He Shut the Belt Off. That's When It Let Go.

▶️ Prefer to watch? Every story’s a short video → https://www.youtube.com/@GetSafetyIntel · 📣 Got a near-miss? Share it anonymously (QR + link at the bottom).

Three years ago today a belt foreman with 13 years underground did everything right — he spotted the hazard first and called to kill the belt. The machine had been quietly failing for a month, and the one safe act on that shift is exactly what turned it loose. This one’s about “off” and “safe” not being the same thing.

At about 6:50 a.m. on August 30, 2023, Aaron Haley, 34, a belt foreman with more than 13 years of mining experience, was killed at Warrior Met Coal’s No. 4 Mine in Brookwood, Alabama, when a component of a longwall belt take-up — the bridle — broke free and struck him.

Here’s the floor version. A longwall belt take-up keeps the conveyor belt tight using a winch cable pulling on a moving carriage — set that day to 33,800 pounds of tension. The bridle is the steel link that connects that loaded cable to the carriage. It’s a coiled spring the size of a room, and it’s under load every second the belt is tight.

The machine had been telling people it was hurt for a month. About four weeks earlier, a pin at the winch connection slipped out and the cable backlashed. Instead of repairing the connection to take the manufacturer’s pin, the crew was directed to install a different pin — roughly twice as long — that stuck several inches below the bridle. Two weeks before the accident the winch cable broke and was replaced; the long pin went right back in. Then on August 28, while winding excess belt, that long pin hooked a cross brace and ripped the track frame apart, weakening the bridle at its connection points. People knew. The report is blunt: management was aware of the damage caused by the longer pin but did not correct the hazard.

On the accident shift, a crew slid in new rails but hadn’t bolted them down yet. Haley’s crew walked in behind the guards, and Haley saw an orange glow — a damaged, overheating roller — and possibly flames near the carriage. He did the right thing. He yelled, “Shut the belt off.” A laborer ran and hit the stop. And as the loaded belt wound down, the shift in stress on that already-weakened bridle was the last straw. It broke free, recoiled through the chain-link guards, tore them loose, and struck Haley. Three other miners were knocked down and pinned under the guards. Haley was flown out and could not be saved.

The rule it shaped — and it’s already on the books: MSHA cited a 104(d)(2) order for violating 30 CFR 75.1725(a) — “Mobile and stationary machinery and equipment shall be maintained in safe operating condition; machinery or equipment in unsafe condition shall be removed from service immediately.” That “(d)(2)” tag means unwarrantable failure — aggravated conduct beyond ordinary negligence — because the damage was known and left in service. One sentence, cold: equipment that is known-damaged doesn’t get watched, it gets locked out and taken out of service. The manufacturer’s pin wasn’t red tape. It was the load path.

The HOP read: the person closest to the work saw the hazard first and acted correctly. The failure wasn’t Haley’s stop command — it was a month of a known-damaged machine being nursed along instead of pulled from service. Blame the last person to touch it and you’ll bury the real lesson with him.

This one lands almost entirely on the S. (E) It was a mechanical struck-by underground — no spill, release, or offsite contamination recorded; contained to the belt entry. The environmental story of a coal mine lives elsewhere (methane and respirable dust management), not in this failure. (H) The health impact was acute and traumatic: one man killed by blunt force, three more knocked down and engulfed in a heavy dust cloud when the guards tore loose. The long-tail health hazard in coal — respirable coal-mine dust and black lung — is real and MSHA-regulated, but it’s a different fight than the one that killed Haley. (H+S together) The through-line is the same phrase in every category: safe operating condition. The maintenance-integrity failure that recoiled a cable here is the same class of failure that, in a chemical unit, poisons a creek or gasses a town. Different consequence, identical root: a known-bad condition left in service.

Silica you can’t see, willfully ignored (Watsontown, PA — Aug 26, 2026). The Department of Labor cited a Pennsylvania brick manufacturer with three willful and four serious violations for knowingly exposing workers to unsafe levels of respirable crystalline silica dust, proposing $496,000 in fines. So what for safety leaders: silica (29 CFR 1910.1053) is the slow-motion version of today’s story — the damage is invisible and cumulative, and willful means somebody had the monitoring data and let the work continue. If you cut, grind, or fire masonry, your exposure assessment and dust controls are the whole ballgame. (Source: OSHA news release, osha.gov)

A second kettle incident says the first lesson didn’t take (Janesville, WI — Aug 20, 2026). DOL cited a Wisconsin food manufacturer after two separate incidents of industrial kettles releasing hot steam and liquid — three workers dead, two seriously injured — finding the company repeatedly exposed workers to the hazard. So what: in HOP terms, a repeat event is the loudest signal there is. How an organization responds to the first failure decides whether there’s a second. If your corrective action after an incident is retrain and reset, you haven’t changed the work. (Source: OSHA news release, osha.gov)

A millwright reached toward a guard on a belt take-up to clear what looked like a jammed roller. A second tech grabbed his shoulder — “winch is still loaded, back out.” They’d killed the drive but not dumped the take-up tension; the cable was still carrying thousands of pounds. Nobody got hurt. Here’s the blameless read: the guard was right there, the fix looked two-seconds-easy, and “the belt’s off” felt like “it’s safe.” That gap between off and safe is where good people get caught — not because they’re careless, but because a de-energized drive and a de-tensioned machine look identical from three feet away. The save was a coworker who knew the stored energy was still live and said something.

Got a fail or a near-miss? Hit reply, or scan the QR below — we’ll feature it anonymously — no names, no company, no blame. Just the lesson, so the next crew doesn’t learn it the hard way.

Walk to one conveyor take-up, tensioner, or counterweight on your site today and ask the person who works on it two questions: how do you know the stored tension is released before you reach in, and where is that step written down? If the answer is "we shut the belt off," you've found the gap Brookwood had. Put the de-tensioning step into the lockout procedure by name — 29 CFR 1910.147(d)(5) requires stored or residual energy to be relieved, disconnected, restrained, or otherwise rendered safe, and 30 CFR 75.1725(c) says no repairs until the machinery is blocked against motion. (Current as of August 30, 2026 — verify at ecfr.gov.) "Off" is a switch. "Safe" is a verification.

Watch: youtube.com/@GetSafetyIntel — subscribe so you never miss one.

Share your near-miss, anonymously: scantoconfess.com — or scan the QR above. No names, no company, no blame.

Forward this to one person on your crew. The lesson only works if it travels.

Please stay Safe & Hydrated!!!