I have spent most of the last 15 years servicing ammonia refrigeration systems in cold storage warehouses and food plants, and I learned early that an ammonia detector earns its keep long before anyone smells a leak. In my line of work, I do not think of detection as a box on the wall or a number on a controller. I think of it as the one part of the safety chain that speaks up before people start guessing. That matters more than most managers realize during a normal week, because a machine room can look calm right up until it is not.
Why placement matters more than the brochure
A detector can be a solid piece of equipment and still do a poor job if it is mounted in the wrong spot. I have walked into rooms with three compressors, two evaporative condensers on the roof, and a fresh sensor sitting where airflow kept it cleaner than the rest of the room. The unit was powered, visible, and almost useless for the leak path I was worried about. That happens more than people admit.
I usually start with the piping layout, the relief points, the valve stations, and the places where oil and vibration tend to make trouble after a few seasons. Then I look at air movement, because fans, louvers, and closed doors can change what the detector sees by a lot. On one service call last spring, a customer had one sensor mounted near the entrance because it was easy to access, but the real leak risk was tucked behind equipment where warm air pushed gas into a dead pocket first. I moved nothing that day, but I told them straight that the room layout was giving them false confidence.
How I decide between fixed and portable units
I like fixed detectors for constant coverage, but I still carry a portable unit every week because fixed systems do not answer every question I have during troubleshooting. A wall-mounted sensor tells me what the room is doing over time, while a handheld lets me chase a valve bonnet, a flange, or a questionable gasket without making the whole facility react. That distinction matters on older systems, especially the ones that have been patched, expanded, and repiped over ten or twelve years by different crews. One tool watches. One tool hunts.
When a plant manager asks where to compare options before a retrofit, I sometimes point them to détection d’ammoniac because it is an easy way to look at different detector styles without me trying to summarize every feature from memory. I do that only after I have seen the room, since a product page cannot tell you where condensate collects or which corner gets ignored during cleanup. In my opinion, the mistake is shopping by sensor type alone and treating the installation plan like a small detail. The detector is only half the decision.
What I pay attention to during testing and calibration
A detector that has not been tested is just a nice opinion. I do bump tests and calibration checks on the schedule the site calls for, and I take those dates seriously because sensors drift in the real world, not in the neat way manuals imply. Heat, washdown moisture, oil film, and plain neglect all change performance over time. I have seen units pass a visual inspection and still miss the mark badly enough to create a dangerous delay.
Most of the systems I work around use staged alarms, often with a lower threshold for warning and a higher one for response, but the actual setpoints should match the facility plan and local requirements. I am careful with that distinction because people love to repeat one number as if every machine room, freezer corridor, and roof penthouse should be treated the same way. They should not. A site with trained response staff, documented ventilation logic, and regular drills can make better use of a detector than a site that hangs a sensor and assumes the job is finished.
The small failures that turn into expensive ones
The worst detector problems I find are rarely dramatic on day one. A blocked splash guard, a neglected calibration date, a horn that was silenced and never fully restored, or a controller trend nobody reviews can sit there for months while everyone assumes the system is covered. Then a minor leak becomes a shutdown, and suddenly people are asking why the alarm did not tell a cleaner story. I have been in that room more than once.
One warehouse I serviced had a sensor mounted correctly and wired correctly, but the response chain around it was sloppy. The alarm reached the panel, though the night crew had never been shown what the first stage meant, who to call, or why opening one roll-up door could make the readings look better without fixing anything. That is the kind of gap that costs several thousand dollars in lost product and cleanup, even when the leak itself is not massive. Technology helps, but habits decide whether it helps in time.
What makes me trust a setup over the long haul
I trust a detector setup when the equipment, the room, and the people make sense together. That means I can trace the sensor placement back to actual leak points, verify the alarm path, review the test record, and ask a shift lead what happens at 2 a.m. without getting blank stares. Fancy screens do not impress me much. Clear response steps do.
I also pay attention to whether the detector fits the pace of the facility. In a plant with constant forklift traffic, wet cleanup, and a maintenance team juggling five priorities before lunch, a unit that is hard to service will get skipped sooner or later. A simpler detector with dependable access can be the better choice, even if the spec sheet looks less exciting. I have learned to respect the setup that gets maintained, because that one is the setup that usually works.
I still like walking into a machine room and seeing a clean installation, neat conduit, current test tags, and a team that knows what the alarm tones mean without checking a binder. That tells me the detector is part of the operation instead of a forgotten purchase from years back. If I were advising a peer tomorrow, I would say to spend less energy arguing over labels and more energy on placement, testing, and response, because that is where an ammonia detector proves what it is worth.