UV Troubleshooting

UV Curing Troubleshooting: 8 Costly Mistakes and How to Catch Them

The eight failures that cause soft cures, over-exposure, and scrap in UV lines, each with the symptom, the real root cause, and the number that catches it before you ship bad parts.

The most frequent UV curing failure is confusing dose with irradiance. Symptom: parts that cured yesterday are tacky today at the same lamp setting. Root cause: irradiance is power (mW/cm2), dose is energy (mJ/cm2 = irradiance times seconds under the lamp). A part needing 800 mJ/cm2 at 400 mW/cm2 needs 2.0 seconds of exposure. If you speced 800 as an irradiance target and ran a 0.5 second dwell, you delivered only 200 mJ/cm2, a quarter of spec. Check the two numbers on the UV Dose and Cure Margin calculator side by side before you blame the resin.

Second: unit errors between mJ and J, and between cm2 and in2. Symptom: a dose that reads 10x or 100x off a supplier datasheet. Root cause: 1000 mJ/cm2 equals 1 J/cm2, and 1 in2 equals 6.45 cm2, so an irradiance quoted per square inch is 6.45x smaller per cm2. A spec of 1.2 J/cm2 is 1200 mJ/cm2, not 1.2. The fix is to force everything into mJ/cm2 and mW/cm2 before any math, then verify the exposure time on the UV Exposure Time calculator lands in a plausible 0.3 to 5 second range.

Third: ignoring lamp decay. Symptom: the recipe was validated in January and fails in June with no setting changes. Root cause: mercury arc lamps lose output steadily, often dropping to 70 to 80 percent of new by 1000 hours, and LED heads drift down over their L70 life of 20,000 hours or more. If new output was 500 mW/cm2 and you are at 380 mW/cm2, dwell that once delivered 950 mJ/cm2 now delivers 722, below an 800 spec. Track it with UV Lamp Life Remaining and UV Intensity Decay vs New, and rebuild the belt speed as output falls.

Fourth: measuring irradiance at the wrong plane. Symptom: the radiometer reads 600 mW/cm2 but parts undercure. Root cause: irradiance falls off with distance, roughly following an inverse relationship that can drop intensity 30 to 50 percent when the part sits 25 mm below the measurement height. A puck read at the lamp face does not represent a part 40 mm down in a fixture. Use UV Irradiance at Part with the actual working distance, and re-measure at part height, not lamp height, whenever you change fixtures or web thickness.

Fifth: radiometer drift and mismatched bandwidth. Symptom: two identical lines report doses 20 percent apart. Root cause: radiometers drift 5 to 15 percent between annual calibrations, and a UVA-band meter reads a different number than a UVV or UVC meter under the same lamp. Comparing a 365 nm LED reading on a broadband mercury meter is meaningless. The fix: calibrate yearly, log the correction factor, and apply it with the UV Radiometer Reading Calibration Correction tool. A meter reading 15 percent high hides a 15 percent dose shortfall.

Sixth: belt speed and dwell time math run backwards. Symptom: doubling line speed to hit throughput quietly halves the dose. Root cause: dwell time under a lamp of illuminated length L at speed v is L divided by v, so a 150 mm window at 30 m/min gives only 0.30 seconds of dwell. Bump the line to 60 m/min and dwell drops to 0.15 seconds, halving delivered energy. Before any speed change, re-run UV Conveyor Belt Speed and UV Dwell Time Under Lamp so the dose stays above spec, or add lamps to compensate.

Seventh: shadowing and non-uniform dose treated as a single average. Symptom: edges cure hard while recessed features stay gummy. Root cause: a reported average of 900 mJ/cm2 can hide a min-to-max spread of 2:1, leaving shadowed pockets at 450. Three-dimensional parts, deep menisci, and pigmented inks all block or absorb UV. Map it with UV Dose Uniformity and UV Shadowing Risk Score, and design to the minimum dose, not the average. A world-class line keeps uniformity within plus or minus 10 percent; anything past plus or minus 25 percent will produce intermittent soft spots.

Eighth: no cure margin, so the process runs at the edge of the cliff. Symptom: yields swing with ambient temperature, batch, and lamp age for no obvious reason. Root cause: running exactly at the minimum cure dose leaves zero headroom, so a normal 10 to 15 percent variation drops random parts below threshold. Build in margin: target 1.3 to 1.5 times the minimum validated dose and confirm it on the UV Dose and Cure Margin and UV Dose Margin calculators. If minimum cure is 700 mJ/cm2, run 1000, and you absorb lamp decay and meter drift without a single tacky part.

Catch these early by re-validating whenever any input changes: new resin lot, new fixture height, lamp swap, or line speed adjustment. Log irradiance at part height, dwell time, and delivered dose after every change, and confirm the cure window on the UV Cure Process Window calculator. The recurring pattern behind almost every UV defect is a single silently changed variable, distance, decay, or speed, that nobody re-entered into the dose math before the part shipped.

Published 2026-07-01.