The concept of a test sequencer is not entirely new. Essentially software subroutines residing outside the PC that execute a sequential list of test operations, test sequencers have long been available in higher-end automatic test and evaluation (ATE) systems that can cost upwards of $1 million. Today, test sequencers are becoming more common in less expensive instruments as well. Here's a look at how test sequencers can speed up tests on optical devices, saving time and money.
A test sequencer is essentially a sequential list of test operations that a system performs on a device. Test sequencer code is pre-loaded into a test instrument, which saves the instrument and the device handler from having to communicate with the PC continuously. The approach automates the test process and shaves precious time off the test cycle.
In a typical test setup without a test sequencer, a PC communicates with a test instrument via an IEEE-488 general-purpose interface bus (GPIB) link and with a device handler through a serial link. The PC receives a signal from the device handler indicating that a component is in place and that the test can begin. The PC then sends a command to the instrument to make a measurement. The instrument sends data to the PC, at which point the PC determines whether the device under test (DUT) is in tolerance or not. If the DUT it is out of tolerance, the PC sends a signal to the handler to reject the part and another test can begin.
The main problem with this type of setup is the slow, continuous data communication between the PC and the instrument. In contrast, the test sequencer approach involves a connection from the PC to the instrument and from the instrument to the device handler (see figure). Eliminating the link from the PC to the device handler removes a significant communications bottleneck by enabling the instrument to communicate directly with the handler. The test sequence is downloaded to the instrument from the PC, which is a one-time event. When the device handler places a part into position, it sends a signal to the instrument to begin the test. The instrument performs the test, makes the pass/fail decision, and sends instructions back to the handler. This flow eliminates the continuous transfer of data over the GPIB, and therefore decreases test time.
In a typical production test setup, a test sequencer residing in the instrument eliminates a link between the PC and the device handler, speeding up the test process.
In most cases, an improvement in overall test throughput of as little as 10 to 100 ms per part has a direct impact on the cost of test, and thus the overall profitability of a manufacturing line. In addition, the quantity of products shipped goes up because testing is quicker, which helps a manufacturer add capacity without investing additional capital.
Instruments with test sequencers can run up to 100 complete test sequences without PC intervention. Each test can include source configurations, measurements, conditional branching, math functions, and pass/fail limit testing with binning capability. Some units can slow down more sensitive measurements and speed up other measurements to optimize overall timing.
The cost of test systems has been a traditional barrier for start-ups, smaller manufacturers, and those with higher-mix, lower-volume products. Now test systems based on instruments and test sequencers can be developed, deployed, and operated for typically tens of thousands of dollars, versus hundreds of thousands of dollars for large ATE systems. oe
Mark Cejer is a business manager for Keithley Instruments Inc., Cleveland, OH.