English
Customer Background
The customer is a fiber optic connector and passive component manufacturer supplying LC, SC and MU patch cord assemblies to data centre integrators in Europe. Their plant runs three assembly cells where ferrules, housings and boots are bonded with two-component UV epoxy, followed by visual inspection and pull test. Annual output is around 1.2 million connectors, shipped in small call-off batches.
The Challenge
Before the upgrade, each cell used one mercury spot lamp with a mechanical shutter, which set a hard ceiling on output and quality:
- Long cure cycle. The epoxy needed 90 seconds under the lamp before the fixture could be released, so each operator could only finish about 120 connectors per shift.
- Heat damage. Mercury lamp heat distorted thin wall ferrules and softened the boot adhesive, producing ovality rejects on roughly 3 percent of parts.
- One wavelength for two adhesives. The plant ran a 365nm surface tack epoxy and a 395nm deep cure epoxy on the same line, so operators had to swap lamps and re-fix the parts mid-cycle.
- No process data. Exposure settings lived in operator memory, so shift variation was untraceable when a batch failed.
- Lamp maintenance. Mercury bulbs were replaced every 1,000 hours, each change needing a fresh calibration and a scrap run.
The Solution
We supplied a multi channel UV LED spot curing system with four independently controlled heads, air cooled and driven from a single LCD controller. Three heads sit on the existing fixtures and the fourth serves the repair bench. Each head carries its own wavelength and its own timed profile, so the 365nm tack step and the 395nm deep cure step now run in the same fixture without a lamp change.
| Configuration Item | Setting Supplied |
|---|---|
| Control channels | 4 independent channels, each with own intensity, timer and trigger |
| Wavelength per head | 365nm, 385nm, 395nm, 405nm (matched to the customer adhesive data sheets) |
| Output intensity | Adjustable, calibrated per channel at installation |
| Trigger mode | PLC / I/O trigger from the cell foot switch, with manual override |
| Cooling | Air cooling, no chiller or compressed air line required |
| Control interface | LCD display with stored process recipes per product family |
| LED lifetime | Above 20,000 hours, no bulb replacement schedule |
Implementation
- Process audit. We measured the existing dose on the ferrule bond and the boot bond, then mapped each adhesive to the wavelength that cured it fastest.
- Fixture integration. The three production heads were mounted on the existing pneumatic fixtures, so no mechanical redesign was needed.
- Recipe build. Each connector family received a stored recipe with channel intensity, cure time and trigger mode, protected from operator edits.
- Operator training. Two days on site covering recipe selection, daily intensity check and light guide cleaning.
- Validation. A 500 piece run per family compared pull strength and ovality against the mercury baseline.
Results
| Metric | Before (Mercury Spot) | After (4 Channel UV LED) |
|---|---|---|
| Cure cycle per connector | 90 seconds | 25 seconds |
| Output per operator shift | 120 pieces | 310 pieces |
| Ovality and heat rejects | 3.1 percent | 0.8 percent |
| Lamp changes per year | 18 bulb replacements | 0 (20,000 hour LEDs) |
| Energy draw on the curing step | Baseline | About 60 percent lower |
| Process traceability | Operator memory | Stored recipes and channel logs |
Why It Works
Two things drive the result. First, independent channel control removes the compromise of one lamp for two adhesives: the 365nm head cures the surface tack while the 395nm head drives the deep section, and both finish inside one fixture load. Second, LED spot output puts the energy on the bond line instead of heating the whole part, so thin wall ferrules keep their roundness and the epoxy reaches full hardness faster than it did under a broad mercury beam.
The stored recipes also removed the human variable. Shift-to-shift cure time differences disappeared from the pull test data within a month, letting the quality team move from full batch inspection to sampling.
Customer Feedback
"We were ready to buy a second mercury line to keep up with orders. The four channel spot system gave us the capacity from the existing benches instead, and the ferrule distortion problem simply disappeared. The stored recipes mean a new operator runs the same cycle as a fifteen year veteran."
Production Engineering Manager, fiber optic connector plant, Europe
Summary
A four channel UV LED spot curing system replaced a mercury spot lamp on a fiber optic connector line, cutting the cure cycle from 90 to 25 seconds, raising output per shift from 120 to 310 pieces and reducing heat related rejects from 3.1 to 0.8 percent. If your line runs more than one UV adhesive, or cures heat sensitive parts on a short cycle, send us your adhesive wavelength, spot size and target cycle time.