinder the standard specifies has π × 7.5² = 176.7 cm². Put the wrong area into the calculation and the coefficient moves by a factor of about 1.27, and because the area sits in the denominator of K = (Q × H) ÷ (A × T × L), the larger figure makes the reported coefficient about 21% low, since 1 ÷ 1.27 = 0.79.
Likely outcome. A mix reported as less permeable than it is, which flatters it against a specification limit, on a specimen the standard does not recognise. Cl 5.1 asks for cylinders with the height equal to the diameter, and the surface preparation of Cl 7.8 has to be done on that specimen. Fix the specimen before you argue about the number.
Frequently Asked Questions
What is the difference between IS 3085 and EN 12390-8?
How is steady state of flow determined, and why does the test run for about 100 hours?
Why is the specimen cured for 28 days before the test?
Related articles
Amit Haridas
Founder & Proprietor, ConcreteInfo. 25+ years experience in concrete technology, durability testing, mix design and quality audit across India. For IS 3085 testing witness, permeability data interpretation, or water-retaining structure QA, contact amit@concreteinfo.in.
Technical updates, not sales emails
One email when a new article goes out on concrete technology, QA/QC and construction quality. No promotions, and your address is never sold or passed to advertising networks. Subscribing is separate from making an enquiry.