The complaint
A residential building kept getting the same complaint: the elevator judders while it runs. It still worked, but the ride was clearly worse, and passengers were starting to worry about safety.
The maintenance team pulled the controller’s fault log: a few 'unstable running' notices, but no fault code. The standard checks — traction machine, ropes, guide rails, guide shoes, safety gear — found nothing.
You can feel it. You cannot find it.
Riding the car up and down, the technicians could feel the jolt only between certain floors, and not every time. It was like an elevator with hiccups: obviously something, impossible to point to. The team slowed the elevator down to keep the ride smooth — but the property manager wanted the root cause fixed, because an elevator that twitches now and then makes residents feel unsafe.
Let the data talk
The team installed a Dr. Elevage Fixed Monitor on the car top. It records acoustic and vibration data on every trip and syncs the data to the cloud, allowing technicians to review the measurements and narrow the inspection scope before returning to site. After 24 hours of continuous monitoring, a pattern appeared.
< 15 HzThe bright blocks all sit in the lowest band, at the same moments as the impacts in the trace above.
Three clues in the data
Several sharp, short impacts in the vertical direction, at fairly fixed positions in the shaft — which points to the guiding system.
Nothing meaningful follows in the horizontal direction. That rules out the car-side rails and shoes, and points to the counterweight side.
The energy sits below 15 Hz — right in the band the human body picks up most easily.
For context: under ISO 2631-1 the human body is most sensitive to vertical whole-body vibration at around 4–8 Hz, with sensitivity falling away above and below that band. The abnormal energy here sat under 15 Hz — right beside that peak, which is why passengers felt it so clearly and a routine check could not reproduce it.
Back on site, with a target
Taken together, the clues pointed at the counterweight system — frame, filler weights, guide shoes and guide rails. This time the team knew exactly where to look:
- Counterweight guide shoes — is the liner wearing evenly? Is the gap to the rail right?
- Counterweight guide rails — verticality, gauge and parallelism, checked with a laser rail alignment tool.
- Filler weights — is every block clamped, with nothing loose or shifted?
The cause
The counterweight ran on hollow guide rails, which are relatively flexible, and years of service had twisted them slightly. The guide shoes had been set too tight, so the rail’s tiny deformation turned into real resistance. That resistance made the counterweight judder, the ropes carried the judder to the car, and the passengers felt it.
Fixed, and verified
The technicians adjusted the guide shoes to open up the gap to the rail. A second round of car-top monitoring showed the previous abnormal vibration no longer present in the follow-up measurement, and a trial ride with the property manager confirmed the jolt had not come back.
Why this is the interesting case
It is the one that shows the method is not simply "put a sensor on the thing you are worried about". The car top is a vantage point onto the whole mechanical system, and a fault does not have to be under the sensor to be visible to it — which matters, because the counterweight is one of the least accessible parts of a working shaft.
You do not need a sensor on every component. You need one sensor in a place where every component leaves a trace.