Developing a Repeatable UV Cure Process

Irradiance versus dose, and why the difference matters

Irradiance is the power arriving at the adhesive, in W/cm². Dose (energy) is irradiance integrated over time, in J/cm². Adhesive data sheets specify one or both, and the system manuals give the first-order relationship: curing time [seconds] = energy [J/cm²] ÷ intensity [W/cm²]. Treat that as a starting point, not a law, for many chemistries the relationship is not linear, and doubling the intensity does not halve the cure time. The manuals note that the fastest, most complete reaction often comes from high peak intensity over a short exposure, which can outperform an equal or even larger dose spread over a longer time. Practical consequence: you cannot freely trade time against intensity and assume the same result. Qualify the combination you intend to run.

Qualify a window, not a point

Reactivity varies between adhesives and even between lots, so the manuals direct you to determine optimum settings experimentally and to consult the adhesive maker for cure times. During qualification, establish a minimum and maximum dose at which parts meet your acceptance criteria, then set the production recipe comfortably inside that window. The margin between your operating point and the window's edges is what absorbs normal drift, a process qualified at a single knife-edge setting will drift out of spec long before anyone notices.

Fix the geometry: distance is an intensity control

Irradiance rises and spot size shrinks as the guide tip approaches the work, on the SunSpot 2, tip-to-substrate distance is literally one of the three intensity adjustments, alongside the front-panel control (60–100% lamp power) and the guide's insertion depth in the receptacle. That is useful during development and dangerous afterward: an operator who holds the guide freehand is re-choosing the intensity on every part. So fixture it. Clamp the light guide at a fixed height and angle, lock the intensity knob and the guide-depth thumbscrew, and keep the tip clear of wet adhesive, resin that cures onto the quartz window becomes a permanent filter. On the SkyBeam, spot size and working distance are set by the lens choice, and each channel stores its own intensity and timer, so the recipe lives in the controller rather than in an operator's wrist.

Measure at the working plane, on a schedule

A radiometer reading only means something where the adhesive lives. Measure UVA irradiance at the guide tip's working position, not at the source, with a spot-cure radiometer, and take the first reading the day the system is new. The SunSpot 2 manual has you record initial intensity together with lamp hours in a maintenance log; every later reading is compared against that baseline. Put the check on a fixed calendar, log every value, and act on the trend: a roughly 20% drop from baseline is the working trigger to clean or service the guide, or swap the lamp. On multi-pole splitters, read every leg, and if legs diverge, rebalance by rotating the guide handle in 30-degree increments until they read even, then lock the thumbscrew.

Revalidate after every hardware change

Any change in the optical path changes the dose at the part, so re-measure and re-baseline whenever you:

• Replace the lamp. A fresh arc lamp restores full output, typically well above what the aged lamp was delivering, so yesterday's exposure time now over-doses. Re-measure and reset the recipe (see replacement lamps for spares).
• Swap or re-route a light guide. Transmission differs guide to guide, drops roughly 10–15% per meter of length, and suffers if the guide is bent near its minimum radius.
• Change a lens or spot size. On the SkyBeam, moving between the 6 and 12 mm lenses changes peak irradiance several-fold at the working distance.
• Re-seat a splitter. Guide rotation affects leg balance, rebalance and re-log every leg.

The habit that makes this painless: every hardware event gets a log entry with the new radiometer reading and the lamp-hour count, so the process file always reflects the machine that is actually on the bench.

Where drift comes from

• Lamp aging. An arc lamp typically lasts over 2,000 hours before output falls to 50% of initial, a slow slide your radiometer schedule is designed to catch. Frequent on/off cycling shortens lamp life, so leave the lamp on with the shutter closed for breaks under thirty minutes.
• Contaminated optics. Dust and adhesive vapors condense on guide tips, lamps, filters, and LED lenses, quietly attenuating output. Inspect and clean end faces per the manual.
• Guide degradation. Liquid guides solarize with UV and heat exposure; closing the shutter between cures materially extends their life.
• Airflow problems. Clogged air filters overheat an air-cooled unit, keep the SunSpot 2's rear clearance and filters clean.
• Geometry creep. Fixtures loosen and operators improvise. Locked knobs, thumbscrews, and a fixtured tip keep the qualified geometry qualified.

LED systems remove some of these terms, the SkyBeam regulates LED current to ±1%, needs no warm-up, times exposures from 0.1 s resolution up to 999.9 s, and keeps a life log per channel against a typical 20,000-hour LED life, but lenses still get dirty and geometry still creeps, so the radiometer schedule stays.

Common questions

How often should output be checked with a radiometer?

On a fixed schedule sized to your risk, a weekly spot-check suits many electronics benches, while validated medical and optical processes verify daily, plus a reading after every lamp, guide, or lens change. What matters most is that the interval is scheduled, the values are logged, and the trend is watched.

If I double the intensity, can I halve the exposure time?

Not reliably. The manuals note the intensity-to-cure-speed relationship is non-linear for some adhesives, even though high peak intensity for a short time is often the most effective cure. Any new intensity/time combination needs its own qualification against your acceptance criteria.

When should the arc lamp be replaced?

The lamp typically exceeds 2,000 hours before output falls to half of initial. With a radiometer, replace when readings approach the bottom of your qualified dose window, a ~20% drop from baseline is a common service trigger. Without one, the manual's fallback is replacement at roughly 2,000-hour intervals.

Do LED systems drift too?

Far less and far slower, typical LED life is 20,000 hours, current is regulated to ±1%, and per-channel life logs track accumulated hours. But dirty lenses, moved fixtures, and swapped guides affect an LED cell exactly as they do an arc cell, so the working-plane measurement routine doesn't go away.