ISO 5208 Leakage Rates A to G: What the Letter on a Brass Valve Certificate Allows, by DN

A pallet of brass ball valves arrives with a certificate that reads "tested to ISO 5208, Class A", and your specification said Rate A. The answer: ISO 5208:2015 grades closure leakage, the leakage past a closed seat, in ten lettered rates, A, AA, B, C, CC, D, E, EE, F and G. Rate A means no visually detectable leakage for the test duration; every other rate is a number multiplied by the valve's DN. The rate that applies comes from the product standard, or from a stricter purchase order, and the standard prescribes the wording: "ISO 5208:2015, Leakage rate A".

One number shows why the fluid matters more than the letter: on a DN15 valve the closure test lasts 15 seconds, and Rate D allows 1.5 mm³/s of water, 0.36 of a single drop in the whole test. An air test separates the rates; water at that size does not. Below: the table resolved for DN15 to DN50, the test a ball valve requires, and the clause wording that makes the rate enforceable.

▶  Play

ISO 5208 Uses Rates A to G, Not Classes

The standard, "Industrial valves — Pressure testing of metallic valves", covers the shell test that proves the body holds pressure, an optional backseat test, and the closure test that measures what passes a closed seat. It sits under a product standard: where the two differ, the product standard wins. The 2015 text is the fourth edition, confirmed by ISO's systematic review in February 2026. Its foreword lists three changes from 2008: definition 2.12, a revised Table 1 (butterfly valve types split, optional high-pressure gas closure test added), and a rewrite of 4.12.4.1 a) for valves rated PN 6 and below. The duration and leakage-rate tables were not touched.

The vocabulary trips up more buyers than the numbers do. Table 4 lists ten leakage rates, A to G with AA, CC and EE between them, and no Class A anywhere. Classes I to VI come from ANSI/FCI 70-2, a control-valve seat-leakage standard. API 598 uses neither letters nor Roman numerals, and the ISO text calls its correspondence with API 598 "loosely defined". "Bubble-tight" is a sales phrase; the standard's term, defined in clause 2.12, is "no visually detectable leakage".

The Rate Table Resolved for DN15 to DN50

Every explainer prints the multipliers; none multiplies. Table 4 is written per DN, so a buyer of half-inch to two-inch brass valves has to do the arithmetic first. Threaded sizes convert through Annex A, 1/2 in being DN15 through to 2 in being DN50, and the closure test for all of them runs 15 seconds under Table 2.

Run the numbers for DN15. Rate B in liquid is 0.01 × 15, so 0.15 mm³/s, which the drop-count column puts at 0.036 of a drop over the test. Rate D is 1.5 mm³/s, still 0.36 of a drop; one whole drop first appears at Rate D on a DN50 valve. In gas the ladder is steeper: Rate B at DN15 is about one bubble, Rate C about ten, Rate D about a hundred.

Source: ISO 5208 Table 4 multipliers (× DN) and Table 2 durations, resolved for a 15-second test; rates AA, CC, E and EE omitted. Computed 12 September 2026.

The consequence is blunt. Up to DN40, no rate from A to D produces a single drop in a 15-second water test, so a hydrostatic seat test on a half-inch valve is pass-fail on visible weeping and nothing finer. Either specify the air test and count bubbles, or specify Rate A and accept that "no visible leakage" is the whole criterion.

The Test Behind the Letter: Fluid, Pressure, Duration

The closure test follows the shell test, in which the body proves it holds pressure with the obturator partly open. The fluid is water or another liquid no more viscous than water, at 5 °C to 40 °C, or a gas, normally air. Test pressures hang off the CWP, the cold working pressure the manufacturer assigns and marks on the valve.

ISO 5208:2008 fixed the gas closure pressure at 6 bar ± 1 bar and the liquid pressure at a minimum of 1.1 × CWP. The 2015 edition revised the gas sub-clause for valves rated PN 6 and lower, so on a low-pressure valve ask which figure was used. Table 2 sets the duration: 15 seconds for isolating valves up to DN50 (60 seconds for check valves), 60 seconds from DN65 to DN150 and 120 seconds from DN200 upward, counted only once the valve is at pressure.

Table 1 is the part buyers get backwards. For a floating ball valve, a gate valve or a plug valve up to DN100 at PN 250 or below, the required closure test is the low-pressure gas test, and the high-pressure liquid test is optional. Clause 4.9 adds the catch: the design must still be able to pass it, and the manufacturer supplies evidence when the order asks. Globe, check and concentric butterfly valves run the other way. An inspector can also check that seats are free of grease, that ball and gate cavities are filled, and that each port is tested in turn.

