The lab night-keeper: guarding the data while the machines sleep

In the materials lab of a research institute, some experiments run for months at a stretch.

High-temperature creep tests, fatigue tests, corrosion observation — none of them can stop: one interruption and months of data are void. So the lab needs a "night keeper": on duty overnight, recording every two hours the state of the equipment, the temperature and humidity of the room, whether the data curves look right.

The night keeper is Lao Ji, a lab technician, twenty-six years at it.

His shift is known as the ten most important hours in the institute. In the small hours the lab is so quiet you can hear the hum of the instruments. Lao Ji makes a round every hour: feeling the equipment casings for heat, checking lines for leaks, verifying the curves on the screens — a sudden jump might mean a specimen has snapped, a machine has failed, or a sensor has gone wrong. Twenty-six years have trained him to diagnose by ear: when a testing machine doesn't sound right, he can tell from down the corridor.

He has averted disaster more than once. Once at two in the morning he heard something wrong in the compressor of a climate chamber, judged a failure coming, and acted early — inspection the next day found the compressor bearing already damaged. "If it had waited until dawn, a whole chamber of specimens at forty days would have been ruined." Another time he noticed data curves that were too beautiful — data too tidy is the suspicious kind — and traced it to a sensor come loose, distorting everything. The graduate students say: "Master Ji is the night keeper of the data. When he's on duty, we dare to sleep."

In the world of research, the daytime produces results and the night produces data. The night keeper's duty: that no researcher's long wait is betrayed by an accident.

Lao Ji's night shift is also the graduate students' late-night canteen. Chasing experiments, they often work past midnight; in his duty room there's a kettle of hot water and a few packs of biscuits. In the lab corridors at midnight, the young ones talk about life with him, and he counsels them in his own way: "Look at that fatigue tester. It takes a million load cycles to get a result. Doing experiments, earning a degree — it's the same. Don't expect things to come in one night. Keep watch night after night, and things will come."

Years after graduating, students still come back to see him. One became a professor; returning to the institute to give a talk, he began with this: "Every piece of data I present today — part of it was spotted at four in the morning, with Master Ji's eyes."

Lao Ji is fifty-five, and the institute moved him to days. After a month on the day shift he asked to go back to nights. "Days are lively, plenty of people. But the lab at night is another creature — so quiet you can hear the equipment breathing. I've listened for twenty-six years. I understand what they say."

In his duty room stands a small cabinet of "failed specimens" gathered over twenty-six years: test pieces whose curves were voided, coupons whose sensors fell off. People don't get it, but Lao Ji does: "These aren't junk. Every one of them reminds the young — data can lie, distort, go bad. The person keeping watch has to be more awake than the data." At the very back of the cabinet sits his old pocket watch from twenty years ago. However long the night, time itself must not go wrong.