“We have NSF 61” and “we have NSF 372” appear on quotations as if they were the same sentence. They are two completely different things: 61 tests what comes out of the product into water (leaching); 372 tests how much lead is in the material (content). A valve can sit at 0.08 % lead and still fail 61 because its brass releases too much at pH 5; conversely, a valve at 0.3 % lead fails 372 no matter how clean its leaching data looks. For the US market, both lines must pass.
In NSF/ANSI 61 a valve falls under Section 9, Mechanical Plumbing Devices — the same section as faucets, flexible connectors and stop valves — and the criterion is the Q value, an average lead leaching dose measured in micrograms. The 2020 edition cut Q to one fifth of its previous value and made the tighter figure mandatory from 1 January 2024. This article runs: what each standard governs, how Q changed, how the extraction test is run, how the 372 weighted average is calculated, the certification route, the Chinese and European mappings, and what can be done in-house.
1. 61 versus 372 in one table
| Dimension | NSF/ANSI/CAN 61 | NSF/ANSI 372 |
|---|---|---|
| Governs | Health effects of materials in contact with drinking water (leaching) | Lead content of wetted surfaces |
| Method | Soak in specified waters, analyse the extract for metals and organics | Calculate lead percentage weighted by wetted surface area |
| Criterion | No contaminant above MCL / MAC; Section 9 additionally applies the Q lead criterion | Weighted-average lead < 0.25 % |
| Nature | Performance-based (result driven) | Composition-based (formulation and area calculation) |
| Legal root | Referenced by state rules and plumbing codes | US Safe Drinking Water Act §1417 (amended 2011, effective 4 Jan 2014) |
| Section for valves | Section 9, Mechanical Plumbing Devices | All wetted parts |
| Typical duration | Months (conditioning plus multiple sampling rounds) | Weeks (material certification can shorten it) |
One-line memory: 61 asks whether lead gets into the water; 372 asks how much lead the product carries. Neither proves the other. Most state plumbing codes want both, and municipal projects want a listing from a certification body, not merely a test report.
2. The Section 9 Q criterion — from 5 µg to 1 µg
Q is the average dose of lead, in micrograms, that a mechanical plumbing device may leach. It is calculated statistically from the extract concentration and the device’s internal volume — it is not a concentration and cannot be compared directly with µg/L. That misreading is the most common error in report review.
| Product category | Previous criterion | 2020 edition | In force |
|---|---|---|---|
| Endpoint devices (faucets, flexible connectors, terminal valves — the “last litre”) | Q ≤ 5 µg | Q ≤ 1 µg | Optional in 2020, mandatory from 1 Jan 2024 |
| Supply stops, flexible connectors and components | Q ≤ 3 µg | Q ≤ 0.5 µg | As above |
That is a fivefold tightening. The practical consequence: products certified before 2020 may not hold their listing after 2024 without a design change. Three engineering levers actually work: move to lead-free copper alloy (bismuth or selenium grades, or C89833-type); take the wetted surface out of leaded brass and into stainless or polymer; or make the internal surface treatment dense — coating pinholes are leaching channels. The third is routinely underestimated: thin plating over a rough substrate keeps releasing lead from micro-pores.
3. Running the extraction test — three waters, two phases
- Wetted parts list: every part, its material grade, its supplier and its wetted surface area. This list drives what is tested and how Q is computed, and it is the document certification bodies examine most closely.
- Choose the exposure waters: metals use pH 5 and pH 10 (one acid-side, one alkaline-side, covering both corrosion tendencies); organics use pH 8, which must be chlorine-free during exposure. Devices containing adsorptive media are tested both with and without the media.
- Conditioning: 1 to 14 days, most commonly 14; at least 10 water changes with no less than 24 h between them. This step removes grease, release agents and machining residue — without it you measure the workshop, not the material.
- Exposure and sampling: after conditioning, the exposure period begins and samples are drawn after the specified stagnation (Section 9 devices follow the 16-hour overnight stagnation, first-draw logic), then analysed for metals and organics.
- Judgement: results against MCL / MAC; Section 9 additionally against Q. Exceeding any limit under any exposure water fails the product.
Note the difference from Chinese practice: GB/T 17219 uses a 24-hour soak, while NSF 61 is “14 days of conditioning plus sampling after a specified stagnation”. Substituting the 24-hour data for NSF 61 is a difference of an order of magnitude in time scale.
4. The 372 weighted average — by area, not by mass
The 0.25 % threshold is easy to remember and easy to calculate wrongly. The rule is a weighting by wetted surface area share, not by mass and not by part count. Here is an example that really happens:
| Part | Material | Lead | Wetted area share | Contribution |
|---|---|---|---|---|
| Body | Leaded brass (CW617N type) | 1.8 % | 70 % | 1.26 % |
| Ball | Nickel-plated brass | 1.5 % | 15 % | 0.225 % |
| Stem | Stainless steel | 0 % | 10 % | 0 % |
| Seals | EPDM | 0 % | 5 % | 0 % |
| Weighted average | 1.485 % → fail | |||
Change only the body to lead-free alloy (lead ≤ 0.1 %) and the weighted value drops to 0.295 % — still above 0.25 %, because the ball’s 15 % of the area still carries 1.5 % lead. Change both (lead-free body plus stainless ball) and it falls to 0.07 %. That also explains why swapping only the ball achieves nothing: the body holds 70 % of the area and is the deciding term. Do this calculation before tooling, not after.
