Why is cost per operating hour the headline?

Because a per-swap total cannot be compared with anything. Nothing else in a cost model is quoted per swap, two lamps with different lives cannot be ranked by it, and a payback case cannot use it. The hourly figure fixes all three: it sits beside energy on the same line, it ranks lamps correctly regardless of price, and it annualises with a single multiplication.

What happened to the scrap risk allowance?

It was removed deliberately. It was a lump sum with no stated derivation, which made the total look more complete without making it more accurate. Scrap risk is real, but it is a defect rate times a part cost times a production volume, and the rework page derives it from those. Rated life took the input slot instead, because it turns this page's answer into something usable elsewhere.

Should I divide by lamps × life for a multi-lamp system?

No, and this is the easy error. The lamps run at the same time, so a four-lamp swap does not buy four lamp-lives of operation. It buys one, during which all four are running. Dividing by lamps times life understates the hourly cost by a factor equal to the lamp count.

Is it worth paying more for longer-life lamps?

Compute it rather than assume it, which is what the hourly row is for. A lamp costing 38% more with 75% more life is meaningfully cheaper to run, and it also buys fewer swaps, which means less downtime and fewer re-qualifications. The catch is that longer-life lamps are only cheaper if they actually achieve their rating on your system. The same cycling and cooling problems that shorten a cheap lamp shorten an expensive one.

Why is my achieved life below the datasheet figure?

Frequent striking is the usual cause: each start costs disproportionate life, so a system that cycles the lamp will never see its rated hours. After that, cooling, an overheated seal ends a lamp early, then low or unstable supply voltage, then handling contamination, where a fingerprint on the quartz bakes in and creates a hot spot. All four are cheaper to fix than to keep paying for.

Does this include disposal?

No. Mercury lamps are regulated waste almost everywhere and carry a handling and disposal cost that varies too much by jurisdiction to assume. Add it into the unit price if it is material to you. It is one of the smaller line items in a retrofit case, but it is not zero, and it is one of the few that disappears entirely with LED.

UV Curing calculator

UV Lamp Replacement Cost Calculator: Cost per Operating Hour

Price a UV lamp change properly and put the answer in a usable unit. Enter the number of lamps in the swap, the delivered price of each, what the downtime costs, and the life the new lamps are expected to give. The calculator returns the cost per operating hour first, then the event cost behind it, the material share, the all-in cost per lamp, and how much of the swap is the stopped line rather than the lamps. Cost per operating hour is the headline because it is the only form in which lamp cost can be compared between lamps, added to energy on a per-part line, or carried into a retrofit case.

What this calculator does

Price a lamp swap including its downtime, and convert it into cost per operating hour. The only form in which lamp cost can be compared, budgeted or carried into a retrofit case.
Use it for comparing lamp options that differ in both price and life, putting lamp cost onto a per-part cost model beside energy, building the mercury side of a UV LED retrofit case, showing what scheduling swaps into planned shutdowns is worth, budgeting annual lamp spend from runtime.
Price a lamp swap including its downtime, and convert it into cost per operating hour. The only form in which lamp cost can be compared, budgeted or carried into a retrofit case.

Formula used

Lamp material cost = lamp count × unit price
True cost per swap = lamp material cost + downtime cost
Lamp cost per operating hour = true cost per swap ÷ rated lamp life
Cost per lamp, all in = true cost per swap ÷ lamp count
Downtime share (%) = downtime cost ÷ true cost per swap × 100

Inputs explained

Number of lamps in this swap: Lamps changed in one event. In a multi-lamp system they normally go together, since a mixed-age set gives a cure surface with a moving profile.
Unit price per lamp: Delivered price of one lamp, including freight and any core charge. Doped and specialist lamps run well above standard mercury.
Downtime cost during the swap: Lost contribution while the line is stopped, plus the labour doing the change. A swap folded into a planned shutdown can honestly be entered as zero.
Rated lamp life: Hours the new lamps are expected to give before the next swap. From the datasheet, or better, from your own swap history on this system.

How to use the result

Best suited to comparing lamp options that differ in both price and life, putting lamp cost onto a per-part cost model beside energy, building the mercury side of a UV LED retrofit case, showing what scheduling swaps into planned shutdowns is worth, budgeting annual lamp spend from runtime.
Ignores the decay that happens across the lamp's life. Cost per hour is flat here; delivered dose is not, and the last quarter of a lamp's life is the least productive part of it. Does not price a mixed-age set, which is a real and common situation with a genuinely non-flat cure profile. Excludes reflectors, quartz sleeves, filters and seals, which are consumables on the same system with their own intervals. Takes rated life at face value. Cycling, poor cooling, low supply voltage and handling contamination all shorten it, sometimes drastically. Says nothing about disposal. Mercury lamps are regulated waste in most jurisdictions and the handling has a cost.

Common questions

Why is cost per operating hour the headline? Because a per-swap total cannot be compared with anything. Nothing else in a cost model is quoted per swap, two lamps with different lives cannot be ranked by it, and a payback case cannot use it. The hourly figure fixes all three: it sits beside energy on the same line, it ranks lamps correctly regardless of price, and it annualises with a single multiplication.
What happened to the scrap risk allowance? It was removed deliberately. It was a lump sum with no stated derivation, which made the total look more complete without making it more accurate. Scrap risk is real, but it is a defect rate times a part cost times a production volume, and the rework page derives it from those. Rated life took the input slot instead, because it turns this page's answer into something usable elsewhere.
Should I divide by lamps × life for a multi-lamp system? No, and this is the easy error. The lamps run at the same time, so a four-lamp swap does not buy four lamp-lives of operation. It buys one, during which all four are running. Dividing by lamps times life understates the hourly cost by a factor equal to the lamp count.
Is it worth paying more for longer-life lamps? Compute it rather than assume it, which is what the hourly row is for. A lamp costing 38% more with 75% more life is meaningfully cheaper to run, and it also buys fewer swaps, which means less downtime and fewer re-qualifications. The catch is that longer-life lamps are only cheaper if they actually achieve their rating on your system. The same cycling and cooling problems that shorten a cheap lamp shorten an expensive one.
Why is my achieved life below the datasheet figure? Frequent striking is the usual cause: each start costs disproportionate life, so a system that cycles the lamp will never see its rated hours. After that, cooling, an overheated seal ends a lamp early, then low or unstable supply voltage, then handling contamination, where a fingerprint on the quartz bakes in and creates a hot spot. All four are cheaper to fix than to keep paying for.

Last reviewed 2026-08-25.