Lisa Taubensee | Teradyne: Why does a microLED's emission profile demand custom optics, and how does one beam splitter test emitters and detectors together?
06:19 - 08:09
Other snippets from this talk
Summary of the clip:
Why does a microLED's emission profile demand custom optics, and how does one beam splitter test emitters and detectors together?
Emission profile of microLEDs is extremely important, and requirements differ by application. For AR with a wide field of view the collection needs differ from tight coupling into a fiber bundle or an FAU. The lens and optics must match the device. In practice customization is often required: different customers present different emission profiles and custom micro-lenses on their microLEDs.
A lens that accurately collects the light from your microLEDs feeds the camera, radiometer or photometer. High-resolution imaging is the typical test methodology: you take an image of your microLED array or micro display and use that image for defect detection. This makes the collection optics central to the measurement, because the emission profile and the micro-lens determine what the camera actually sees.
A beam splitter with a port on the lens enables spectral measurements at the same time, sending some light to a spectrometer. That is an interesting concept because it adds a second measurement channel without changing the insertion. For optical interconnects, a light source can be popped on, so photodetectors and microLEDs are tested in one insertion.
In this short video, you can learn:
* MicroLED emission profiles vary by application, so AR wide-field-of-view collection and tight fiber-bundle or FAU coupling place different demands on the lens.
* Customization is often required because different customers have different emission profiles and custom micro-lenses on the microLEDs.
* A beam splitter with a lens port allows simultaneous spectral measurement to a spectrometer, and with a light source added, photodetectors and microLEDs are tested in one insertion.
📋 **Clip Abstract** Testing microLEDs starts with a lens whose collection matches the emitter's emission profile, because AR wide-field-of-view and tight fiber or FAU coupling pose different requirements. High-resolution imaging gives defect detection on arrays and micro displays, while a beam splitter port adds simultaneous spectral measurement and, with a light source, tests photodetectors and microLEDs in one insertion.
About the speaker:
* Speaker: Lisa Taubensee
* Company: Teradyne
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#MicroLEDEmissionProfile, #HighResolutionImaging, #BeamSplitter, #OpticalInterconnects, #MicroLEDDisplays, #ARDisplays
This is a highlight of the presentation:
Challenges of MicroLED testing
More Highlights from the same talk.
02:55 - 04:46
Why do millions of micron-scale microLEDs force fine-pitch probing, fast measurement and absolute calibration together?
Why do millions of micron-scale microLEDs force fine-pitch probing, fast measurement and absolute calibration together?
MicroLED wafers now carry millions, even billions, of emitters. Each device is on the scale of one or two microns square, and some companies push below one micron, while five to ten microns remains standard for many applications. Every one of those emitters still has to be tested, so the probing system needs very fine pitch and very fine resolution to reach devices at that density.
The second demand is speed. Measurements must be fast and accurate enough to support production throughput; sitting for a day to test a single wafer is not practical. Calibration is the third: production test needs systems whose measurements can be trusted, whether the tool sits beside another system or somewhere else in the world, so material can be loaded and brightness confirmed, defects detected and testing run seamlessly across the world.
Because microLEDs are still evolving, a setup that delivers accurate, absolute measurement also helps track the next generation of product. It answers the question of what improvements have been made: how do these microLEDs compare with material from two years ago, and with the changes implemented since? Absolute calibration turns a measurement into a trend line across product generations rather than a one-off number.
In this short video, you can learn:
* Wafer-level microLED test must resolve emitters one to two microns square, with some companies going sub-one micron while five to ten microns is standard for many applications.
* Production throughput rules out spending a day on one wafer, so measurement must be fast and accurate.
* Trusted calibration lets identical material be measured beside another system or anywhere in the world, so brightness and defect results transfer and product generations can be compared.
📋 **Clip Abstract** Testing microLED wafers means probing millions of one-to-two-micron emitters, so fine pitch and resolution, fast accurate measurement for throughput, and trusted calibration are demanded at once. That calibration also gives absolute numbers that let teams compare current material with devices made two years ago and track what improvements have been achieved.
About the speaker:
* Speaker: Lisa Taubensee
* Company: Teradyne
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#MicroLEDWaferTest, #FinePitchProbing, #AbsoluteCalibration, #ProductionThroughput, #MicroLEDDisplays, #DisplayMetrology
11:29 - 13:07
Can microLED links hit four-gig rates, and what tests prove it — eye diagrams, BER, rise time, insertion loss?
