Does UV LED really use less energy than mercury?

Usually yes, for three separate reasons worth keeping apart: a larger share of input power leaves as usable UV in the band being emitted, the array draws nothing between parts instead of idling shuttered, and it can be driven below full power on undemanding jobs. How much less in your case depends on your duty cycle and on what dose the material actually needs, not on the ratio of nameplate ratings.

What counts as wall-plug power?

Everything the supply sees: the emitters, the driver electronics, and the cooling. Air-cooled arrays hide most of their cooling load in the same figure; water-cooled ones usually do not, and a chiller left out of the number can be a substantial fraction of the total. Measure at the panel if you can.

Why can't I just compare kilowatts with the mercury page?

Because the two technologies deliver very different fractions of their input as usable UV, in different spectral distributions. A mercury lamp emits broadly across UVA, UVB and UVC; an LED array emits a narrow band, typically at 365, 385 or 395 nm. What matters is dose delivered in the band your photoinitiator absorbs, at the line speed you run. Two systems are comparable when they cure the same parts at the same rate, and at that point cost per part is the honest comparison.

Will my existing chemistry cure under LED?

Not necessarily, and this is the failure that ruins retrofit cases after the energy sums are already agreed. Formulations relying on short-wavelength output for surface cure can gel underneath and stay tacky on top under a long-wavelength narrow-band source, because the photoinitiator that handled the surface is not being excited at all. Trial the actual material before committing to any of these numbers.

Does the array's output fall over time?

Yes, though far more slowly than a mercury lamp's. As the emitters age, holding the same irradiance takes more drive current, so the wall-plug figure creeps up across the array's life. For a first-pass energy estimate the effect is small; for a ten-year cost model it is worth carrying explicitly.

Is the annual figure a measurement?

No. It is this page's shift cost multiplied by 250, a stated convention for an ordinary single-shift year. Double it for two shifts, scale it for seasonal work, replace it for continuous operation. It exists because the payback page needs an annual number and the earlier version gave nobody a way to compute one.

UV Curing calculator

UV LED Energy Cost Calculator: Cost per Part

Price the electricity a UV LED array consumes, in the form that can be compared with anything else. Enter the array's wall-plug power at the drive level you actually run, the hours it emits per shift, your blended electricity rate, and the parts produced. The calculator returns cost per part first, then the shift, annual and hourly figures. Cost per part leads because it is the only basis on which this page and the mercury energy page describe the same job. Installed kilowatts are not comparable between the two technologies, and a per-shift figure just rewards whichever system was assumed to run fewer hours.

What this calculator does

Turn a UV LED array's wall-plug power and runtime into energy cost per part, per shift and per year, on a basis that can honestly be set against a mercury system.
Use it for establishing the energy line for a UV LED installation, producing the annual figure a retrofit payback case needs, comparing LED and mercury on a common per-part basis, pricing the effect of running an array below full drive, quantifying what part-gating saves on a line with poor flow.
Turn a UV LED array's wall-plug power and runtime into energy cost per part, per shift and per year, on a basis that can honestly be set against a mercury system.

Formula used

Energy used per shift (kWh) = wall-plug power × runtime hours
Shift energy cost = energy used per shift × electricity rate
Energy cost per part = shift energy cost ÷ parts cured per shift
Hourly energy cost = wall-plug power × electricity rate
Annual energy cost = shift energy cost × 250 shifts (a disclosed convention: substitute your own year)

Inputs explained

Array wall-plug power: Power drawn from the supply at the drive level actually used, including the driver and any chiller. Not the sum of the diodes' optical output.
Runtime per shift: Hours the array is actually emitting. Unlike a mercury lamp, an LED array switches instantly, so this is curing time rather than energized time.
Blended electricity rate: All-in rate including demand charges and delivery, from an actual invoice.
Parts cured per shift: Good parts through the same shift. The only basis on which this page and the mercury page describe the same job.

How to use the result

Best suited to establishing the energy line for a UV LED installation, producing the annual figure a retrofit payback case needs, comparing LED and mercury on a common per-part basis, pricing the effect of running an array below full drive, quantifying what part-gating saves on a line with poor flow.
Cannot verify that the array delivers the dose the material needs. An energy figure for a system that does not cure the parts is not a saving. Assumes constant draw while emitting; arrays with dynamic power control and part-gating will differ, generally in LED's favour. Ignores array degradation, which raises the drive current needed for a given irradiance over the array's life and with it the wall-plug power. Says nothing about capital cost, which is where the mercury-versus-LED decision usually turns; the payback page carries that. Comparisons to mercury are valid only at equal dose, equal band and equal throughput, and this page cannot check any of the three.

Current U.S. benchmarks

As of May 2026, industrial electricity averages 8.7 cents per kWh across the U.S. (EIA), up 5.1% from a year earlier. State averages range widely, so plants should confirm against their own tariff.

Common questions

Does UV LED really use less energy than mercury? Usually yes, for three separate reasons worth keeping apart: a larger share of input power leaves as usable UV in the band being emitted, the array draws nothing between parts instead of idling shuttered, and it can be driven below full power on undemanding jobs. How much less in your case depends on your duty cycle and on what dose the material actually needs, not on the ratio of nameplate ratings.
What counts as wall-plug power? Everything the supply sees: the emitters, the driver electronics, and the cooling. Air-cooled arrays hide most of their cooling load in the same figure; water-cooled ones usually do not, and a chiller left out of the number can be a substantial fraction of the total. Measure at the panel if you can.
Why can't I just compare kilowatts with the mercury page? Because the two technologies deliver very different fractions of their input as usable UV, in different spectral distributions. A mercury lamp emits broadly across UVA, UVB and UVC; an LED array emits a narrow band, typically at 365, 385 or 395 nm. What matters is dose delivered in the band your photoinitiator absorbs, at the line speed you run. Two systems are comparable when they cure the same parts at the same rate, and at that point cost per part is the honest comparison.
Will my existing chemistry cure under LED? Not necessarily, and this is the failure that ruins retrofit cases after the energy sums are already agreed. Formulations relying on short-wavelength output for surface cure can gel underneath and stay tacky on top under a long-wavelength narrow-band source, because the photoinitiator that handled the surface is not being excited at all. Trial the actual material before committing to any of these numbers.
Does the array's output fall over time? Yes, though far more slowly than a mercury lamp's. As the emitters age, holding the same irradiance takes more drive current, so the wall-plug figure creeps up across the array's life. For a first-pass energy estimate the effect is small; for a ten-year cost model it is worth carrying explicitly.

Last reviewed 2026-08-25.