Almost ThereCycling, Bikes, Parts, Accessories, Clothing and More…

Account limit of 2197 requests per hour exceeded.

High Brightness LEDs Require High-Speed Production Testing

HBLED testing involves different types of test sequences at various stages of production, such as during design research and development, on-wafer measurements during production, and final tests of packaged parts. Testing "recipes" often include a multitude of steps, but this article focuses on the basic "ingredients" for these recipes—essential tests that illustrate how to probe for LED characteristics and examples of test setups. The emphasis is on electrical device characterization, but includes light measurement techniques where appropriate. Moreover, according to Keithley Instruments, http://www.keithley.com/ , by applying new test technologies, test times and costs can be reduced. 

 

Basic electrical tests involve current-voltage (I-V) measurements to get a picture of an LED's operating characteristics. The resulting I-V curves look pretty much like those of a conventional diode. A complete test could include a multitude of voltage values versus current operating points, but a limited sample of points is generally sufficient to probe for an LED's figures of merit.

 

Some tests require the application of a dc current source to the LED, and then measuring the resulting voltage; other tests require the application of a dc power supply and measuring the current through the LED. A source-measure unit (SMU or SourceMeter® instrument) is ideal for these types of test, because it combines a precision power supply with high accuracy measurements similar to those that can be achieved with a precision digital multimeter (DMM). Simply put, these instruments can function as a dc power source to supply voltages or currents, and also measure each of these signal types.

 

Forward Voltage (VF) Test and Optical Tests. The VF test verifies the forward operating voltage of an LED. A VF test is performed by sourcing a known current and measuring the resulting voltage drop across the diode. When the forward current is applied to the diode, it begins to conduct. During the initial low current source values, the voltage drop across the diode increases rapidly, but the slope begins to level off as drive current increases. The diode normally operates in this region of relatively constant voltage, which is why it is useful to test the diode under these conditions. Typical test currents range from tens of milliamps to amps, while the resulting voltage measurements are typically in the range of few volts. The results of this test are typically used by manufacturers for binning purposes as the forward voltage is directly related to the chromaticity (color characteristic) of the LED.

 

Forward current biasing is also used for optical tests because electrical current flow is closely related to the amount of light emitted. Optical power measurements can be made by using an integrating sphere or placing a photodiode close to the LED under test to capture the emitted photons. This light is then converted to a current, which can be measured by an ammeter or one channel of a SourceMeter instrument.

 

In many test applications, the voltage and light output of the diode can be measured simultaneously using a fixed source current value. In addition, details such as spectral output can be obtained by using the same drive current value and a spectrometer.

 

Reverse Breakdown Voltage (VR) and Leakage Current (IL) Tests. Applying a negative bias current to the LED will allow probing for the so-called Reverse Breakdown Voltage (VR). The test current should be set to a level where the measured voltage value no longer increases significantly when the current is increased slightly more. At levels higher than this voltage, large increases in reverse bias current result in insignificant changes in reverse voltage. The specification for this parameter is usually a minimum value. The test is performed by sourcing a low-level reverse bias current for a specified time, then measuring the voltage drop across the LED. The measurement result is typically in the range of tens of volts.

 

Normally, moderate voltage levels (volts to tens of volts) are used to measure a Leakage Current (IL). The IL test measures the low-level current that leaks across the LED when a reverse voltage less than breakdown is applied. It is a common practice for leakage measurements, and more generally for isolation measurements, to make sure a certain threshold is not exceeded in production. There are two reasons for this. First, low current measurements require longer settling times, so they take longer to complete. Second, environmental interference and electrical noise in the test environment exert greater influence on low level signals, so extra care in shielding is required. This extra shielding complicates the test fixture and may interfere with automated component handlers.

 

Single LED Test System. An automated single LED test station would typically consist of a PC, component handler and dual-channel SourceMeter instrument, such as a Keithley Model 2602A. If the tests are being done at the wafer stage of production, then a wafer probe station would also be involved.