Source: ISO 5208:2015 Table 1, Table 2 and clause 4.12.4; gas pressure as fixed in the 2008 text, revised in 2015 for PN ≤ 6. Read 12 September 2026.

Which Rate Applies to a Brass Water Valve, and the PO Wording

ISO 5208 does not pick the rate for you: Table 4 Note 2 says the rate that applies is the one in the valve product standard, or a more stringent rate in the purchaser's order. For copper-alloy ball valves in building water supply that product standard is EN 13828:2025, covering DN6 to DN100 at PN10 or PN16 (10 bar or 16 bar nominal), which replaced the 2003 edition on 17 March 2025. Isolating valves in water-supply networks fall under EN 1074-2. The European test standard cited alongside is EN 12266-1:2012, whose values Note 4 says match Rates A to G; the shell and seat tests a gate valve certificate should list are covered separately.

Three buyers, three clauses. An importer stocking DN15 to DN50 brass ball valves for distributors has no project specification to inherit, so the order carries the rate. The clause reads: "Closure test to ISO 5208:2015, Leakage rate A, gas, per Table 1; certificate of compliance per clause 4.13 with each batch." A contractor buying against a European tender that names EN 13828 writes the product standard first and the rate second. A buyer of brass check valves for a booster set should expect a numeric rate rather than Rate A: the required test there is liquid at 1.1 × CWP over 60 seconds, and a metal seat will show something.

Rate A is the practical choice for small-bore soft-seated valves because it is the only rate with no numeric allowance, and the water test cannot distinguish the numeric rates at DN15 to DN40 anyway. Clause 4.12.7 b) prescribes the form of words; clause 4.8.2 makes any optional test you list binding; clause 3.1 gives you the right to witness, with five working days' notice from the factory.

Reading the Certificate: Five Things It Must Say

Clause 4.13 is one sentence, and most orders skip it: the manufacturer provides a certificate of compliance when the purchaser's procurement document specifies it. No clause, no certificate; "tested to ISO 5208" on a datasheet describes a factory habit, not your batch. Five fields decide whether the certificate is worth the paper.

• Standard and edition: "ISO 5208:2015". A 2008 citation is a withdrawn edition; a bare "ISO 5208" says nothing about which Table 1 was applied.
• The rate letter, spelled as a rate: "Leakage rate A". A certificate that says "Class A" or "Class VI" is mixing FCI 70-2 vocabulary into an ISO document.
• Test fluid: water or air. On a ball valve the required test is air; a water-only certificate means the required test may not have run.
• Pressure basis: the CWP the figure hangs off, and the marked value on the valve should agree with it.
• Duration and size: 15 seconds for DN15 to DN50, tied to the DN on your order.

Two traps recur: an ISO 5208 heading with API 598 drop counts underneath, which do not transfer, and a rate quoted for a different size, which matters because every numeric rate scales with DN.

What IFAN Publishes on Its Brass Valves, and What to Ask For

The published position is specific enough to check. IFAN makes ball valves from DN8 to DN100 in PN16, PN25 and PN40 and gate valves from DN15 to DN50 with BSP or NPT threads. Every body is CW617N, forged or cast, and the seats are PTFE; the alloy is covered in the guide to CW617N and lead-free brass.

Testing runs in a CNAS-accredited laboratory on every production batch, with hydrostatic, thermal and impact tests before shipment, and the material certificate states CW617N per batch. Certifications listed are CE, ACS, WRAS and SASO; there is no minimum order on the valve range and the quoted lead time is about ten days from stock. The product catalogue lists the variants by size and end type.

What is not published is the part this article is about. IFAN's pages do not state which ISO 5208 or EN 12266-1 leakage rate the valves are released against, whether the batch closure test runs in air or water, at what pressure or for how long, and no sample certificate is posted. Do not read "hydrostatic test" as "Rate A in gas"; those are different tests. Put the rate on the order, ask for the clause 4.13 certificate with the first batch, and treat the answer as the supplier assessment. The forged versus cast body question belongs in the same email.

Conclusion

For brass ball and gate valves up to DN50, specify "ISO 5208:2015, Leakage rate A" in gas, name the product standard first if the project has one, and ask for the certificate in the same clause. The answer flips only where a metal-seated check valve makes a numeric rate the honest requirement. Six checks before the order goes out:

• The standard reads 2015, not 2008.
• The word is Rate, not Class.
• A ball-valve certificate shows the air test.
• Duration is 15 seconds at DN15 to DN50.
• The rate is quoted for your DN, not a larger valve.
• The clause 4.13 certificate and the witness option are written in, not assumed.

If your current supplier's datasheet says "Class A", send them the clause wording from this page; the reply tells you more about the factory than the datasheet did.

Frequently Asked Questions