The colour wheel was the defining component of the first generation of laser phosphor projectors. It was cheap, it was understood, and it solved the problem of getting three primaries out of a blue laser. It also threw away light on every rotation, added a moving part with a finite life, and limited brightness by mechanical speed. The generation now shipping in volume has largely replaced it with static dichroic combiners, and the reasons are optical rather than mechanical.

The architecture that replaced it
A blue laser bank emits at 450 to 465 nm. Part of that light passes through a dichroic and goes straight to the light pipe as the blue primary. The remainder is reflected onto a phosphor layer, which converts it to a broad yellow-green emission spanning roughly 500 to 700 nm. A second dichroic then recombines the converted light with the unconverted blue, and the combined beam enters the illumination relay.
Two static filters replace one spinning wheel. The bill of materials is smaller, the alignment is fixed at assembly, and the only moving parts left are the phosphor substrate's cooling and, in some designs, a diffusion element for speckle reduction.
Efficiency arithmetic
Each dichroic pass costs something. A well-made combiner transmits 95 to 97 percent in its passband and reflects 95 to 97 percent in its stopband, but the two numbers do not multiply favourably for a design that uses both paths: the blue that goes straight through pays one pass, while the converted light that reflects pays another, and any light that is recycled through the phosphor pays twice.
Set against that, the wheel's losses are not small either. A segmented wheel interrupts the beam, so the average transmission depends on duty cycle and segment geometry, and the phosphor conversion itself is 50 to 60 percent efficient for the yellow band. In practice the static combiner wins on system efficiency by a few percent, and wins much more decisively on lifetime and on brightness uniformity across the field, because there is no rotating segment boundary sweeping through the pupil.
Edge steepness is the whole specification
The difficult requirement is not high transmission, it is a steep transition. The blue laser line at 450 nm must be transmitted, and the phosphor emission starting at around 500 nm must be reflected, with as little overlap as possible. That 50 nm gap is generous by dielectric standards, but the combiner works at 45 degrees, where the edge shifts toward shorter wavelengths and softens for p-polarised light.
The consequence is that the filter must be designed against an angle-wavelength-polarisation map, not a single curve. A combiner that measures perfectly at 45 degrees in the design polarisation will leak converted light back toward the laser if the source is unpolarised and the beam diverges, and that leakage shows up as a colour shift at the edge of the image rather than as a brightness loss.
Thermal shift and diode drift
Blue laser diodes shift wavelength with temperature at roughly 0.05 to 0.08 nm per kelvin. Over a 30 K operating range that is a 2 nm move, which sounds small until it is compared with the edge width of a steep combiner. A filter whose edge sits 5 nm from the laser line has very little margin, and the margin shrinks when the coating itself shifts with temperature at another 0.02 nm per kelvin.
Designs that survive field service put the edge further away and accept a slightly softer transition, or they stabilise the diode temperature. Both are cheaper than a service campaign.
Absorption, not just reflection
A projector puts several watts per square millimetre through the combiner. At 50 ppm absorption on a 30 by 30 mm part, that is still tens of milliwatts deposited in a coating a few micrometres thick, and the resulting local temperature rise shifts the edge and, at higher levels, degrades the coating. Absorption below 30 ppm is achievable with ion-beam-sputtered oxide stacks, and the measurement belongs in the specification with a stated method.
Substrate choice matters here as well. A 1.1 mm borosilicate plate spreads heat adequately; a thin 0.5 mm plate does not, and a thick plate introduces beam displacement if it is not parallel. Wedge below one arc minute keeps that displacement from turning into a registration error between colour channels.
What to specify
Peak transmission above 95 percent in the blue passband, reflectance above 95 percent from 500 to 680 nm, transition width under 15 nm from 10 to 90 percent at 45 degrees, absorption below 30 ppm at 455 nm, wedge under one arc minute, flatness lambda/4, and transmission measured at the actual incidence angle and polarisation.
None of these are unusual numbers. What is unusual is a supplier who will measure all of them and report the angle-resolved data, which is the only way to know whether a combiner will hold colour across an image or only in the centre of it.
Product page: dichroic combiners for projection — specifications, coating options and lead times.