1. First — Understand What “CADR Certified” Actually Means (And What It Doesn’t)
CADR = Clean Air Delivery Rate — the volume of truly cleanair a purifier delivers per unit time. It’s a product of two things, not one:
CADR = Airflow (m³/h) × Filtration Efficiency (at the target particle size)High airflow with a thin mesh = high airflow but low efficiency = low CADR.
Dense True HEPA with a weak fan = high efficiency but choked airflow = low CADR.
The sweet spot is engineering, not a sticker.
The Three Pollutant Categories (per AHAM standard)
| Category | Particle Size Proxy | What It Tells You |
|---|---|---|
| Smoke | 0.09–1.0 µm — the hardest category; penetrates deep into lungs | Tests fine-particle performance (PM2.5 / PM1.0 realm) |
| Dust | 0.5–3.0 µm | Mid-range particulate |
| Pollen | 5–11 µm | Larger allergen-scale particles — easiest to capture, highest CADR of the three |
AHAM (ANSI/AHAM AC-1 standard) requires all three tested on three separate units with averaged results.
China’s GB/T 18801 framework tests particulate CADR (via standard smoke/NaCl/KCl aerosol) and separately tests formaldehyde CADR for gaseous — but the key compliance point is: the claimed CADR must be within 90% of tested value or it’s a violation.
2. The Three “Certificate Tracks” — And Why Most Are Worthless
This is where B2B buyers get burned. There are three tiersof “CADR documentation,” and only one is defensible:
⭐ Track 1 — AHAM Verifide® (Gold Standard, North America / Global Premium)
- The manufacturer sends 3 production units to an AHAM-recognized independent lab
- Tested per ANSI/AHAM AC-1 in a calibrated 28.5 m³ test chamber
- Results go into AHAM’s public online database: ahamverifide.org→ “Identifying AHAM Verifide Products”
- If it’s in the database with your exact model number, it’s real
Verification step: Type the brand/model into the AHAM directory. If the factory says “we have AHAM certified” but the model isn’t listed → it’s fake. The seal can be printed by anyone; the database can’t be faked.
AHAM Verifide mark = the only CADR claim that survives retailer compliance teams at Amazon, Best Buy, Costco.
⭐ Track 2 — GB/T 18801 Test Report from a CMA/CNAS-Accredited Lab (China Domestic / Export)
Under China’s system, legitimate CADR data comes from a CMA-accredited (China Metrology Accreditation) testing lab — the report carries a CMA stamp and certificate number you can cross-check.
| What to Check on the Report | Pass | Fail |
|---|---|---|
| Lab name & CMA license number | 中国家用电器检测所, 中检集团, SGS-CSTC, Intertek, TÜV, CTI, etc. with visible CMA 章 | Unknown lab name, no CMA stamp, or the “stamp” looks like clip art |
| Sample description | Exact model number, voltage, fan speed setting tested (usually max speed), filter model/spec | Says “air purifier” generically — no model lock |
| Test standard cited | GB/T 18801-2015 (or newer revision) — Appendix B (particulate) and Appendix C (gaseous/formaldehyde) | Vague: “tested per factory method” or no standard cited |
| Test chamber volume stated | 30 m³ or 81 m³ (the two standard sizes in CN) | Not stated — impossible to validate the math |
| Three values reported | Particulate CADR (m³/h), Formaldehyde CADR (m³/h), CCM grade (P1–P4 / F1–F4) | Only one number (“CADR 800!”) with nothing else |
| Attached photos | Show yourunit, yourfilter, yourlabel | Stock photo from Google |
Critical cross-check: The CADR number obeys a sanity limit. A purifier’s max CADR ≈ 60–75% of its mechanical airflow at max fan speed, because even H13 HEPA leaks a few percent. If they claim 800 m³/h CADR but the unit is a 7-inch desktop shell with a 90mm fan — the math physically can’t work. Ask for the fan curve / airflow spec and do a back-of-envelope.
⚠️ Track 3 — “Factory Test Report” / “In-House Data” / No-Name Lab (The Danger Zone)
“We tested in our own lab — here is the report.”
Walk away. This is the source of >80% of inflated CADR claims on B2B platforms. Common tricks include:
| Cheat | How It Works | How to Catch It |
|---|---|---|
| Testing without the filter installed | Run the fan bare — airflow is maxed, “CADR” looks huge | Demand test photos showing the actual production filterin place with serial/model visible |
| Testing at 100% fresh-air bypass | Open intake/exit flaps to let air shortcut around the filter | Verify test setup diagram in the report matches sealed operation |
| Using a nano-sized test chamber and scaling up | Small chamber = less dilution → artificially high decay rate → inflated CADR | Check: chamber volume must be ≥ 28.5 m³ (AHAM) or 30 m³ standard (GB/T) |
| “Peak” reading, not steady-state average | Laser particle counter spikes then decays — they report the spike | Proper CADR uses the exponential decay slopeover 15–20 min, not a momentary number |
| One cherry-picked unit | They hand-built one golden sample, tested it, then mass-produce a cheaper version | This is why you need 3rd-party retesting of random production pulls, not just the initial cert |
| Photoshopped PDF | Entire report fabricated | Call the lab. Ask to speak to the testing engineer. Reference the report number. Fake labs hang up; real ones transfer you. |
3. The Formaldehyde / VOC Gap — Where “CADR” Alone Misleads
Particulate CADR (dust/smoke/pollen) gets all the marketing ink, but for most buyers the real air quality complaint is VOCs — paint fumes, formaldehyde off-gassing, cooking odors, pet smells.
