A hearing protector only protects to the labelled number when the noise survey, the derating math, and the on-head fit are all aligned; mismatches in any one of those three steps routinely leave workers above the OSHA 90 dB(A) 8-hour action level even with the device in place [S1][S7].
The reference below covers the four accepted derating methods (NRR, SNR, HML, Octave Band), the products that map to them, and the field checks that decide whether a hearing protector is installed correctly or just worn.
Step 1 — Quantify the noise before you open the catalogue
Every derating equation starts with a measured number, not an estimate, and the type of meter you carry dictates which method you can actually use downstream [S1]. A Class 2 sound level meter returning dB(A) is the minimum for the OSHA NRR route; dB(C) is required for SNR; both dB(A) and dB(C) are required for HML; and octave-band filters (typically 63 Hz to 8 kHz) plus the protector's APV table are needed for the Octave Band route [S1].
For broadband industrial noise above roughly 100 dB(A) — common in stamping, foundry knockout, and compressed-air venting — the Octave Band method is the only one that catches strong low-frequency components that defeat flat NRR or SNR figures [S1]. Sample the worker's actual position, log TWA over the shift, and record the dominant frequency bands; regulators and the ISO 4869 series treat this dataset as the input to the protector selection step [S1][S4][S8].
Step 2 — Match the derating method to the regulation, not the habit
OSHA jurisdictions use the NRR equation Lprot = L − (NRR − 7) / 2 for dB(A) measurements, and Lprot = L − NRR / 2 for dB(C) measurements — that −7 dB safety margin is built into the OSHA formula, so do not subtract it twice [S1]. EU and UK practice typically uses SNR or HML with an additional +4 dB real-world correction on top of the calculated ear level to compensate for poor fit, eyewear-glasses breakage of the muff seal, and intermittent removal [S1].
A worked example for a 102 dB(A) TWA in an OSHA site: a protector labelled NRR 30 derates to Lprot = 102 − (30 − 7) / 2 = 102 − 11.5 = 90.5 dB(A), which sits at the OSHA action level; the same protector under the SNR method (assume SNR 35, C-weighted LC = 105 dB(C)) gives Lprot ≈ 105 − 35 + 4 = 74 dB(A) protected level — the two routes will not agree, and that is normal [S1]. Pick the method your regulator recognises, not the one your spreadsheet likes.
Step 3 — Confirm the labelled rating actually applies to your product
Ratings are product-specific, not family-wide, and the part number on the box is what counts. The 3M E-A-Rcaps Model 200 (Fisher Scientific catalog 17-380-9) is a banded foam-tip protector rated NRR 17 dB, CSA Z94.2 Class BL, with ABS/polyurethane headband, yellow foam pods, and a one-size-fits-most fitment intended for intermittent use and for visitors in noisy areas [S3]. That NRR 17 figure feeds straight into the OSHA equation above: at 95 dB(A) TWA it gives Lprot = 95 − (17 − 7) / 2 = 90 dB(A) — borderline for a 90 dB(A) action-level workplace, so this is a visitor/short-duration product, not a primary engineering control substitute [S1][S3].
For higher-noise zones, spec higher-rated muffs or double protection (earplug + earmuff) and remember the rule of thumb that combined protection does not add arithmetically — derate each unit through the same equation and combine the protected levels, not the NRRs. Software-side hearing-protection apps (HearSafe, 2026 release) operate on a different layer: they cap headphone consumer output and do not substitute for industrial PPE selection [S2].
Step 4 — Fit the protector; the rating assumes a perfect seal
The +4 dB real-world correction in the SNR/HML/Octave Band methods exists because lab ratings assume trained subjects; field fit routinely loses 5–10 dB [S1]. For foam earplugs, roll the foam into a thin cylinder, reach over the head to pull the pinna upward and back, insert deep, and hold 30–60 s while the foam expands; for banded caps like the E-A-Rcaps Model 200, the band must sit under the chin (not behind the head or on top of a hard-hat shell) so the foam tips seal the canal entrance at a consistent angle [S3].
Run a qualitative fit-check at issue: speak — your own voice should sound muffled and 'boomy'; shake your head sharply — the device should not move or whistle; if the worker wears Rx spectacles, sidearms must not break the muff cushion seal. The Octave Band method with APV values is the only route that lets you spot frequency-specific leakage (typically the 500 Hz–2 kHz speech band, where most industrial hearing loss occurs) [S1][S4].
Step 5 — Decide when to issue, escalate, or replace
Issue the protector when calculated protected level is at least 5–10 dB below the workplace exposure standard (so a 90 dB(A) action level needs a protector derating to ≤80 dB(A) protected, factoring the +4 dB EU correction where it applies) [S1]. Escalate to dual protection or engineering controls when no single commercial protector derates below the standard, or when the noise is dominated by low-frequency content below 250 Hz — earmuffs physically cannot attenuate that band well, and Octave Band data will show it [S1].
Replace when the foam tips harden, crack, or fail a visual squeeze test (E-A-Rcaps pods are foam and degrade with skin oils), when the headband loses tension, or after any impact damage to the cup. Re-test the noise survey annually and re-train each worker at least once a year; a one-off fit-test session pays back inside the first recordable case it prevents. For site-level installation planning across other equipment rooms, the procedure walkthrough style in embedded part installation and the field steps in vertical lift module installation follow the same survey-then-fit-then-verify logic. For wider 2026 PPE spec sourcing context, the cost-band breakdown in safety barrier price and cost guide tracks the same supplier-quote pattern you'll see for hearing-protector contracts.
Spec-level background on the components involved: linear guide, and crossed roller guide.