Can microLED links hit four-gig rates, and what tests prove it — eye diagrams, BER, rise time, insertion loss?
At-speed testing for optical interconnects is a different and new topic for the micro LED community. The alignment system required will differ: aligning to a camera may not be far off for micro LED, and aligning to an FAU is still not the same. Arrangement and build of DUTs will be slightly different. Multi-site is another question—can you more easily test as many as possible in one shot?
Bit error rate testing and eye diagrams are new tests to consider. Customers and people ask about rise and fall time: how fast can these be toggled? Can four-gig links be made stably, or is three gig about what can be reached at the moment with the technology? Insertion loss into fibers must also be considered. Many of these tests are very similar to any time you are sending high-speed data.
Protocol-aware engines and controls are needed to control your micro display or modulate your micro array. For optical interconnects, at-speed testing brings questions around bit error rate, eye diagrams, rise and fall time, stable four-gig links versus three gig, and insertion loss into fibers. The DUT arrangement and build change, alignment to FAU differs from camera alignment, and multi-site one-shot testing becomes a practical target. These are new tests for high-speed data paths.
In this short video, you can learn:
* At-speed optical-interconnect testing requires a different alignment system than camera alignment and is not the same as FAU alignment.
* Bit error rate, eye diagrams, rise and fall time, stable four-gig links, and insertion loss into fibers are the tests being raised.
* Multi-site testing asks whether as many DUTs as possible can be tested in one shot, with different DUT arrangement and build.
📋 **Clip Abstract** At-speed testing of optical interconnects introduces new alignment, DUT arrangement and multi-site questions for micro LED. The tests under discussion include bit error rate, eye diagrams, rise and fall time, stable four-gig versus three-gig links, and insertion loss into fibers.
About the speaker:
* Speaker: Lisa Taubensee
* Company: Teradyne
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#AtSpeedOpticalTesting, #BitErrorRateTesting, #EyeDiagramAnalysis, #InsertionLossIntoFibers, #OpticalInterconnects, #MicroLEDDisplays
09:48 - 11:30
What exact test list does a microLED array need — brightness, FWHM, dominant wavelength, leakage — and must it speak MIPI?
What exact test list does a microLED array need — brightness, FWHM, dominant wavelength, leakage — and must it speak MIPI?
Per micro LED brightness or luminance is the most common measurement. It supports DMIRA on micro displays, and in optical interconnects it identifies which pixels or micro LEDs cannot transmit data. It exposes defects, cluster defects and pixels that may be shorted together. Uniformity is the next typical test across the array, because those pixel-level failures and variations determine whether an array can be used for display or data transmission.
Spectral characterization matters: you want an idea of the wavelength the array emits at, plus full width half maximum. In AR/VR you also want dominant wavelength and photopic measurements filtered through the human eye. Electrically, the list includes opens, shorts, leakage current, trim, shmoo and any margin characterization. That requires a power supply to bias and exercise the micro LEDs while the optical and electrical data are captured.
Protocol-aware testing is the other requirement. Many micro displays these days speak natively MIPI, so you want the ability to talk to your display. That means protocol-aware engines and whatever is needed to control the micro display or modulate the micro array. The test list therefore spans optical, electrical and protocol layers, and the same photonics test approach already applied elsewhere can be reused here.
In this short video, you can learn:
* Per-microLED luminance testing supports DMIRA on micro displays and identifies pixels that cannot transmit data in optical interconnects.
* Spectral tests cover emission wavelength, FWHM, dominant wavelength and photopic measurements, while electrical tests cover opens, shorts, leakage, trim, shmoo and margin.
* Many micro displays natively speak MIPI, so protocol-aware engines are needed to control the display or modulate the micro array.
📋 **Clip Abstract** MicroLED arrays require a per-pixel test list that starts with brightness or luminance and extends to uniformity, spectral and photopic measurements, plus electrical opens, shorts, leakage, trim, shmoo and margin characterization. Because many micro displays speak MIPI natively, test also becomes protocol-aware, using engines that talk to the display and modulate the micro array.
About the speaker:
* Speaker: Lisa Taubensee
* Company: Teradyne
* Event: Eindhoven 2026
* Location: High Tech Campus, Eindhoven
#MicroLEDTesting, #MIPIProtocolTesting, #DominantWavelength, #LeakageCurrent, #MicroLEDDisplays, #OpticalInterconnects