Why does the expected decline matter more than today's margin?

Because today's margin is the best it will ever be. Mercury output falls continuously, so the process is on a downward slope from the moment it is measured, and a single number tells you where it is rather than where it is heading. A 30% margin that becomes 10.5% before anybody next looks is a process that spends part of its life below its own floor, and no one finds out because nobody measured during that window.

Where do I get the decline figure?

From your own radiometer history, which is the reason to keep one. Published decline curves are nominal, individual lamps vary widely, and a lamp in a dirty housing declines faster than the curve predicts because the housing is fouling at the same time. If you have no history, start conservative and let the first few intervals calibrate it.

My margin expires before the next check. What should I change?

The interval, usually. Shortening a radiometer check from quarterly to monthly costs a few measurements; raising the dose costs line speed or a lamp purchase, permanently. The instinct is to slow the belt because it feels like a process fix, but it is paying forever for a scheduling gap. Raise the dose only if the interval is already as short as it can practically be.

Is a very large margin a problem?

Sometimes, which is why the page flags it rather than celebrating it. Excess dose embrittles some coatings and yellows others, and on a mercury line the infrared that accompanies the UV scales with dwell, so a heat-sensitive substrate can distort at a dose that is acoustically harmless. More often a very large margin is simply unclaimed line speed, which is worth knowing.

Does this cover reflector fouling or a dirty window?

Not properly, and it is a real limitation. A single decline percentage models the lamp ageing; quartz clouding and reflector fouling can outpace it substantially and are not separable here. If a measured dose falls faster than the decline figure predicted, suspect the optics before the lamp: the irradiance page splits those losses out.

Should the margin be measured at the worst point on the part?

Yes, and this page cannot tell you whether you did. A margin computed from a centre-line puck reading describes the centre line; if irradiance falls towards the edges of the belt or a shaped part shadows itself, parts of the surface are already below what the centre suggests. Run the uniformity and dose-mapping pages first, then compute margin against the coldest point that matters.

UV Curing calculator

UV Dose Margin Calculator (Today and at the Next Check)

Check whether a UV process has enough dose margin: today, and at the point you next plan to measure. Enter the required cure dose, the dose a radiometer actually measured, the minimum margin the process must hold, and how much lamp output you expect to lose before the next check. The calculator returns the margin now, the margin then, and the headroom over your floor at both. Margin is not a property of a process, it is a snapshot of one, and a line that passes on commissioning day can be failing weeks later without anything having been changed.

What this calculator does

Check the dose margin a process has today and whether it still clears the minimum after the lamp declines before the next radiometer check.
Use it for deciding how often a line needs a radiometer check, checking whether a commissioning margin will survive to the next scheduled measurement, justifying a shorter measurement interval instead of a slower line, reviewing whether a process has margin available to trade for speed, investigating a field failure by reconstructing the margin at the time of manufacture.
Check the dose margin a process has today and whether it still clears the minimum after the lamp declines before the next radiometer check.

Formula used

Dose margin today (%) = (measured dose − required dose) ÷ required dose × 100
Headroom today (points) = margin today − minimum acceptable margin
Dose after decline = measured dose × (1 − expected decline ÷ 100)
Margin at the next check = (dose after decline − required) ÷ required × 100
Headroom at the next check = margin at the next check − minimum acceptable margin

Inputs explained

Required cure dose from datasheet: The material's cure-dose requirement, in a named spectral band.
Measured dose at part surface: What a radiometer puck integrated on a real pass, in the same band. A measured dose, not a computed one.
Minimum acceptable process margin: The floor the process must stay above. This is a quality decision rather than a physical one, and 20% is a common answer.
Expected output decline before the next check: How much output the lamp will lose between now and the next radiometer measurement, from your own trend. Set it to zero to see today's picture alone.

How to use the result

Best suited to deciding how often a line needs a radiometer check, checking whether a commissioning margin will survive to the next scheduled measurement, justifying a shorter measurement interval instead of a slower line, reviewing whether a process has margin available to trade for speed, investigating a field failure by reconstructing the margin at the time of manufacture.
Models lamp decline only. A reflector that fouls or a window that clouds falls faster than a lamp ages and is not captured by a single decline percentage. Says nothing about uniformity. A margin computed on a centre-line reading can hide edges that are already below requirement. One-sided. It protects against undercure and treats large margins as merely informational, though overcure and substrate heat are real limits. Assumes linear proportional decline over the interval, which is a planning approximation rather than a lamp's true curve. A dose figure is a process setting; gel content or a rub test is cure evidence.

Common questions

Why does the expected decline matter more than today's margin? Because today's margin is the best it will ever be. Mercury output falls continuously, so the process is on a downward slope from the moment it is measured, and a single number tells you where it is rather than where it is heading. A 30% margin that becomes 10.5% before anybody next looks is a process that spends part of its life below its own floor, and no one finds out because nobody measured during that window.
Where do I get the decline figure? From your own radiometer history, which is the reason to keep one. Published decline curves are nominal, individual lamps vary widely, and a lamp in a dirty housing declines faster than the curve predicts because the housing is fouling at the same time. If you have no history, start conservative and let the first few intervals calibrate it.
My margin expires before the next check. What should I change? The interval, usually. Shortening a radiometer check from quarterly to monthly costs a few measurements; raising the dose costs line speed or a lamp purchase, permanently. The instinct is to slow the belt because it feels like a process fix, but it is paying forever for a scheduling gap. Raise the dose only if the interval is already as short as it can practically be.
Is a very large margin a problem? Sometimes, which is why the page flags it rather than celebrating it. Excess dose embrittles some coatings and yellows others, and on a mercury line the infrared that accompanies the UV scales with dwell, so a heat-sensitive substrate can distort at a dose that is acoustically harmless. More often a very large margin is simply unclaimed line speed, which is worth knowing.
Does this cover reflector fouling or a dirty window? Not properly, and it is a real limitation. A single decline percentage models the lamp ageing; quartz clouding and reflector fouling can outpace it substantially and are not separable here. If a measured dose falls faster than the decline figure predicted, suspect the optics before the lamp: the irradiance page splits those losses out.

Last reviewed 2026-08-25.