LEDディスプレイチップの信頼性と光出力安定性をテストする方法

Linsn TS921 PCIe送信カード前面回路基板

To test an LED display chip, measure its light output, color and forward voltage before and after a controlled aging run. Use the same measurement current, temperature and optical setup for both readings. Compare each LED with its own starting values and record any devices that went dark or became intermittent.

In this guide, “chip” means the light-emitting die in a display pixel. A supplier may test the bare die or the packaged LED that goes onto the module. Ask which one was tested. A package test also captures changes in its optical material, internal connections and thermal path.

LEDシステムの回路を正しく接続してください。

1. Set the LED Life Test Conditions

(1) Define a pass before powering the samples

A pixel can still light up after aging and fail a color or brightness requirement. Define stability in terms you can measure, such as the allowable change in luminous intensity or color coordinates. For reliability, specify what counts as a failure and the operating period over which each sample must meet the requirements.

(2) Specify the electrical and thermal conditions

Use the datasheet for the exact part and color channel. For a scanned display, record peak current, pulse width and duty cycle. Two drive patterns can have the same average current but expose the LED to different peak currents. For temperature, identify the sensor location so that someone repeating the test knows which temperature to reproduce.

LED信頼性試験計画チェックリスト
テスト項目指定するものどのように解釈するか
サンプルIDLEDのモデル、色、パッケージ、ウェハーまたは製造ロット、サンプル数、および選択方法。この結果は、記録された条件下における、サンプリングされた集団に適用される。
電気ストレス順方向電流、駆動モード、およびパルス動作の場合は、ピーク電流、パルス幅、デューティサイクル。機種固有の定格値を使用してください。平均電流値だけでは、パルス状のストレスを正確に表すことはできません。
熱条件周囲温度、設置方法、気流、および温度測定点。周囲温度はLEDの接合部温度とは異なります。
テスト期間経過稼働時間、測定間隔、および中断状況。短時間のスクリーニング検査は、検証済みの寿命予測ではありません。
合否基準試験前に合意された、許容される光学的な変化、電気的なドリフト、および機能的な故障。該当する製品仕様書または認定計画書を使用してください。普遍的な合格基準はありません。

(3) Choose a duration that answers the test question

A burn-in run can reveal early failures. A qualification test asks whether a defined sample meets agreed limits after a specified exposure. If a supplier quotes a projected lifetime, ask for the measured hours, the performance endpoint and the method used to extrapolate from the measurements.

LEDテスター回路図

2. Measure LED Performance Before and After Aging

(1) Select and identify the samples

For an incoming shipment, select devices from the production lots you intend to accept. For a wafer study, record where on the wafer each die came from. Give every sample a fixture position and an ID. Record any defects found at inspection, even if only initially conforming devices will enter the aging run.

(2) Measure light output, color and forward voltage

Take baseline readings, repeat them at the planned intervals, and let the devices settle at the specified measurement temperature each time. For an RGB package, record red, green and blue separately. A white pixel can change tint when one channel loses output faster than the other two.

If you measure luminous intensity in candelas, fix the viewing angle and distance. A small tilt can change the reading for a narrow-beam LED. If you measure total luminous flux in lumens, use flux again for the final reading. The two quantities describe different aspects of the emitted light.

LEDテスター回路図

LEDの劣化前後に記録する測定値
サイズ測定USBレコーディング比較ルール
光出力初期および後期の光束、または固定角度における光度。光学量と測定ジオメトリは全体を通して同じものを使用してください。
順方向電圧指定された測定電流および温度における電圧。各サンプルをそれぞれの基準値と比較し、予期せぬ変動を調査する。
色の一貫性LEDの種類に応じて、色座標または主波長を示します。同じ機器設定と合意済みの色域制限を使用してください。
機能障害開回路、短絡回路、断続的な出力、および検出時間。報告書には、生存したサンプルだけを平均するのではなく、失敗したサンプルも記載してください。
試験条件電流、温度、中断、機器識別情報、および校正状態。治具や測定値の変化と、実際の機器の劣化を区別する。

Light output retention (%) = final output ÷ initial output × 100. For example, an LED that measured 10.0 mcd initially and 9.2 mcd after aging retained 92% of its initial intensity. Report the measurement interval alongside that figure. List failed devices separately so a strong average among survivors cannot hide a failure.

