How is this different from the UV uniformity page?

Different subject, and therefore a different fix. The uniformity page surveys a LAMP's cure surface: the spread comes from reflector condition, alignment, working distance and end-of-arc fall-off, and the remedies are optical. This page maps a shaped PART: the spread comes from faces at different angles and distances, and the remedies are fixturing, rotation, or another head. the earlier version computed the identical formula on both pages and never said which problem you were looking at.

Why does raising the dose make the hot spot worse?

Because a lift is proportional, not additive. More lamp power or more dwell multiplies every grid point by the same factor, so closing a 20% shortfall at the cold point puts an extra 26% on the hot point too. That is arithmetic, not pessimism, and it is why the page prints both numbers together. The question is never just whether the cold point can be cured but whether the hot one survives it.

My cold point is in a recess. What actually helps?

Change what the feature sees, not how much the lamp emits. Rotating the part between passes, adding a second head at a complementary angle, tilting the fixture, or reflecting light into the recess all raise the cold point without raising the hot one in proportion. A photoinitiator with a longer-wavelength response can also help where the shortfall is partly absorption rather than pure shadow.

How fine should the grid be?

Fine enough to land on the features you suspect, which usually means placing readings deliberately rather than on a regular pitch. A coarse regular grid biases in the dangerous direction: the true minimum is below the lowest point you happened to measure, so a map that just clears requirement on a coarse grid has probably not cleared it at all.

Is a high variation figure automatically a problem?

No, and this is the trap in reporting variation alone. A map with 40% variation whose coldest point still has 25% margin is producing fully cured parts. A map with 10% variation sitting entirely below requirement is scrapping everything. Variation describes the shape of the map; only the cold point against the requirement says whether parts cure.

Can I use the average dose as evidence of cure?

No. Cure is local, each area of the surface either received enough dose or it did not, and no averaging happens between one face and another. On a shaped part the average is routinely well above requirement while a specific feature is short, which is precisely the failure that reaches a customer, because the part looks finished everywhere the eye goes first.

UV Curing calculator

UV Dose Mapping Calculator: Cold Points on a Shaped Part

Take a grid of dose readings over a shaped part and find out what it actually tells you. Enter the coldest and hottest grid points, the average across the map, and the dose the material requires. The calculator returns the earlier version's variation figure plus the two numbers that decide what to do next: how far the coldest point sits from the requirement, and what the average, and with it the hot spot: would have to be raised to before that cold point clears. On a shaped part those two move together, which is why a shadowed feature is rarely fixed with more lamp.

What this calculator does

Turn a grid of dose readings taken over a shaped part into the two answers that decide anything: did the coldest point cure, and what does fixing it cost the hot spot.
Use it for qualifying a new shaped part before it goes to production, explaining why a recessed feature fails while flat faces pass, deciding between more dose, rotation, and a second lamp head, setting the acceptance basis for a fixture change, checking whether a proposed lift would overcure the top face.
Turn a grid of dose readings taken over a shaped part into the two answers that decide anything: did the coldest point cure, and what does fixing it cost the hot spot.

Formula used

Dose spread = hottest grid point − coldest grid point
Variation from average (%) = spread ÷ average × 100
Coldest point against requirement (%) = (coldest − required) ÷ required × 100
Scale needed to clear the cold point = required ÷ coldest
Average dose needed = average × scale; hot spot after scaling = hottest × scale

Inputs explained

Coldest grid-point dose (cold spot): Lowest dose recorded anywhere on the mapped part. On a shaped part this is normally a face turned away from the source, not a distant one.
Hottest grid-point dose (hot spot): Highest dose recorded on the same map. Usually the face square to the lamp at the shortest distance.
Average dose across the grid: Mean of every grid point. It describes the map; it is not a dose any point on the part received.
Required cure dose: What the coating or adhesive needs, in the band the map was measured in. the earlier version had no field for this, so it could report a spread and never say whether the part cured.

How to use the result

Best suited to qualifying a new shaped part before it goes to production, explaining why a recessed feature fails while flat faces pass, deciding between more dose, rotation, and a second lamp head, setting the acceptance basis for a fixture change, checking whether a proposed lift would overcure the top face.
Three summary numbers cannot describe a map. Two parts with identical cold, hot and average readings can have very different amounts of surface in the cold region, and that fraction is what decides scrap. Says nothing about WHERE the cold point is, and on a shaped part the location is the whole fix. The scaling rows assume a proportional lift. Moving the lamp closer is not proportional. It changes the geometry that caused the spread in the first place, helping some faces and not others. A coarse grid overstates the cold point, because the true minimum is almost certainly below the lowest reading taken. Ignores rotation and multi-pass exposure, which are exactly the remedies a geometry-driven map usually calls for.

Common questions

How is this different from the UV uniformity page? Different subject, and therefore a different fix. The uniformity page surveys a LAMP's cure surface: the spread comes from reflector condition, alignment, working distance and end-of-arc fall-off, and the remedies are optical. This page maps a shaped PART: the spread comes from faces at different angles and distances, and the remedies are fixturing, rotation, or another head. the earlier version computed the identical formula on both pages and never said which problem you were looking at.
Why does raising the dose make the hot spot worse? Because a lift is proportional, not additive. More lamp power or more dwell multiplies every grid point by the same factor, so closing a 20% shortfall at the cold point puts an extra 26% on the hot point too. That is arithmetic, not pessimism, and it is why the page prints both numbers together. The question is never just whether the cold point can be cured but whether the hot one survives it.
My cold point is in a recess. What actually helps? Change what the feature sees, not how much the lamp emits. Rotating the part between passes, adding a second head at a complementary angle, tilting the fixture, or reflecting light into the recess all raise the cold point without raising the hot one in proportion. A photoinitiator with a longer-wavelength response can also help where the shortfall is partly absorption rather than pure shadow.
How fine should the grid be? Fine enough to land on the features you suspect, which usually means placing readings deliberately rather than on a regular pitch. A coarse regular grid biases in the dangerous direction: the true minimum is below the lowest point you happened to measure, so a map that just clears requirement on a coarse grid has probably not cleared it at all.
Is a high variation figure automatically a problem? No, and this is the trap in reporting variation alone. A map with 40% variation whose coldest point still has 25% margin is producing fully cured parts. A map with 10% variation sitting entirely below requirement is scrapping everything. Variation describes the shape of the map; only the cold point against the requirement says whether parts cure.

Last reviewed 2026-08-25.