How to Measure Train Horn Decibels: SPL Meter vs Phone App

Every train horn maker prints a decibel number on the box, and almost none of them tell you how they got it. The good news: with a sub-$200 handheld SPL meter — or, within real limits, a free phone app — you can measure your train horn’s actual decibels yourself in an afternoon.

Why You Should Measure Instead of Trusting the Box

Advertised ratings in this market are famously inflated. Some vendors measure right at the bell opening where sound pressure peaks, some measure inside the bell, and a few use distances of an inch or less — methods that produce huge numbers that mean nothing at real-world distances. Others quote a momentary startup spike instead of sustained output. HornBlasters, one of the few companies that publishes its methodology, tests every horn at both 3 feet and 100 feet with calibrated professional meters, and flatly calls claims above 150 dB physically impossible for vehicle horns.

That’s why a horn honestly rated 135 dB at 3 feet often out-blasts a “155 dB” horn in the same parking lot: per HornBlasters, move a bell-opening measurement out to 3 feet and the real output can be 125 dB or lower. We broke down the marketing math in our guide to the advertised vs real-world decibel gap — this article is the follow-up: how to get your own trustworthy number.

The Right Tool: What SPL Meter Classes Mean

Real sound level meters are graded by an international standard, IEC 61672-1, into Class 1 (precision) and Class 2 (general purpose). The difference is tolerance: at the 1 kHz reference frequency, a Class 1 meter must hold ±1.1 dB while Class 2 allows ±1.4 dB, and the gap widens at frequency extremes — at 20 Hz it’s ±2.5 dB for Class 1 versus ±3.5 dB for Class 2, per Cirrus Research’s breakdown of the standard.

Here’s the part that should relax your budget: a Class 2 meter is plenty for horn testing. The FRA itself specifies Class 2 instruments for certifying real locomotive horns. Two widely sold handheld examples:

Meter	Range	Accuracy	Standard
REED R8050	30–130 dB	±1.4 dB	IEC 61672-1 Class 2 (Type 2)
Extech 407730	40–130 dB	±2 dB	General purpose

The REED runs about $189, offers A and C weighting, and has fast/slow response — everything this job needs. Notice the ceiling, though: both meters top out at 130 dB. A serious train horn kit can exceed that at 3 feet, which matters for how you’ll run the test (more below).

One more habit worth copying from the pros: the FRA requires an acoustic calibrator (IEC 60942) to verify the meter before use. Consumer meters drift too. If your meter came with a calibration cap or you can borrow a calibrator, check it — and if you can’t, at least treat single-decimal precision with humility.

Phone Apps: What NIOSH Actually Found

NIOSH — the federal occupational-noise research agency — evaluated 192 smartphone sound apps in 2014 and found most lacked the accuracy and features for occupational noise measurement. Out of that research came the agency’s own free NIOSH Sound Level Meter app for iOS, validated to within ±2 dBA of a reference Type 1 laboratory meter. Paired with an external calibrated microphone, it measured within ±1 dB of the reference over a 65–95 dB test range and meets Type 2 requirements of the meter standard.

Android is a different story. NIOSH declined to recommend any Android app: hardware varies enormously between handsets, and Android doesn’t let developers reliably bypass the phone’s built-in audio compressor the way iOS does. Independent testing published in Canadian Audiologist found variations of 3 dB or more from a reference meter were common across apps, with the NIOSH app the most consistent of the group.

Now the train horn problem: that validation range was 65–95 dB. A train horn at 3 feet can be pushing 140+ dB, and the tiny MEMS microphones in phones are built for speech, not artillery — at horn levels they saturate and compress, so the app under-reads by an unknown amount. Where that leaves you:

• A phone app is fine for relative comparisons — before/after a mod, horn A vs horn B from the same spot, or drive-by loudness at 50–100 feet
• A phone app is not fine for an absolute close-range number to stack against an advertised rating
• If you use one, use the NIOSH app on an iPhone — it’s free and the only one with published federal validation

Meter Settings That Change Your Number

Two people can measure the same horn with the same meter and report numbers several dB apart, purely from settings.

Weighting. A-weighting (dBA) discounts low frequencies roughly the way your ear does; C-weighting is nearly flat and reads higher on deep-toned horns (train horns play low-frequency chords — see our trumpet tuning guide). The FRA specifies A-weighting, and honest manufacturer ratings use it too. Measure in dBA.

Response. Fast response averages over 125 ms; slow averages over 1 second. The FRA uses slow response because a locomotive certification blast is sustained for 10 seconds. For a short aftermarket blast, slow smears the reading down — use fast response with max hold.

Max vs peak. Some meters also show a “peak” value that captures instantaneous pressure spikes. Peak always reads dramatically higher and is one of the inflation tricks HornBlasters calls out. The honest number is the maximum A-weighted fast reading, not the peak.

