The Grit Chamber Entry That Went Bad Four Minutes In

First sixty seconds:

1. Give a clear, unambiguous order to evacuate - "Get out now, come up the ladder" - and keep talking to the entrant so you have a continuous status. Do not discuss, do not ask him to finish the task.
2. Take up the retrieval line as he climbs, keeping it taut but not hauling him off the ladder. If he becomes unable to climb, haul him out with the winch/tripod - that is exactly what the retrieval system is for.
3. Do not leave the opening. The attendant's post is the opening; abandoning it to go find a phone removes the only person who knows where the entrant is.
4. Simultaneously raise the alarm by the means that works - air horn, shouting to a second worker, plant radio on a different channel, or a runner. If your radio cannot reach the control room, the pre-entry plan should already have identified an alternate; if it did not, that is a finding for the investigation.
5. Call 911 and state confined space, hydrogen sulfide, one entrant with the exact location and access route.

What must not happen:

• The attendant must not enter the space. Most confined-space multiple fatalities are would-be rescuers. The atmosphere that overcame a man with a monitor will overcome you in the same number of breaths.
• No one else goes in without SCBA, a harness, a line, and a second attendant. At 22 ppm and rising, entry is a respiratory-protection operation, not a rescue by nerve.
• Do not send the entrant back down for equipment, and do not re-enter after an apparent recovery - recovery from H2S exposure does not mean the atmosphere has changed.
• Do not silence the monitor alarm.

The pre-entry test was not wrong so much as unrepresentative and one-dimensional.

1. Sampled in the wrong place. The reading was taken through a line from the top. Hydrogen sulfide is heavier than air and stratifies; the breathing zone at 4.8 m is a different atmosphere. The pre-entry test must be taken at the bottom, and at multiple depths.
2. A single snapshot was treated as a stable condition. One clean reading before entry does not establish that the space will stay clean. The rising trend during entry proves there was an active source.
3. The source was never identified or eliminated. Disturbed grit releasing trapped H2S, or - more likely here - a leaking isolation gate admitting sewage, which keeps generating sulfide. A slide gate that is closed, locked, and tagged is not necessarily sealing. Verify with a downstream observation, a secondary plug, or by confirming no flow arrives.
4. Ventilation was not verified. Twenty minutes through a 200 mm duct says nothing about whether air reached the bottom of a 4.8 m chamber with a corner. The blower intake may also be sitting near a sewer vent or idling truck exhaust.
5. Communications were not tested before entry. The radio cannot reach the control room from this location, and the entry proceeded anyway. That is a pre-entry checklist item.
6. No atmospheric monitoring of the entrant beyond the personal monitor's own alarm, and no escape respirator - a five-minute escape SCBA would have changed the outcome.
7. Single entrant, no backup entrant, no rescue plan with a trained team and a retrieval system anchored above the opening. The harness and line exist, which suggests a tripod arrangement, but nothing confirms a rescue-capable standby.
8. The alarm setpoints are questionable. An H2S alarm should be triggering well before 22 ppm; that is already above the 15-minute ceiling and represents a failed evacuation trigger.

Note also that H2S deadens the sense of smell at the concentrations where it is most dangerous - the entrant's nose is not a monitor.

First: reassess whether the entry is needed. The job is grit removal. Ask whether it can be done without entry - a vacuum truck working from above, mechanical removal, or deferring until the chamber can be fully drained and cleaned under better conditions. The safest confined space entry is the one you do not make.

Reclassification. The space has now demonstrated a hazardous atmosphere that developed during occupancy. It is not a routine entry with a clean pre-test; it is a permit-required space with a known hazard source, and the entry permit must reflect that.

Before re-entry:

• Eliminate the source. Prove the upstream gate seals - install a secondary plug or a blind, and confirm visually or by flow observation. If sewage can reach the chamber, gas will keep being generated no matter how long you ventilate.
• Remove the fuel. Clean out the accumulated grit and sludge that is releasing sulfide, working from above.
• Ventilate to the bottom and verify. Continuous forced air with the discharge at the bottom, exhaust at the top, and monitor at multiple depths before and throughout.
• Continuous monitoring with correct alarm setpoints - O2 alarms at 19.5% and 23.5%, H2S at 5–10 ppm, LEL at 10%. The current trigger point of 22 ppm is far too late.
• Test communications at the opening before anyone descends, and locate an alternate means that works from that spot.
• Full standby and rescue provisions: attendant at the opening, retrieval system anchored and rated, second entrant/attendant, and a rescue plan with a team equipped for the hazard.

Respiratory protection decision. If ventilation and source control cannot hold the atmosphere below the alarm setpoints, entry is an IDLH-adjacent operation and requires SCBA or supplied air with an escape pack - or it does not happen. Do not split the difference with a half-mask APR; particulate-and-acid-gas cartridges are not adequate for hydrogen sulfide in an oxygen-deficient or rising-concentration environment, and the exposure is not predictable enough to rate.

Finally: investigate and document. This was a near-miss with a clear lesson. Report it through your safety program, revise the pre-entry procedure to require bottom sampling, communication checks, and verified isolation, and retrain before the next entry.