Operations Director Yuki Tanakaruns operations across six SMT lines in Osaka and Penang. Here's what changed when anomaly detection, technician dispatch, and parts ordering started running as one closed loop — and what the numbers look like after two quarters.
Headline result
reduction in unplanned SMT-line downtime
across six SMT lines, two plants, first two quarters of operation.
Across our SMT lines, the closed-loop scheduling means a flagged component goes straight to the technician with parts on order. Our MTBF has nearly doubled since rollout.
Yuki Tanaka, Operations Director · Lumen Electronics
Lumen Electronics runs six surface-mount technology lines across two plants in Osaka and Penang. Each line carries roughly 180 components per board — pick-and-place heads, reflow ovens, stencil printers, AOI stations — and a single line stop mid-shift burns an entire board batch before the fault is even triaged.
Before rollout, the operations team was living in a maintenance ticket queue. A flagged anomaly usually surfaced as a downtime event first and a service call second: the technician was dispatched after the line was already bleeding output, parts were ordered reactively, and the next shift inherited a fix that had to spill over into their hours. Mean time between failures sat below 320 hours on the SMT lines, and roughly a third of those failures surfaced without any prior warning at all.
The cost pattern was the same story every quarter: emergency vendor expedite fees, weekend overtime on the maintenance crew, recovery runs on already-rushed batches, and a slow bleed of yield on boards that re-entered the line with incomplete rework documentation. Yuki's team had the data — current draw, vibration on conveyors, reflow profile drift — but no tooling to see it integrate.
After ProactiveIQ
Three shifts across the first 6 months of running ProactiveIQ on Lumen's SMT lines — measured against the same six lines, same board mix, same maintenance roster:
Edge AI detection flagged developing faults on pick-and-place heads and reflow fans 2–6 weeks ahead of failure. Replacement parts land on the bench with the technician, not after the line is dark. Across the six SMT lines, unplanned downtime hours fell roughly 62% in the first two quarters of operation.
Mean time between failures on the SMT lines moved from ~320 hours to ~640 hours within six months. Routine inspection windows are still scheduled, but condition-based work orders now carry the load — the calendar-driven overhaul cadence drops because the wear signal precedes the failure.
On the first SMT line that went live, recovered output plus eliminated weekend overtime covered the line's annualised ProactiveIQ cost within the first five months. Lines two through six rolled out on the back of that number — the operations team was the first ask, not the last.
See it on your floor
Walk us through your line count, your failure history, and the assets you most want protected. We'll come back with a scoped pilot on one of your lines and a projected downtime delta you can defend in front of your operations team.
Tell us about your plant, the buying process at your company, and any procurement deadlines you're working against. We'll reply within one business day with a confirmed slot and an engineer who can answer the technical questions on the call.