ISO 15848 Explained: Reading BH, CO3 and SSA0 on a Valve Certificate

A low-emission string is four independent claims, not one grade. In the form ISO FE - BH - CO3 - SSA0 - t(-29 °C/+230 °C) - Class 300 - ISO 15848-1, it says the stem seal held at or below 10⁻⁴ mg·s⁻¹·m⁻¹ of stem circumference, across 2 500 mechanical cycles and four thermal cycles, with no packing adjustment permitted during the test, over a tested range of −29 °C to +230 °C (−20 °F to 446 °F).

Each of those four fields can be improved or degraded on its own, and a supplier that quotes only the first one has told you a quarter of the story.

This article decodes each field, gives the tables behind them, and states what the certificate does not cover.

What does each field claim?

Each field is a separate claim: the test family (FE), the tightness class (BH), the endurance class (CO3), the permitted stem-seal adjustments (SSA0), the temperature range actually qualified (t), the pressure class of the tested specimen, and the part of the standard.

Keep in mind that the four middle fields are measured together but specified separately. A buyer who asks for "ISO 15848 certified" has asked for nothing actionable; a buyer who asks for a tightness class, an endurance class and a temperature class has asked a question a manufacturer can answer with a document.

What does the tightness class mean?

It is how much the stem or shaft seal is allowed to leak, stated against the tracer gas used — AH at ≤ 10⁻⁵, BH at ≤ 10⁻⁴ and CH at ≤ 10⁻² mg·s⁻¹·m⁻¹ of stem circumference when helium is the tracer, and AM, BM and CM at ≤ 50, 100 and 500 ppmv when methane is.

These are the limits in ISO 15848-1:2015 for the stem or shaft seal when helium is the tracer gas:

When methane is the tracer gas, the same standard expresses tightness as a concentration instead of a mass flow:

Body seals — the joints between pressure-retaining parts, not the stem — sit at ≤ 50 ppmv under either tracer gas.

Three things about this table are more useful than the numbers.

The helium classes and the methane classes are not convertible. The standard states that no correlation is intended between the tightness classes achieved with helium and those achieved with methane, and no correlation between the total-leak-rate method and the sniffed concentration method. A valve carrying a BH certificate has not been shown to be BM. If your specification names one tracer gas, a certificate for the other is not an answer.

The three mass-flow columns are the same limit expressed three ways — per metre of stem circumference, per millimetre of stem diameter, and as a volumetric flow. They are not three different stringency levels. When two documents quote "different" numbers for the same class, this is usually why.

Class AH is a different stem-sealing architecture, not a tighter version of Class BH. The standard's own remark points at bellow seals, which is why the jump from BH to AH changes the valve's construction rather than its packing grade.

What does the endurance class mean?

It is how many mechanical and thermal cycles the valve completed while still holding its tightness claim, counted separately for isolating valves (CO1 at 205 cycles up to CO3 at 2,500) and control valves (CC1 at 20,000 up to CC3 at 100,000).

Two details decide whether a quoted endurance class means what a buyer thinks it means.

The classes build on each other rather than replacing each other. For isolating valves, CO1 is 205 full-stroke cycles with two thermal cycles, run as 50 cycles at ambient, 50 at test temperature, 50 at ambient, 50 at test temperature and a final 5 at ambient. CO2 adds 1 295 mechanical cycles with one thermal cycle; CO3 adds a further 1 000 with one more thermal cycle. A certificate that states C01-C02-C03 has therefore run the sequence, not three separate test programmes.

CO and CC cycles are not the same motion. An isolating-valve cycle is a full stroke. A control-valve cycle is defined around ±10% of the total stroke. The 100 000 cycles of CC3 are not 40 times the endurance of CO3; they are a different movement repeated more often. Comparing a CO class against a CC class is the endurance version of comparing two different units.

What does the temperature class mean?

It is the temperature range the qualification actually covers, not the valve's catalogue range — t-196 °C, t-46 °C, tRT, t200 °C or t400 °C in the coded form, or an explicit range such as t(-29 °C/+230 °C).

