Production testing of HB-LEDs and LED modules demands the right hardware and techniques (MAGAZINE)

Nov. 30, 2011
Cost-effective testing of LEDs and LED modules in a production environment can have a strong influence on manufacturing efficiency, as MARK CEJER explains.
+++++

This article was published in the November/December 2011 issue of LEDs Magazine.

View the Table of Contents and download the PDF file of the complete November/December 2011 issue.

+++++

The latest high-brightness LEDs (HB-LEDs) offer ever-higher luminous flux, longer lifetimes, greater chromaticity, and more lumens per watt. To reduce the cost of LEDs to consumers, manufacturers are working to improve their yields and their manufacturing-efficiency levels. This requires cost-effective testing of LEDs and modules in a production environment. The key is to find the right combination of test equipment and the knowledge of how to use it effectively.

HB-LED testing fundamentals

FIG. 1. A typical diode’s electrical I-V (current-voltage) curve is shown in Fig. 1. Although a complete test sequence could include hundreds of points, a limited sample is generally sufficient to probe for the figures of merit. Many HB-LED tests involve supplying a known current and measuring the resulting voltage, or vice-versa, so a single piece of hardware that synchronizes both functions can result in quicker system setup and enhanced throughput.

Testing can be done at the die level (both wafer and package) or the module/subassembly level. In the latter case, HB-LEDs are connected in series and/or parallel; therefore, higher currents are typically involved, sometimes up to 50A or more, depending on the application. Some die-level testing can require 5 10A, depending on die size.

Forward-voltage (VF) test: A forward-voltage test verifies the device’s forward operating voltage. When a forward current is applied to the diode, it begins to conduct. During the initial low-current values, the voltage drop across the diode increases rapidly but levels off as the drive current increases. The region of relatively-constant voltage is where the diode normally operates. Results are often used in binning devices because an HB-LED’s VF is related to its chromaticity.

Optical tests: Forward current biasing is also used for optical tests because current flow is closely related to the amount of light emitted by an HB-LED. A photodiode or integrating sphere can be used to capture the emitted photons to measure optical power. This light is converted to a current that’s measured using an ammeter or one channel of a source-measure unit (SMU).

Reverse breakdown voltage (VR) test: A negative bias current applied to an HB-LED allows probing for its reverse breakdown voltage. The test current should be set to a level where the measured voltage value no longer increases significantly when current is increased slightly. At voltages higher in magnitude than the breakdown voltage, large increases in reverse-bias current produce insignificant changes in reverse voltage. The VR test is performed by sourcing a low-level reverse-bias current for a specified time, then measuring the voltage drop across the HB-LED. Results are typically in the tens of volts.

Leakage current testing: Moderate voltages are normally used to measure the current that leaks (IL) across an HB-LED when a reverse voltage less than breakdown is applied. In production testing, it is common practice to ensure only that leakage doesn’t exceed a specified threshold.

Boosting test throughput in production

At one time, manufacturers used an external PC to control all aspects of HB-LED production testing; in each element of a test sequence, the sources and instruments had to be configured for each test, perform the desired action, and then return the data to the PC, which evaluated the pass/fail criteria and determined where to bin the DUT (device under test). Transferring commands from the PC and results back to it ate up a lot of test time.

FIG. 2. The latest generation of smart instruments, including Keithley’s new Model 2651A High Power System SourceMeter instrument (Fig. 2), is optimized to boost throughput substantially by minimizing communication traffic. The majority of the test sequence is embedded in the test instrumentation within a script and is executed by a microprocessor that allows control of the test sequence, with internal pass/fail criteria, calculations, and control of digital I/O. The microprocessor stores a user-defined test sequence in memory and executes it on command to all SourceMeter instruments in the test configuration, reducing set-up and configuration time.

Communication between units takes place via TSP-Link technology, a high-speed trigger-synchronization/inter-unit communication bus, which connects multiple instruments in a master/slave configuration. This eliminates time-consuming GPIB (general purpose interface bus) traffic, and greatly enhances system throughput.