How the FRA Measures a Real Locomotive Horn

The federal rule for actual locomotives, 49 CFR § 229.129, is a masterclass in specifying conditions. A locomotive horn must produce 96–110 dB(A) at 100 feet forward of the locomotive. The microphone sits 15 feet above the top of the rail (4 feet for low-mounted horns), within 20 degrees of the track centerline. The meter must be an IEC 61672-1 Class 2 instrument, checked against an IEC 60942 calibrator, set to A-weighting with slow response, averaging over 10 seconds. Testing is only valid between 32 °F and 104 °F, 20–95% humidity, and wind no more than 12 mph.

You don’t need to replicate all of that. The lesson is what a citable measurement looks like: distance, height, weighting, response, and weather are all pinned down. A marketing claim of “155 dB” with none of those conditions attached tells you nothing.

Step-by-Step: Measuring Your Own Horn

1. Air the system up to full cut-out pressure and let the compressor stop. A half-charged tank blows a weaker, quieter note.
2. Pick a large empty paved area at least 50 feet from buildings and walls — reflections add decibels — on a calm, quiet day.
3. Set the meter: dBA, fast response, max hold, high range (e.g., 60–130 dB).
4. Position the microphone 3 feet from the horn bells, level with them and pointed at them. A tripod beats a hand, and keep your body out of the sound path.
5. Put on hearing protection. At 3 feet from a train horn this is not optional.
6. Blast for a full 3 seconds. Record the max-hold value, reset, and repeat three times; use the middle value.
7. Repeat the same procedure at 30 feet and, if you have room, 100 feet.
8. Log everything: distance, weighting, response, tank pressure, and rough weather.

If the meter pegs at 130 dB at 3 feet, congratulations — your horn genuinely exceeds the consumer-meter ceiling. Don’t guess beyond it; use your 30-foot reading as the comparison number and say so.

Reading Your Results Honestly

Expect your number to come in below the box. Sound falls off steeply with distance — HornBlasters’ own example is a horn measuring 149 dB at 3 feet dropping to roughly 133 dB at 30 feet — and most advertised figures were never 3-foot numbers to begin with. A 3-foot reading in the low-to-mid 130s dBA means you own a genuinely loud train horn, whatever the sticker said; for what those numbers actually feel like to the ear, see how loud 150 dB really is, and for what your readings mean down the road, see how far a train horn can be heard.

Two closing rules. First, only compare readings taken at the same distance with the same settings — a 3-foot dBC peak and a 30-foot dBA max are different universes. Second, your consistent backyard method is worth more than any spec sheet: it’s exactly the tool you need to test whether a pressure bump or new valve actually made your horn louder.

Sources

• 49 CFR § 229.129 — Audible warning device (Cornell LII) — FRA 96–110 dB(A) at 100 ft requirement, microphone placement, Class 2 meter and IEC 60942 calibrator, A-weighting/slow response, weather limits
• HornBlasters: Why Fake Decibel Ratings Mislead Buyers — 3 ft/100 ft testing methodology, 149 dB → ~133 dB falloff example, bell-opening and peak-spike inflation tactics, 135 vs 155 dB comparison
• NIOSH Sound Level Meter on the App Store — app platform, ±2 dBA validated accuracy, A/C/Z weighting, free
• CDC/NIOSH Science Bulletin: New NIOSH Sound Level Meter App — ±1 dB accuracy with external calibrated microphone over the 65–95 dB test range
• CDC/NIOSH Science Bulletin: How Accurate Are Smartphone Sound Measurement Apps? — 192 apps evaluated in 2014, most inadequate for occupational use
• Canadian Audiologist: How Accurate Are These Smartphone Sound Measurement Apps? — 3 dB+ variability across apps, MEMS microphone limits, iOS vs Android compressor access
• Cirrus Research: Class 1 vs Class 2 Sound Level Meters — IEC 61672-1 tolerances: ±1.1 vs ±1.4 dB at 1 kHz, ±2.5 vs ±3.5 dB at 20 Hz
• REED R8050 Dual Range Sound Level Meter — Type 2 rating, 30–130 dB range, ±1.4 dB accuracy, A/C weighting, fast/slow response
• REED R8050 at Global Test Supply — $189.00 price for the standalone meter
• Extech 407730 Digital Sound Level Meter — 40–130 dB range, ±2 dB accuracy, 0.1 dB resolution

Keep reading

• Advertised vs real-world decibels: why “150 dB” horns aren’t
• Decibels explained: how loud is 150 dB really?
• How far away can you hear a train horn?
• How to make a train horn louder: what actually works
• Farbin dual trumpet “150 dB” kit review — a case study in inflated ratings