Valve packing used to be an afterthought — the soft rings in the stuffing box that someone re-tightened when a stem wept. Fugitive-emissions rules changed its status: on valves in VOC service, the stem seal is a compliance surface, and the paperwork around it is as important as the rings. The good news for buyers is that the vocabulary is finite — two materials, one spring trick, four standard names — and writing it on the RFQ takes five lines.
Key takeaways
- PTFE packing serves to about 200–260 °C with the lowest friction; graphite serves to roughly 450–650 °C class and takes the heat PTFE cannot.
- Live-loading — Belleville disc springs on the gland — holds packing preload as the set consolidates, and is the standard companion to high-cycle and low-emission service.
- ISO 15848-1 is the type test with tightness classes (methane classes AM/BM/CM at 50/100/500 ppmv; helium classes AH/BH/CH at 10⁻⁵/10⁻⁴/10⁻² mg·s⁻¹·m⁻¹); ISO 15848-2 is the production acceptance test.
- API 624 type-tests rising-stem graphite-packed valves (310 cycles, three thermal excursions to 260 °C, 100 ppm pass/fail); API 641 does the same family for quarter-turn at 610 cycles.
- TA-Luft’s 2021 revision points to ISO 15848-1 for valves; EPA Method 21 is screening, not certification.
Packing: PTFE vs graphite
The two materials divide at temperature. PTFE packing — braid or V-ring sets — runs to about 200–260 °C in the catalogs, with the lowest friction of any engineering solid, which keeps stem torque down and suits control-valve duty; its stem-seal surface wants the finer finish — published guidance for polymer stem seals calls for about 0.4 µm Ra or better in the packing travel area. Graphite is the heat answer: flexible graphite rings serve the 450–650 °C class depending on atmosphere, at moderate-to-high friction, and tolerate a slightly rougher stem — around 0.4–0.8 µm Ra per the packing-standard guidance. Neither is universally right: PTFE loses its corner above its class and can degrade with fumes near its ceiling; graphite costs torque and can consolidate. The trim-side context for what the packing protects is in valve trim materials.
Live-loading
Packing is a wear item under preload, and preload is the whole seal. As the valve cycles, packing consolidates — even dense graphite sets can consolidate by a couple of percent over their life — and the stud-bolt stretch that clamped it relaxes. Live-loading replaces that slack with a stack of Belleville disc springs on the gland follower: the spring stack’s deflection dwarfs the stud stretch, so the sealing load survives consolidation and thermal cycling. The literature points at the same service list every time: high cycle counts, thermal swings, remote or unmonitored locations, and anywhere fugitive-emission performance must hold between maintenance windows. Spec it as a construction note — live-loaded gland — and the valve arrives with its preload spring built in.
The standards
Four names carry the compliance conversation; each answers a different question:
- ISO 15848-1 — the type test: a valve model is cycled and thermally exercised, then classified by measured tightness. Methane classes run AM, BM, CM at 50, 100 and 500 ppmv; helium classes run AH, BH, CH at 10⁻⁵, 10⁻⁴ and 10⁻² mg·s⁻¹·m⁻¹ of stem diameter. The two class families do not convert into each other — specify which medium your project quotes.
- ISO 15848-2 — the production acceptance test: the routine test each delivered valve passes. A type-tested design plus production-tested units is the full claim.
- API 624 — the rising-stem graphite-packed test: 310 mechanical cycles with three thermal excursions to 260 °C, methane at pressure, pass/fail at 100 ppm. API 641 is the quarter-turn sibling at 610 cycles.
- EPA Method 21 — the screening field test: a sniffer reading in ppm that locates and classifies leaks. It certifies nothing; the leak definition it tests against (commonly 500 or 10,000 ppm) comes from the regulation, not the method.
- TA-Luft — Germany’s clean-air rule; its 2021 revision references ISO 15848-1 for valve stem testing, replacing the older VDI 2440 basis.
What to write on the RFQ
packing material— PTFE or graphite, to the temperature band.live-loaded gland— yes, for cycling or thermal-swing service.emission standard and class— e.g.,ISO 15848-1, methane BM, orAPI 624 compliant.test documentation— type test certificate and per-unit production test, named.maintenance expectation— the cycle count and temperature swings that justify live-loading. Our needle valves and ball valves quote packing as a line item, and the isolation-side counterpart of this spec is the double-block-and-bleed pattern from double block and bleed valves.
FAQ
Does graphite packing need live-loading?
Not to function — non-live-loaded graphite is normal practice, and a valve that can be monitored and re-tightened tolerates graphite’s slow consolidation fine. Live-loading earns its cost where preload cannot be maintained by hand: high cycle counts, thermal swings and low-emission service that must hold between outages. For a frequently cycled low-emission valve, the pairing is standard; for a static manual valve, plain graphite is honest value.
What is the difference between ISO 15848-1 and ISO 15848-2?
Part 1 type-tests a valve design and assigns its tightness class — methane AM/BM/CM or helium AH/BH/CH — under cycling and thermal excursions. Part 2 is the production acceptance test each delivered valve passes. A credible emissions claim needs both: the class belongs to the design, the test certificate belongs to the unit. Quoting the class without the production test is half the paperwork.
Is EPA Method 21 a certification?
No. Method 21 is a screening procedure: a portable instrument sniffs for VOCs at the stem and reports a reading, and the leak definition that reading is judged against — commonly 500 or 10,000 ppm — comes from the applicable regulation, not from the method. It locates leaks and triggers work orders; it does not certify a valve as low-emission. Certification lives in the ISO 15848 and API test families.
PTFE or graphite packing for my temperature?
Under roughly 200–260 °C, PTFE serves with the lowest friction and finest stem finish. From there through the 450–650 °C class, graphite is the material — with the torque and consolidation behavior that come with it. The temperature line is the first divider; cycle count and emission class decide live-loading; and the stem finish follows the material.
What changed under TA-Luft 2021?
The 2021 revision tightened German clean-air expectations and, for valves, moved the stem-seal test basis from the older VDI 2440 guidance to ISO 15848-1, with its own tightness limits layered on top. For buyers the practical effect is simple: a valve sold into German VOC service should quote ISO 15848-1 classification, and the vintage of the TA-Luft reference on the datasheet is worth a glance.
My valve leaks around the stem — what should I check first?
The gland, then the packing, then the stem, in that order. First tighten the gland follower nuts evenly, in a staggered cross pattern, about a quarter turn at a time — uneven tightening distorts the gland and over-tightening crushes the packing and raises stem torque without sealing anything. If a measured re-tighten does not stop the weep, the packing is worn or consolidated and the fix is replacement, not more torque. Before the new set goes in, inspect the stem through the packing travel area: scoring or pitting will cut fresh packing and the leak returns within weeks. A scored stem needs repair or replacement, not another ring set.
Can I tighten the packing while the line is pressurized?
The published default is no: isolate and depressurize before any packing work, and several makers state it as a flat prohibition under pressure. Some manufacturers do authorize minor in-service gland adjustment on specific designs, but the permission lives in that maker’s manual for that valve — there is no general low-pressure exemption, and the documented hazards are real: the leak path can enlarge abruptly, and hot or injected fluid causes burns and worse. If the weep is tolerable and the line cannot be isolated, monitor it and plan the outage; live-loading exists precisely so preload survives between adjustments.