Note that one test does not cover both ends of a wide range. To qualify a valve from −46 °C to 200 °C, two tests are required: one at −46 °C and one at 200 °C. A certificate carrying only t200 °C tells you nothing about cold performance, and a specification that assumes otherwise will pass a valve that was never tested at the cold end.

Certificates also appear in a second form that quotes the tested range directly rather than a class code — for example t(-29 °C/+230 °C), as issued for German TA Luft attestations. Both forms report what was tested. The explicit range is simply the narrower statement, and it is worth reading it as exactly that.

Why is SSA the field buyers should watch?

Because it says whether the packing had to be tightened during the type test, and that is the field that speaks to maintenance interval: SSA0 means nobody touched the gland at any point, and it is the field most often left out of a summary.

SSA is the number of stem-seal adjustments permitted during the type test, and it is the least understood field on the certificate.

SSA0 means the valve completed the whole endurance programme — every cycle, every thermal swing — without anyone touching the gland. SSA1 allows one adjustment, SSA2 two, SSA3 three, each tied to an endurance class.

A valve that needed an adjustment mid-test is not a failed valve, but it is a different claim. If the maintenance interval is what you are buying — a low-emission valve specified because it will not need attention between turnarounds — then SSA0 is the field that speaks to that, and it is the field most often omitted from a summary. Ask for the full string, not the tightness class alone.

Which part applies — Part 1 or Part 2?

Part 1 if a design is being qualified, Part 2 if production valves are being accepted — and a valve has to be qualified to Part 1 before a Part 2 test can be run on it, so a Part 2 result never stands on its own.

ISO 15848 has two parts, and they answer different questions:

• Part 1 is the classification system and qualification procedure for type testing. It is the source of every class in this article.
• Part 2 is the production acceptance test, intended for standard production valves where a fugitive emissions requirement is specified.

A valve must have been qualified to Part 1 before a Part 2 test can be performed on it, so a Part 2 result does not stand on its own. The two parts also use tightness class letters that are not the same scale — a Part 2 Class B and a Part 1 Class BH are different claims expressed with similar-looking labels.

If a supplier's certificate says "tested to ISO 15848" without a part number, the first question is which part.

Why is the question being asked now?

Low-emission valve specifications are not a fashion. They are downstream of work-practice rules that define a leak numerically.

Two observations follow, and both matter more than the individual figures.

The regulators' numbers and the standard's classes are answers to different questions. A 500 ppmv reading under Method 21 is a screening trigger in a work-practice rule: it starts a repair clock on a component in service. 500 ppmv is also the ISO 15848-1 Class CM limit — but that is a type-test acceptance criterion on a test bench. The number matching does not make the two interchangeable, and a certificate does not create a compliance position. This article is not a compliance determination.

Note that TA Luft does not simply adopt ISO classes. It uses ISO 15848-1's test parameters but sets its own leakage limits, and its limits are pressure- and temperature-dependent. A valve can hold a Part 1 class that satisfies one regime's table and not another's, which is why the specification, not the certificate, decides.

Which applies — API 624 or API 641?

The fork is stem movement: API 624 is written around rising and rising-rotating stem valves with graphite packing, API 641 around quarter-turn valves.

The split is where most specifications go wrong.

Be aware of the most common misapplication in this field: requesting API 624 for a ball valve or a butterfly valve. API 624 is written around rising-stem designs; the quarter-turn standard is API 641. A specification that names the wrong one is a specification that cannot be met, and both are separate from API 622, which tests the packing set on a fixture rather than a complete valve.

One version detail worth confirming with the supplier rather than assuming: the API 624 scope boundary for valve size changed between editions, so the size at which a valve leaves the standard's scope depends on which edition is being cited.

What does a Chinese-made valve's certificate look like?

China has converted both parts of ISO 15848 into national standards, and the conversions are modified adoptions rather than verbatim translations — which changes the markings on the certificate.

This is the practical point for a buyer of Chinese valves: the classification logic, the tightness classes and the endurance classes are the same family of requirements, but the string printed on the document is not the string a European or American specification will be looking for. When an enquiry asks for ISO FE - BH - CO3, and the works certificate says GB FE - BH - CO3 - GB/T 40079, the two are speaking the same language with a different label — and a buyer who does not know that will read the certificate as non-compliant.

