Leak tests can be a frustrating part of an automotive manufacturer’s quality assurance process. Whether it’s a DIY setup or a third-party platform, there are often challenges around cycle times, basic quality thresholds, and tracing the root cause of a flaw, or even just being confident that a problem isn’t in fact a test malfunction.
That’s because many manufacturers continue to rely on isolated scalar data points for their leak tests, often with DIY test systems with little or no data visualization capability. This approach can red flag potential problems early, but it simply can’t give the comprehensive data insight that’s required to cut failure rates and to increase the repeatability, accuracy and speed of the leak test. For that, more test data must be collected and visualized. One approach is waveform, or signature, analysis.
Figure 1: In this example, a die-cast aluminum transmission case warmed by a wash is generating false negatives on the subsequent leak test. To compensate for this kind of result, manufacturers will often raise their allowable leak limit to maintain production volume, but this undermines confidence in the test. With waveform analysis, the manufacturer could correlate part temperature to test results and make the adjustments needed to arrive at an actual and repeatable leak test rate. The blue line shows the unadjusted leak rate of the transmission case, which steadily declines as it cools. The red line is the temperature-compensated leak rate.
Signature analysis captures the entire leak test waveform that’s hundreds of thousands of data points per cycle. It won’t just catch if the part leaks and how quickly for a simple pass/fail. It visualizes in real time if there are anomalies that may indicate a problem, from an obstruction or delay in filling the part, to a seal failure during the test, or a delay in emptying at the end of the test. It even reveals where there may be an opportunity to shorten the test cycle.
Slashing cycle times and test costs
A common leak test issue on many lines is adjusting for differences in temperature between the part and the ambient air. This is common when parts are coming in cold off a truck, or are warm from having just come out of a wash. Either scenario can skew leak test results.
In one case, a manufacturer was having trouble achieving repeatable leak test results in the required time for a die-cast aluminum transmission case that was warm from a wash. Waveform analysis was used to correlate part temperature with the measured leak rate, to eliminate the false failures and arrive at the actual leak rate within the required test time (Figure 1). By using waveforms to track, analyze and adjust for the effect of part temperature on the test, the manufacturer substantially improved the leak test’s accuracy and repeatability.
With the actuator leak test at another manufacturer, waveform analysis and excellent test repeatability allowed the test team to determine, based on the leak rate of air under pressure, if actuators were meeting spec for piston to bore clearance with test specifications ranging from ~0.0000 to 0.003 inches (Figure 2 above). This boosted quality and cut in half the number of leak testers required to keep up with production.
Figure 2: In this waveform, the leak rate of air under pressure is being used to determine if actuators are meeting spec for piston to bore clearance, with four discrete clearance specifications
Avoiding pricey warranty repairs
At another manufacturer, the problem was oil leaks in axles that often wouldn’t show up until after final installation in the vehicle. In this case, that vehicle was off-highway a tractor but an axle is an axle. Waveform analysis was crucial to catch a potential leak on the production line, to avoid aftermarket warranty repairs that could cost well over US$5,000 per unit. The manufacturer also gained a better understanding of how environmental conditions in the plant, such as temperature, could impact test process and results, and adjust accordingly.
The waveform doesn’t lie
A waveform just isn’t capable of keeping a secret. Signature analysis can greatly improve the speed, accuracy and repeatability of a leak test, to boost production line yield and reduce costs by ensuring your test is always optimized. And when a problem does arise, quality engineers have the big data insight they need to take corrective action in just minutes or hours, instead of spending days or weeks wading through piles of spreadsheets.
Robert Ouellette, product launch manager at Sciemetric Instruments, works firsthand with customers to solve pressing technical challenges in leak testing and other technical areas of manufacturing.
October 6, 2016