Two more practical points. Solder and brazing filler are wetted parts — leaded solder is a classic failure source. And platings are assessed as platings, but where the substrate contains lead it still enters the calculation, because plating is not a barrier unless it can be shown to be continuous and pore-free.
5. The certification route — who issues, what a listing covers
- Issuer: an ANSI-accredited third-party body (NSF, UL, IAPMO, CSA). In-house data cannot support a listing.
- Scope: a listing covers a product family plus material formulation plus wetted-area structure — not an unlimited range of models. Change the brass supplier, the seal compound or the internal coating and it must be re-evaluated.
- Surveillance: annual factory audit plus sample retesting. A listing is a continuing status, not a one-off certificate.
- Marking: the 61 / 372 marks may be used only within the listed scope. Marking a component’s approval as a complete-device approval is a common overreach.
6. The Chinese and European mappings
| System | Standard | Approach | Difference from NSF |
|---|---|---|---|
| China | GB/T 17219 (hygienic evaluation of drinking-water equipment and protective materials) | 24-hour soak, hygiene indicators on the soak solution, plus a hygiene licence | Material hygiene evaluation rather than product certification; short time scale; no Q-type statistical criterion |
| Europe | EN 15664-1 (metal release, dynamic rig) | Dynamic circulation per national rules, long duration | Closer to real hydraulic conditions and longer; no unified 0.25 % lead gate |
| Europe | EN 681-1 (elastomer seals) | Compound and hygiene requirements | Elastomers only, no metal release |
| Europe | EN 13828 (potable-water ball valves) | Product standard including a dezincification gate | Dezincification is a corrosion mechanism, not a lead-content measure |
| US | NSF/ANSI 61 + 372 | Leaching plus content, third-party listing | — |
The commonest export mistake is answering a US customer with a GB/T 17219 report. Method, duration and criteria all differ, and a technical reviewer sees it immediately. Keep the Chinese report for Chinese projects, obtain an NSF 61/372 listing for US projects, and follow national hygiene rules plus the EN 15664 family for Europe.
7. What can be done in-house, and what cannot
| Step | In-house | Must go to a third party |
|---|---|---|
| Sample preparation | Cut by wetted area, clean, label, retain | — |
| Conditioning | 14-day water-change schedule, pH and chlorine monitoring, blank control | — |
| Exposure and sampling | Constant temperature, dark, specified stagnation | — |
| Analysis | — | Trace metals by ICP-MS, organics by GC/MS |
| Evaluation | — | Toxicological assessment, Q statistics, listing decision |
| 372 calculation | Wetted parts list and area-weighted table can be built in-house | Statutory material composition testing, certifier review |
Turning “wetted parts list + surface area + material grade + lead content” into a fixed template is the highest return on effort available: you can compute the weighted value at design stage and avoid discovering an over-limit after tooling, and you arrive at the lab with the paperwork complete instead of spending two more rounds supplying it.
Where pressure and tightness data are needed, shell and seat tests follow GB/T 13927-2022 or the applicable product standard — an entirely separate line from hygiene certification. For the equipment side see the water valve and faucet comprehensive test bench and the valve pressure test machine.
FAQ
If we have NSF 61, do we still need 372?
Yes. 61 proves the leaching is acceptable; 372 proves the lead content is acceptable. They are judged independently and projects normally require both; missing one means no listing. Nor does 372 substitute for 61 — a material free of lead is not necessarily free of other contaminants.
Is the Q value the same as µg/L?
No. Q is an average lead leaching dose in micrograms, computed from extract concentration and internal volume. µg/L is a concentration. The same Q corresponds to a lower concentration in a device with a larger internal volume. Reading Q as a concentration in a report leads to entirely wrong conclusions.
Are certificates obtained before 2024 still valid?
Depends whether the listing has been updated to the tightened 2020 criteria. Going from Q ≤ 5 µg to 1 µg (and 3 µg to 0.5 µg for supply stops) is a fivefold step, and older designs often cannot meet it without change. Check the certifier’s current listing to see whether the tightened criterion is stated.
Will a stainless ball get us through 372?
Not on its own. 372 weights by wetted area, and the deciding term is usually the body, which holds the largest share. Swapping the ball moves the weighted value only slightly. Compute the area-weighted table first, then decide which part to change.
Specific clauses follow NSF/ANSI/CAN 61 (including the 2020 Section 9 revision) and NSF/ANSI 372; the tables above are for orientation in selection, quoting and certification discussions. For the pressure-test requirements on the same valve see AWWA C515 and C509; for the European supply-valve system see EN 1074; for dezincification and acoustics on building potable-water ball valves see EN 13828; and the full map is in the valve test standards index.
Related test equipment
Each test above maps to a bench. Pick by valve type and by the item you need to prove:
- Valve pressure test machine — shell at 1.5x and seat at 1.1x, with a report per valve
- Valve comprehensive test bench — pressure, tightness, flow, torque and life cycling on one clamping system
- Valve testing equipment range — browse all models by valve type and test item