What does this standard not cover?

Four things by the standard's own scope — end connection joints, vacuum application, corrosion effects and radiation — plus three further boundaries that matter when a certificate is read: it is not a seat-tightness standard, not a fire-safety standard, and not a batch release.

The standard's own scope is explicit about four exclusions:

• End connection joints — the flange or weld face is not part of the qualification
• Vacuum application
• Effects of corrosion
• Radiation

Three further boundaries matter to a buyer reading a certificate:

It is not a seat-tightness standard. Stem and body-joint leakage are what it measures. Seat leakage is a different acceptance question and a different standard — the one that decides whether a metal-seated design can hold Rate A is not this one.

It is not a fire-safety standard. Fire testing is a separate qualification with its own procedure.

A type-test certificate is not a batch release. It qualifies a design and its tested configuration. Whether the valve in the crate matches the tested configuration is a production acceptance question, and that is Part 2 — or the national equivalent.

When does this go to a human?

Five cases for low-emission specifications: any duty below −101 °C (−150 °F), where automatic coverage ends unconditionally; any requirement for a shut-off class the valve type cannot support at that size and temperature; any specification that names a tightness, endurance or temperature class but not the tracer gas or the part number; any duty that mixes a regulatory threshold with a type-test class; and any case where the works certificate and the enquiry use different marking conventions.

• Any duty below −101 °C (−150 °F), where this site's automatic coverage ends unconditionally. A t-196 °C qualification belongs to that band: the certificate can exist for the valve, and the duty is still outside the range the selector covers, which is why every LNG cryogenic valve enquiry here goes to a person
• Any requirement for a shut-off class that the valve type cannot support at the size and temperature asked for
• Any enquiry where the specification names a tightness class, an endurance class or a temperature class and does not name the tracer gas or the part number
• Any duty where the specification mixes a regulatory threshold with a type-test class
• Any case where the works certificate uses one marking convention and the enquiry uses another

What people assume, and what the standard says

Common questions about fugitive emissions certificates

Which class should I specify?

That depends on the service and on the regime that applies to your site, and working it out is a compliance determination, which is outside what this site does. What we can tell you is what each class means and which fields the specification must name — the tightness class, the tracer gas, the endurance class, the temperature class and the part.

Why does my supplier's certificate show numbers that do not match the table?

Most often because the table is expressed in three equivalent mass-flow columns. Check which column the certificate uses before concluding there is a discrepancy.

Does a low-emission certificate add cost?

We do not publish prices, and the cost of a qualification route depends on the design, the size and the test house. What can be said factually is that Class AH typically implies a bellow seal rather than a packing set, which is a construction change.

My valve is a ball valve. Can it get ISO 15848-1?

ISO 15848-1 covers isolating and control valves and includes quarter-turn designs, so the route exists. The American equivalent for quarter-turn valves is API 641 rather than API 624, which is a common specification error.

Do I still need a seat-tightness standard if I have a fugitive emissions certificate?

Yes. They measure different leak paths, and one certificate does not stand in for the other.

What this article does not cover

Note that this article does not determine whether a valve complies with any regulation, does not select a tightness class for a duty, does not select a packing or a stem-sealing architecture, and does not calculate emissions. It explains what each field on a low-emission certificate claims, gives the tables behind them, and states the boundaries of the standard and of this site's coverage.

Where the next constraint sits

A fugitive-emission certificate says nothing about seat leakage, which is a separate acceptance question with its own standard. Where the seat rather than the stem sets a valve's limit, and when the seat material rather than the certificate decides the choice, is set out in Soft-Seated vs Metal-Seated Valves: Where the Temperature Line Is.

Two of the documents this certificate family sits beside are in motion, and the standard most often named in cryogenic specifications was withdrawn in 2015. BS 6364: Withdrawn in 2015 — What a Cryogenic Valve Specification Should Name Instead and GB/T 24925-2019: What a Chinese Cryogenic Valve Test Report Actually Contains cover the cryogenic end of the same question.

Sources