Karl Guttag | KGOnTech: Why are reflective waveguides and two-grating designs displacing the three-grating standard, and what causes that rainbow reflection?
02:58 - 04:32
Other snippets from this talk
Summary of the clip:
Why are reflective waveguides and two-grating designs displacing the three-grating standard, and what causes that rainbow reflection?
This year brought a visible shift toward geometric or reflective waveguides, appearing mostly from China but also from Europe and elsewhere. In 1D architectures, LCOS dominates production, while most DLP examples remain prototypes. Several 1D designs now pair reflective expanding waveguides with OLEDs; brighter OLEDs are what let those combinations work reasonably well, which is why a few are reaching the market.
Three-grating waveguides dominated for years β the prior speaker covered them β but newer designs increasingly use two gratings, what was first seen as a wave-optics design. Instead of three, there is an input grating and then a combined expansion and output grating. That arrangement makes the layout more compact, which is a large part of why two-grating architectures are replacing the long-standing three-grating standard.
The other payoff is optical. That rainbow reflection people dislike is caused by the triangular third grating, so switching to a two-grating layout removes the element responsible for it. Separately, the discussion notes a lot of talk about silicon carbide, another thread running alongside these grating-architecture changes in the same set of conversations.
In this short video, you can learn:
* Reflective 1D expanding waveguides are spreading, with LCOS dominating production while most DLP designs stay as prototypes.
* Newer designs increasingly use two gratings β an input grating plus a combined expansion and output grating β replacing the three-grating standard and making the layout more compact.
* The rainbow reflection users dislike comes from the triangular third grating, so removing it eliminates that artifact.
π **Clip Abstract** Reflective and geometric waveguides expanded this year, mostly from China with some from Europe, while LCOS carries 1D production and brighter OLEDs make reflective 1D designs work reasonably well. Newer two-grating layouts β input grating plus combined expansion and output grating β are displacing three-grating designs, cutting size and removing the rainbow reflection traced to that third grating.
About the speaker:
* Speaker: Karl Guttag
* Company: KGOnTech
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#ReflectiveWaveguides, #TwoGratingDesign, #RainbowReflection, #SiliconCarbide, #AROptics, #WaveguideDisplays
This is a highlight of the presentation:
Latest Trends OST AR Hardware and Optics Technology
More Highlights from the same talk.
00:56 - 02:04
How do 1D and 2D reflective waveguides compare to diffractive waveguides in efficiency and color uniformity?
How do 1D and 2D reflective waveguides compare to diffractive waveguides in efficiency and color uniformity?
One notable trend from Lumus: OLEDs are coupling into 1D waveguides more than expected. With 1D pupil expansion handled by conventional refractive optics, the architecture can be fairly efficient. Lumus reported roughly 8% efficiency β nits in to nits out β and that figure holds at about a 50-degree field of view. Compared with typical waveguides, that is described as very efficient, which is why this combination is drawing attention.
For 2D waveguides, the same source reported about 1% efficiency, nits in to nits out. That is several times better than what is typically achieved with diffractive waveguides. The gap matters because pupil expansion architecture and combiner type determine how much light reaches the eye. The source framed the reflective approach as holding a real advantage here, with 1D designs using refractive optics and 2D designs still beating typical diffractive results on throughput.
Color uniformity is the other differentiator. Reflective waveguides are described as having fairly good color uniformity β not perfect, but much, much better than what is typically seen with diffractive waveguides. That combination of efficiency and color performance is the core argument for reflective architectures. The transcript notes agreement in the room on this point, with no correction offered, so the comparison stands as presented: 8% for 1D with refractive optics, about 1% for 2D, both ahead of diffractive alternatives.
In this short video, you can learn:
* The cited source reports roughly 8% nits-in-to-nits-out efficiency for 1D waveguides with 1D pupil expansion and conventional refractive optics at about a 50-degree field of view.
* Their 2D waveguides are said to deliver about 1% efficiency, several times better than typical diffractive waveguides.
* Reflective waveguides' color uniformity is not perfect but is much better than what is typically seen with diffractive waveguides.
π **Clip Abstract** The source describes OLEDs coupling into 1D waveguides at a higher rate than expected, with 1D pupil expansion plus conventional refractive optics reaching about 8% nits-in-to-nits-out efficiency at roughly 50-degree field of view. Their 2D waveguides are reported at about 1% efficiency, several times better than typical diffractive waveguides, while reflective waveguides also offer fairly good color uniformity that, though not perfect, is much better than diffractive alternatives.
