Test protocol · version 1.0
Machine noise protocol
We measure A-weighted equivalent sound level (LAeq, 60 s) at 1 m from the machine with a calibrated Class 2 or better sound level meter, idle and while working, and correct for room background.
- Version
- 1.0 (September 25, 2026)
- Applies to
- diode lasers, galvo lasers, CO2 lasers, fiber lasers, fume extractors, enclosures, air assist, vinyl cutters, embroidery machines
- Results so far
- None. Testing starts with our first purchased machines.
What we measure
| Measurement | Unit | Better | Notes |
|---|---|---|---|
| LAeq while working, 1 m front | dBA | Lower | Main number. |
| LAeq idle (fans/pumps on) | dBA | Lower | What you hear between jobs. |
| LAeq while working, 1 m side | dBA | Lower | Some machines vent to the side. |
| LAFmax | dBA | Lower | Loudest moment (fast time weighting). |
Why this test
Air assist pumps, exhaust fans and embroidery heads can be loud enough to matter in an apartment, a shared studio or a market stall.
Noise ratings, where given at all, rarely state distance or method.
Equipment
We list instrument classes and specifications rather than brands, because the exact units are still being purchased. Each result page will name the exact model and serial-level calibration date.
| Item | Minimum spec | Why |
|---|---|---|
| Sound level meter | IEC 61672-1 Class 2 or better, A and C weighting, LAeq and Fast time weighting, data logging | Class 2 is the standard general-purpose accuracy class; Class 1 is used if available. |
| Acoustic calibrator | 94 dB at 1 kHz, IEC 60942 compliant | Field check before and after every session. A drift over 0.5 dB voids the session. |
| Tripod + windscreen | Microphone at 1.0 m horizontal distance, at the height of the machine’s work surface + 0.3 m | Fixed, repeatable geometry. |
| Measuring tape / laser distance meter | ±2 mm | Positions are measured, not guessed. |
Setup
- Quiet room, doors closed, other equipment off. Machine on its usual stand or table, at least 1 m from walls where possible.
- Measure background LAeq for 60 s with everything off.
- Calibrate the meter with the acoustic calibrator and record the reading.
Procedure
- Idle: machine powered, fans/pumps running as they would between jobs; LAeq over 60 s at the front position.
- Working: run the standard job for the category (KP-ENG-01 for lasers; a standard 10,000-stitch design for embroidery; a 300 mm vinyl decal for cutters); LAeq over 60 s at front and side.
- Repeat each measurement three times; report the median.
- Where the machine has fan/air-assist levels, measure default and maximum.
- Re-check calibration at the end of the session.
Controls
- Background must be at least 10 dB below the measurement. If it is 3–10 dB below, we apply the standard energy correction and flag the value; under 3 dB the value is not published.
- Same room, positions and table for every machine.
- Readings taken by the meter, not a phone app.
Scoring
The noise sub-score (0–10) uses a fixed scale so that scores stay comparable across categories and over time.
| Component | Weight | Based on |
|---|---|---|
| Working LAeq, front | 70% | dBA |
| Idle LAeq | 30% | dBA |
| Total | 100% |
Formula. Each component scores linearly from 10 at 45 dBA to 0 at 75 dBA (clamped). Weighted sum, one decimal.
Scores are a convenience. The raw measurements are always published next to them, and our picks explain which measurement mattered for which kind of seller.
What we publish
- Background, idle and working levels at each position, calibration readings and meter model (CSV).
- A photo of the measurement layout with distances.
Limitations
- A 1 m level in our room is not the same as sound power (ISO 3744-type methods). Rooms with hard walls will sound louder.
- Tonal whine (for example a small pump) can be more annoying than its dBA suggests; we describe it in words when we hear it.
Where results will appear
On every product page this protocol applies to, a lab slot like the one below stays empty until we have measured that machine:
noise_dba_1mLAeq while working, 1 m front
Measurement pendingWe'll measure this on a unit we buy ourselves and publish the raw data. Read the protocol for LAeq while working, 1 m front
Protocol changelog
When we change a protocol, we bump its version, explain why here, and re-run or flag any results measured under the old version.
| Version | Date | Change |
|---|---|---|
| v1.0 | September 25, 2026 | First public version, published before any machine was tested so readers can critique the method first. Not yet run on any machine. |
Other protocols
Sources
Every spec and number on this page links back to where it came from.