(3) Check the package when output falls

When light output drops, first repeat the reading at the original alignment and temperature. Then examine the package for discoloration or other visible changes. An encapsulant that transmits less light can lower the reading even while the die emits light. Record any sunlight or ultraviolet exposure if the samples were evaluated for outdoor service.

(4) Investigate assembly-related failures

Group unusual results by lot and fixture position. If several outliers occupy one part of the bench, check the local temperature and channel current before attributing the pattern to the LEDs. Keep failed parts for inspection. Die attachment, wire bonds and mounting can each affect the result, and some defects require analysis beyond a visual inspection.

3. Set Up and Check the LED Life Test Bench

(1) Verify the current in each branch

LEDs connected in series carry the same current. Parallel strings can draw different currents because their forward voltages differ. Use branch current regulation or a verified balancing circuit, then measure each branch with the fixture populated. Check that the supply can cover the combined forward voltage of a string plus the voltage its current regulator needs.

LM317定電流駆動回路図

For the LM317 circuit shown above, check the regulator’s voltage headroom and heat dissipation at the intended string current. Select component values for the LED under test.

(2) Trace a dark string to the failed position

One open connection can extinguish a whole series string. Check the supply and fixture connections, then isolate the suspect device. With a shorted LED, one position can stay dark while current still flows through the others. Record which sample failed and how long its neighbors were without current.

(3) Review the bench log and supplier report

Use powered hours from the log. A fixture shut down for a weekend adds elapsed time without adding operating exposure. Check the sample IDs against the report and look for missing final readings, substitutions or exclusions before accepting the reported pass count.

Cree製LED XLamp LEDの信頼性概要 lists different test conditions and failure criteria for different product groups. Its Long-Term Lumen Maintenance document distinguishes measured output from projected maintenance life. For a display pixel, use the qualification plan for that pixel’s own part number.

関連するディスプレイレベルのチェックについては、 LEDディスプレイの輝度を測定および検証するためのガイドをご覧ください。

4 よくある質問

(1) Can I test an LED display chip with a multimeter?

A meter in diode-test mode can help check polarity and forward conduction if its test voltage is sufficient for the LED. Some meters cannot light a blue or green LED. A single reading gives no measurement of light loss or color drift over time.

(2) What does a 96-hour LED burn-in test prove?

It tells you how the tested samples behaved during those 96 hours at the stated current and temperature. If a report says only “passed,” ask how many LEDs entered the test, how many failed and whether light output was measured. Those details determine what the result supports.

(3) What current should I use for an RGB display LED?

Identify the part number, the color channel and whether the test uses continuous or pulsed drive. Use a test point supported by the datasheet and qualification plan. A peak pulse rating applies only under its stated pulse conditions.

(4) Why does an LED look dimmer after it warms up?

Light output can change with junction temperature. Allow the device to return to the baseline measurement conditions and remeasure it at the same current. Any remaining difference warrants checking the optical setup and the sample for lasting changes.

(5) Does a passing LED chip test cover the whole screen?

No. Module solder joints, driver electronics, power distribution and cooling introduce other failure points. Check the assembled screen under its intended scan and brightness settings, including solid red, green and blue patterns to reveal missing or uneven pixels.

5. 結論

A useful reliability report lets you follow one LED from its first reading to its last. Look for its part and lot, drive conditions, powered hours, initial and final optical readings, and any recorded failure. If a supplier gives only a lifetime figure or an average brightness value, ask for the underlying sample results before using that figure to qualify a display module.

コメント送信