About the speaker:
* Speaker: Karl Guttag
* Company: KGOnTech
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#PupilExpansion, #ReflectiveWaveguides, #ColorUniformity, #NitsInNitsOut, #ARWaveguides, #NearEyeDisplays
07:36 - 09:49
How can a two-way combiner solve the blue bleed issue in quantum dot microLEDs while simplifying optical alignment?
How can a two-way combiner solve the blue bleed issue in quantum dot microLEDs while simplifying optical alignment?
The usual combiner is an X-cube or another route for red, green and blue. A two-way approach instead pairs a blue LED with a second panel carrying pumped red and green. Blue pumps both red and green, and because that panel contains no blue, a blue filter can block it hard. This matters because blue-pumped quantum-dot microLEDs consistently show blue bleed, and UV quantum-dot examples remain unseen.
The alternative uses a blue pump but keeps green direct, then uses that green to pump red. According to Hyoke, pumping red from green is almost as efficient as blue pumping; the wavelength difference is not large, so efficiency stays fairly good. This yields direct green, pumped red and a direct blue LED. A simple two-way cube replaces the complex X-cube arrangement.
He thinks that, at a theoretical level, this looks like a potentially good way to go as opposed to an X-cube. It is much simpler and should have very good efficiency. The alignment job is simpler as well: only one panel has to be aligned, not three. Those are the efficiency advantages he highlights for the two-way combiner.
In this short video, you can learn:
* Blue-pumped quantum-dot microLEDs have consistently shown blue bleed, and UV quantum-dot examples were not yet seen.
* A blue-pumped red/green panel can receive a solid blue block because it contains no blue to leak.
* Green-pumping red is almost as efficient as blue-pumping red, so direct green, pumped red and direct blue become possible.
π **Clip Abstract** A two-way combiner can separate a blue LED from a red/green pumped panel, allowing a solid blue block where no blue exists. The clip contrasts that with green-pumped red and argues a simple two-way cube may be simpler to align and more efficient than an X-cube.
About the speaker:
* Speaker: Karl Guttag
* Company: KGOnTech
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#TwoWayCombiner, #BlueBleed, #QuantumDotMicroLEDs, #XCube, #MicroLEDDisplays, #DisplayOptics
02:14 - 04:40
Why does optimizing micro lenses on microLEDs individually fail to improve overall system efficiency, and how can we fix it?
Why does optimizing micro lenses on microLEDs individually fail to improve overall system efficiency, and how can we fix it?
The message from Lumus is that the microLED engine must be considered part of a system, not optimized in isolation. Putting micro lenses on top of a microLED may improve the microLED's efficiency or produce more nits, but it may not optimize the system. Uniform illumination will not result. Illumination is less effective than when micro lenses are designed in conjunction with the entire optical system.
If telecentric optics are absent, you may think you have done well by putting micro lenses over your LED, but particularly towards the periphery you will get very dark images. You will not get the coupling efficiency you expect. The whole micro lens and optics and everything must be designed as a complete unit, instead of doing them individually and then slapping them together.
As photons emerge from a panel with micro lenses, their translation matters, and the micro lenses may need to vary. At the periphery, varying their shape slightly aims light differently. Simulations show shifting, centering, and omitting the micro lens array. Doing it right could yield 6x more efficiency, or 600% better efficiency. Micro lenses help couple out more light, but the array should be designed in conjunction with the whole system.
In this short video, you can learn:
* Micro lenses on microLEDs can improve LED efficiency or nits but may not optimize the full optical system.
* Without telecentric optics, peripheral images become very dark and expected coupling efficiency is lost.
* Designing the micro lens array with the entire optical system can yield 6x or 600% better efficiency, while varying lens shape toward the periphery aims light differently.
π **Clip Abstract** Lumus argues that microLED engines must be designed as complete optical systems, not optimized piece by piece. The clip shows that micro lenses over microLEDs can help couple out more light, but without telecentric optics and periphery-varying lens shapes, uniform illumination and expected coupling efficiency are lost, with correct co-design potentially giving 6x or 600% better efficiency.
About the speaker:
* Speaker: Karl Guttag
* Company: KGOnTech
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#MicroLensArray, #TelecentricOptics, #CouplingEfficiency, #PeripheryDarkImages, #AROptics, #MicroLEDDisplays