Page 26

EETE APR 2014

TEST & MEASUREMENT Signal management for functional test By Bob Stasonis Functional test has always been used for verification of overall functionality, calibration information, and certification for high-risk applications such as medical devices. Newer test implementation strategies are also driven by factors such as budgets, throughput, BIST (Builtin Self-Test), and Unit-Under-Test (UUT) designs. It is the latter that has the most influence on what can be tested. Budgets and throughput limit what will be accomplished. Fig. 1: Increased package density can also lead to increased I/O One area that often does not get enough consideration is what I call “Signal Management”. Basically, this is the design and/ or configuration of the signal switching system for test. Signal management allows the system to share test resources and test multiple UUTs simultaneously. Properly configured, it helps the test system designer minimize the ATE (Automatic Test Equipment) Fig. 2: Basic switching configurations – SPST, DPST, SPDT are available in shielded and un-shielded versions. footprint and save on the bottom line. What are your UUT’s test requirements? Before looking at designs and processes, as well as signal management, you should know the UUT. This includes the type of product, its configuration (single PCB, panelized PCBs, final product), the test specifications, the planned test points/pad size, the anticipated quantity (per line/day/shift, etc.) and the anticipated fault spectrum. Obviously, I left the word “Budget” out. But you can’t determine how much something will cost to test until you understand it. Negotiations for funding should not start until you understand what it will take to test the UUT. Only then can you reach the necessary compromises. High I/O and functionality I’m not revealing anything new when I say that electrical access is getting more difficult. The increasing functionality per UUT, coupled with shrinking geometry adds a problem. Because of the small PCB size, these devices may be tested while still in the pallet. So, there will either be a set of parallel instrumentation or signal management to route the instruments to the UUT that needs it. Component density does not appear to be an issue for functional Fig. 3: Example of a 2-pole by 99 multiplexer. test. After all, the primary concern here is “Something goes in, something comes out”. While simplistic, this is often the case. Stimulus is applied to the inputs of the UUT, and a specific data set should come out. But component density will be a factor for several reasons. First, higher density is a good indication of broad functionality – be it complex functionality, high I/O count, or a combination of both. Looking at the sample PCB in figure 1, you have to answer these questions first; What is the I/O count? What types of signals (Digital/Analog) are present at the connector(s)? What instrumentation is required? Is access for calibration necessary? Are diagnostics critical? Will any probing be performed by a human or robotic means? Will automated handlers be used? Are the I/O connectors used easily probed or connected to? Once we understand these questions, we can look at the signal management configuration. But, before we discuss this, let’s look at some of the signal switching choices available. Signal switching types Switching can be as simple as an on/off connection. SPST (Single Pole, Single Throw), DPST (Double Pole, Double Throw), and other switch configurations are used to connect devices including power, loads, and mechanical actuators, as seen in figure 2. In most cases, the primary issues here are voltage, current, and power – bandwidth is hardly ever a problem in this application. Multiplexers, or muxes, are used to route a single instrument or stimulus to multiple connections on the UUT, one connection at a time - see figure 3. This is very flexible and cost effective for sharing a single resource with multiple test points, or vice versa. Again, the parameters of voltage, current, and power must be considered as well as the number of simultaneous connections. In most cases, the frequency response of the mux is also important. The most flexible switching solution in many applications is the cross point matrix. Looking at figure 4, a matrix is essentially a series of relays that connect a particular Y-axis electrical connection with the appropriate X-Axis connection. By closing the proper relays, virtually any Y-axis connection can be connected to any X Axis connection. Multiple simultaneous connections are allowed as well. This can be the easiest way to make connections. But matrices do have their limitations. The first issue is bandwidth. Most matrices are limited because the number of possible connections and potentially long un-terminated Bob Stasonis is USA Marketing Manager for Pickering Interfaces - www.pickeringtest.com 26 Electronic Engineering Times Europe April 2014 www.electronics-eetimes.com


EETE APR 2014
To see the actual publication please follow the link above