| Metric | What It Measures | Why It Matters for Sourcing |
|---|---|---|
| Formaldehyde CADR (m³/h) | Removal rate of gaseous HCHO under GB/T 18801 Appendix C | Essential if selling into China (new furniture/off-gassing is top concern) and increasingly scrutinized in EU/US |
| CCM — Cumulate Clean Mass (P1–P4 for particles, F1–F4 for formaldehyde) | How much total pollutant mass the filter can adsorb before CADR drops to 50% | F4 = best (≥1500 mg formaldehyde). Low-grade carbon = great CADR on Day 1, dead by Week 4 |
| Carbon weight & type | Activated carbon by weight in grams, not “carbon filter” as a description | <100g of loose granule carbon in a mesh pocket = cosmetic. Real chemisorption needs 200–600g+ packed bed or pelletized carbon |
| Ozone emission | Must be ≤5 ppb (CARB limit) / ≤50 ppb (UL 867) for electrostatic/ionizer units | California CARB certification is the gatekeeper — no CARB = can’t sell in CA; ionizer-only units are regulatory landmines |
Red flag phrases:
- “Negative ion purification”as the primary mechanism — usually means no mechanical HEPA, with ozone risk
- “Nano silver / photocatalytic TiO₂”without ozone test data — UV+P25 TiO₂ can generate ozone/nano-particle concerns under certain conditions
- “Formaldehyde removal rate 99%”with no CADR or CCM backing — removal % without a time/volume basis is a nonsense number
4. Safety Certificates vs. Performance Claims — Two Separate Chains
B2B buyers confuse these constantly. You need both, but they come from different places:
| Type | Covers | Issued By | What to Verify |
|---|---|---|---|
| Safety cert (ETL/UL/CSA / CE LVD+EMC / CCC) | Electrical shock, fire, mechanical, abnormal operation | NRTL (Intertek/UL/TÜV/SGS) or national authority | File number searchable on the certifier’s public directory. Photo of the actual label with file number |
| Performance claim (CADR, CCM, HEPA class) | How wellit cleans air | CMA/CNAS lab (CN) or AHAM-recognized lab (US) | Lab name + accreditation number + sample model lock + test-condition transparency |
| Materials / Emissions (RoHS, REACH, CARB, OEKO for any textiles) | Hazardous substances, ozone, VOC off-gassing | Dependent on market | CARB attestation letter with model listed; not just a generic “compliant” statement |
A product can have a valid CE safety mark and a completely fake CADR claim simultaneously. Safety = “it won’t shock you.” CADR = “it actually cleans air.” You need both verified independently.
5. The Supplier Verification Playbook — 7 Checks That Expose Fake CADR
✅ Check 1: Demand the lab’s accreditation, then verify it yourself
- China: CMA certificate number → cross-check at the provincial market supervision bureau or cma.org.cn style directories
- International: CNAS logo + certificate number → searchable at cnas.org.cn
- If the report is from SGS/Intertek/TÜV → call that office branch and confirm report number
✅ Check 2: Confirm the tested sample = your production unit
The report must show:
- Your model number (not “AP-001 generic”)
- Photos of the unit with your filter part number
- Fan speed setting explicitly stated (“maximum speed” / “highest fan gear”)
- Filter description: H11/H12/H13 with ISO 29463 or EN 1822 basis
Ask: “Can you ship us 3 random production units — not hand-picked — and we’ll send one to [independent lab of our choice] for a blind CADR check?”A real factory says yes and builds it into the PO. A fake one stalls.
✅ Check 3: Do the sanity math
Rough reality check:
Approx. max CADR ≈ Chamber airflow (m³/h) × Filter efficiency
A 200mm diameter fan at max speed = roughly 300–500 m³/h mechanical airflow
Through an H13 pleat pack = ~95–99% efficient → CADR ~280–490
If they claim CADR 800 on a compact tower with a single 120mm fan...
the claim is physically impossible regardless of what the PDF says.✅ Check 4: Audit the filter supply chain (this is where CADR silently dies)
Your CADR isn’t set by the motor — it’s set by the filter media. If the factory switches from H13 borosilicate glass fiber to cheap PET “HEPA-like” mesh to save $2/unit, CADR at 0.3 µm drops from 99.97% to 60–80%, and the real-world PM2.5 performance collapses.