 

In such a test system, the main purpose of the PC is to store measurement data in a database for documentation. A secondary purpose is to reconfigure the test sequence on the SourceMeter instrument for different parts. However, a Keithley Model 2602A can actually operate independently from the PC controller. It has a Test Script Processor (TSP) that supports writing a complete test plan to operate on the instrument itself. In other words, a user can write a complete PASS/ FAIL test sequence script and run it from the front panel of the unit without instrument reprogramming.

 

In a more production-oriented setting, the component handler would transport individual LEDs to the test fixture where electrical contact to the SourceMeter is established. The fixture is shielded from ambient light and houses a photo detector for light measurements. In this setup, a single Dual-Channel SourceMeter can be used for both source and measure connections. Channel A can supply the test signal to the LED and measure its electrical response, while Channel B monitors the photo detector during optical measurements.

 

The test sequence can be programmed to begin using a digital signal line from the component handler, which provides a "start of test" (SOT) signal. When the SourceMeter instrument detects the SOT signal, the characterization tests for the LED begin. After all electrical and optical tests are completed, a digital line from the SourceMeter instrument sends a "measurement complete" signal to the component handler. In addition, the SourceMeter instrument's built-in intelligence can perform all pass/fail operations and send a command through its digital I/O port to the component handler instructing it to bin the LED based on pass/fail criteria. Usually, two subsequent actions take place synchronously: data transfer to the PC for statistical process control (SPC), and mechanical placement of a new LED in the test fixture.

 

LED Test System for Multiple Devices/Arrays. Large scale production testing of multiple LEDs and LED arrays may involve a burn-in process. In these tests, multiple parts are measured over a specified time period. A continuous current flow is usually mandatory to drive the LEDs, but multiple light detectors might be multiplexed to a current meter by a switch system. The appropriate instrument and switching system and will be dictated by the dynamic range of electrical currents used in the tests, and the number of devices being tested in each sequence.

 

For example, Keithley's Model 3706 System Switch/Multimeter offers six switch card slots, which can handle up to 576 multiplexed channels or 2688 switch matrix cross-points. It also has TSP capabilities, making it a good companion to the company's SourceMeter instruments. Their TSP-Link® feature allows both instruments to be quickly and easily integrated for tight synchronization of operations, and both can be operated from a single test script to maximize test throughput.

 

Programming Tests for Speed Using TSP. With many instruments, the PC has to control all aspects of the test. If so, in each element of a test sequence the instruments must be configured for each test, then perform the desired action, and subsequently return data to the controlling PC. The controlling PC must then evaluate the pass/fail criteria and perform the appropriate action for LED binning. Each command sent and executed consumes precious production time and lowers throughput.

 

Obviously, a large percentage of this test sequence time is consumed by communicating information to and from the PC. Instruments with TSP capabilities dramatically decrease the amount of traffic over the communications bus, and thereby increase system throughput for complicated test sequences

 

With TSP, the majority of the test sequence is embedded in the instrument. TSP is a full-featured test sequence engine that allows control of the test sequence, with internal pass/fail criteria, math, calculations, and control of digital I/O. TSP can store a user-defined test sequence in memory and execute it on command. This limits the set-up and configuration time for each step in the test sequence and increases throughput by lessening the amount of communications to and from the instrument and PC

 

Using TSP involves a simple 3-step process:

1)   Create the script.

2)   Download the script to the instrument.

3)   Call the script to run.

The test script can be written with Test Script Builder software provided with each SourceMeter instrument, or it can be downloaded to the instrument using another program, such as Visual Basic or LabVIEW.

 

More details on HBLED testing, including test system diagrams, sample TSP scripts, and tips on avoiding measurement errors are available in an application note, "High Speed Testing of High Brightness LEDs", available for free download at

 www.keithley.com/data?asset=50328.

 

About the Author

Charles Cimino has been with  <a href="http://www.keithley.com/">Keithley Instruments</a>  in Cleveland, OH for more than 20 years in various product and business development roles. Currently he is a Business Manager responsible for the company's Precision DMM and Signal Switching products. He has an undergraduate electrical engineering degree and an MBA degree from Case Western Reserve University's Weatherhead School of Management. He can be reached at ccimino @keithley.com, or at 440-248-0400.