LED test system for a single device

A system configuration for testing one HB-LED at a time is shown in Fig. 3. The component handler transports the individual HB-LED to a test fixture, which is shielded from ambient light and houses a photodetector (PD) for light measurements. Two SMUs are used: SMU #1 supplies the test signal to the HB-LED and measures its electrical response, while SMU #2 monitors the photodiode during optical measurements.

FIG. 3. The test sequence is programmed to begin using a digital line from the component handler that serves as a start-of-test (SOT) signal. After the instrument detects this signal, the test sequence begins. Once completed, a digital line signals measurement-complete status to the component handler. In addition, the instrument’s built-in intelligence performs all pass/fail operations and sends a digital command through the instrument’s digital I/O port to the component handler to bin the HB-LED based on the pass/fail criteria. Then, two actions can be programmed to take place simultaneously: data is transferred to the PC for statistical process control, and a new DUT moves into the test fixture.

Production testing of HB-LEDs

To achieve acceptable throughput, production test systems measure multiple parts simultaneously. Fig. 4 illustrates a device test system for three HB-LEDs that has one photodiode (PD) channel.

FIG. 4. Junction self-heating can contribute significant measurement error in HB-LED production testing. As the junction heats over time, for a constant forward-bias current, the forward voltage drops, so it’s crucial to manage device self-heating to ensure accurate, repeatable measurements. Self-heating can be minimized by reducing the amount of time the test will take, which in turn reduces the amount of time necessary for the test current to be applied to the device. Smart instruments can simplify configuration of the device soak time (which allows any circuit capacitance to settle before the measurement begins), as well as the integration time (which defines how long it takes the analog-to-digital (ADC) to acquire the input signal), because both factor into how long the test will take. New SMU instruments, including Keithley’s 2651A, have digitizing ADCs, which can sample at speeds up to one microsecond per point or up to 50 times faster than high-performance integrated ADCs. These higher measurement speeds further improve overall test times.

The use of pulsed measurements minimizes test times and junction self-heating. Modern SMUs with high pulse-width resolution ensure precise control over how long power is applied to the device. Pulsed operation also allows these instruments to output current levels well beyond their DC capabilities.

High-power LED module testing

The demand for a lot of light in a small package has led lighting manufacturers to develop high-power LED modules, which often consist of one or more large-die LEDs. When multiple die are present, they’re either wired in parallel or in series, depending on the application and the available power source. The die of these LEDs can be much larger than those of typical HB-LEDs and can handle much higher currents. In fact, it’s common for a single die to be required to withstand current levels as high as 10A.

Obviously, testing high power HB-LED modules demands hardware that can deliver a lot of power to the DUT. Although SMUs’ ability to handle both sourcing and measurement normally makes them the best solution for testing LEDs, most SMUs on the market simply can’t deliver the level of power that testing high power HB-LED modules requires. Most instrument-based SMUs are capable of delivering only 20W of power or less, but this application often requires 100W or more. Keithley’s 2651A instrument is capable of delivering up to 200W of continuous DC power and up to 2000W of pulsed power.

Pulse-width modulation and HB-LEDs

FIG. 5. Pulse-width modulation (PWM) offers a way to control the brightness of LEDs. Although it’s also possible to control an LED’s brightness simply by lowering the forward drive current, this method is undesirable because the color of the light produced will change slightly with current level. PWM is the preferred technique because it uses a constant current level for each pulse and therefore offers greater consistency in the color of the light produced. PWM also offers more-linear control over brightness, and greater power-conversion efficiency.

In this technique, the current through the LED is pulsed at a constant frequency with a constant pulse level, but the width of the pulse is varied (Fig. 5), which changes the amount of time the LED is in the ON state, as well as the perceived level of brightness. The LED is actually flashing but at such a high frequency that the human eye can’t distinguish it from a constant light level.

Given that LEDs are often used with PWM, it’s only appropriate that they be tested with PWM techniques. As part of PWM testing, an LED is usually tested by running a series of pulses through it while using a spectrometer to take an integrated measurement of the light output over the course of many pulses. This measurement may take tens or hundreds of milliseconds to complete. During the pulsed output, the forward voltage is measured on every pulse to look for changes as the temperature of the LED rises.