Require:
- Filter media manufacturer name (e.g., HV, Ahlstrom, Hollingsworth & Vose, or known CN makers like Zhuge/Shenzhen Xingyuan) — not “high-efficiency media”
- HEPA class per ISO 29463-3 or EN 1822 with test report on the media roll, not just the assembly
- Carbon spec: weight in grams, type (pelletized vs. granular vs. impregnated sheet), and CCM F-rating
✅ Check 5: Check for the silent killer — pressure drop & noise at claimed CADR
A factory can hit a big CADR number by over-speeding the fan — but then:
- Noise blows past 55–65 dB(A) at max → unsellable for bedrooms
- Motor lifetime crashes → warranty claims spike
- Pressure drop across a cheap filter at high flow → motor whines, bearings fail
Legit suppliers will give you a CADR vs. dB(A) vs. power curve for each speed setting. If all they have is one “CADR 800” number and no noise/power breakdown → incomplete data.
✅ Check 6: Watch the “CCM skip”
Plenty of factories will gladly sell you a unit that tests great Day 1 but whose filter loads up in 3 weeks because it’s under-media’ed.
| Grade | Particulate CCM (P) | Formaldehyde CCM (F) | Practical Meaning |
|---|---|---|---|
| P1 / F1 | 3,000–5,000 mg | 300–600 mg | Budget — replace filter monthly |
| P4 / F4 | >60,000 mg / >150,000 mg | >1,500 mg | What you want for consumer products |
Ask for the CCM test page in the same report as CADR — if it’s missing, they likely didn’t run it or it graded poorly.
✅ Check 7: Pre-shipment performance spot-check
Write this into your QC plan:
“Buyer may randomly select 1 unit from every 500-unit batch for third-party CADR verification. If measured CADR falls below 85% of spec’d value, entire batch is subject to re-inspection at supplier cost.”
6. RFQ Spec Template — CADR & Performance Section
AIR PURIFIER RFQ — PERFORMANCE & CERT VERIFICATION
【Target Performance — Minimum Acceptable】
■ Particulate CADR (smoke/PM): ≥ ___ m³/h @ max fan speed (measured per ___ )
■ Formaldehyde CADR (if claimed): ≥ ___ m³/h @ max fan speed
■ CCM particulate grade: P___ (min P3 / prefer P4)
■ CCM formaldehyde grade: F___ (min F3 / prefer F4)
■ Noise @ max speed: ≤ ___ dB(A) | @ sleep speed: ≤ ___ dB(A)
■ Effective coverage area claim: ___ m² (show calc: CADR × 0.07–0.1)
【Filter Spec — Lock These in Writing】
■ HEPA class: ISO 29463-3 ___ (e.g. H12 / H13) — require media MFG name
Acceptable media sources: [list or "buyer approves"]
■ Carbon: ___ g minimum (pelletized / granule) — no "carbon-coated foam" substitute
■ Filter frame seal: EVA gasket or urethane pour — no unsealed friction-fit only
【CADR Test Requirements】
■ Test standard: □ GB/T 18801-2015 (or later) □ ANSI/AHAM AC-1 □ Both
■ Lab requirement: Must be CMA-accredited (CN) or AHAM-recognized (US)
Preferred labs: [SGS / Intertek / TÜV / 中国家用电器检测所 / CTI — your pick]
■ Report must include:
└─ Chamber volume & temp/RH log
└─ Sample model #, fan speed setting, filter part # + photo
└─ Particulate CADR, Formaldehyde CADR, CCM grade, noise, power
└─ Lab CMA/CNAS stamp + cert # + engineer signature + report date
■ AHAM Verifide (if NA market): Model MUST appear in AHAM public directory
【Safety & Materials Certs Required】
□ Electrical safety: [UL / ETL / CE LVD+EMC / CCC] — file # searchable
□ Ozone emission report (especially if ionizer/electrostatic): ≤5 ppb for CARB
□ RoHS / REACH declaration
□ CARB attestation (if shipping to California)
【Change Control — Critical】
■ Filter media source, carbon weight ±10%, motor/fan model: ALL require
written re-approval + re-test at supplier cost before shipping changed units.
■ Supplier warrants tested CADR ±10% holds across production within filter
service life — not just on Day 1 golden sample.7. The Cost Reality — What You’re Paying For
| Component | What Drives CADR | Where Factories Cut to Fake It |
|---|---|---|
| Motor/fan (~25–35% of BOM) | Brushless DC, 200–270mm impeller, good aerodynamic design | Swap to cheaper AC shaded-pole or underpowered BLDC → airflow collapses under filter load |
| Filter media (~20–30%) | Correct H13 glass fiber pleat depth, media area, proper seal | Replace with PET “HEPA-style” mesh or thin 10mm pleats → efficiency drops, pressure rises |
| Carbon bed (~10–20% if included) | 300–600g+ pelletized carbon, proper thickness | Foam dipped in carbon slurry — looks black, adsorbs nothing long-term |
| Lab testing | 1,500–4,000 for real CMA/CNAS full report | $150 “internal report” from a no-name shop that’ll print whatever number you need |
The pattern: When a quote comes in 20%+ below market and they’re waving a “CADR 700” report — the savings came out of the filter media, the motor, and the lab integrity. You’ll pay for it in returns and bad reviews.
Bottom Line
Don’t buy a CADR number. Buy a verifiable test chain: accredited lab → locked filter spec → production sample integrity → performance warranty. The certificate is only as good as the lab behind it and the factory’s discipline in not swapping media after the test passes.