Quality Test & Inspection: Data Leads to Good Leak Test Decisions
Leak testing is an important and widely deployed nondestructive test methodology used by many manufacturers to assess the quality of fabricated parts. The fact is that nothing is 100% sealed; everything leaks, whether it is supposed to or not. The challenge for manufacturers is to determine whether the leakage is acceptable from product quality and regulatory compliance perspectives.
The majority of leak testing systems that are available today focus primarily on determining a single value that characterizes whether the leak rate is acceptable under internal standards and regulatory directives. While this traditional method may detect 80% of the typical defects, the other 20% will typically represent 80% of the after sale warranty cost impacts.
Manufacturers also need to be able to determine the cause or origin of the leak, which is not possible with just the leak rate figure. More data must be collected and then analyzed in order to determine the characteristics of the leak. Unfortunately, most systems are not set up to gather the necessary data, nor do they offer detailed analysis of the leak characteristics.
Through the adoption of new leak testing approaches that provide comprehensive data about the entire leak test cycle, manufacturers can get more out of their leak test.
Taking the old standards to a new levelThe efficacy of a manufacturer’s leak test is determined by a number of factors, including the condition of the test setup, how long the test cycle takes, how much information the test provides, whether all parts are being tested or if just a statistically representative sample is being collected, and whether manufacturers can determine from their test not only the leak rate, but the characteristics of the leak and what caused it.
How well the test setup is working can greatly impact the results. Take, for example, pressure decay leak testing where the part being tested is filled with air or some alternative gas and then monitored as the gas leaks, measured in standard cubic centimeters per minute (sccm). If the reservoir was not filled to the correct pressure, the assumed pressure differential would be incorrect, rendering the test result inaccurate. Inaccuracies like this can be the result of a test system not functioning properly-for example, the inlet hose was crimped or damaged-resulting in a pass, but the part would not have been properly tested.
The key determinant of a test’s overall effectiveness is how much information the test provides. With limited data, one has limited insight into results that deviate from the standard. Why did one part pass and another fail? If a part has been shipped and is deemed defective after it is in use, yet it reached the desired endpoint of testing, where did it go wrong and why? What additional information about the leak characteristic is required in order to determine the root cause of the quality issue?
In a perfect scenario, the manufacturer has complete visibility into its manufacturing processes and can pinpoint specific variables that impact part quality as each part is tested.
Most of the specialized leak testers on the market are focused on determining the leak rate through capture and analysis of a few precise moments of the test. But what if there were anomalies that took place at other points, due to noise or disturbances, for example, not captured by the test? Those defects would not be caught and faulty product could still make it to market.
By capturing the full leak test pressure curve including fill, stabilization, decay and exhaust, manufacturers can recognize anomalies in a leak pattern and use that data to determine their cause.
Optimization of the test setup is crucial to getting accurate results. Manufacturers are tasked with picking the right portion of the decay curve to get a repeatable and consistent leak rate measurement and establishing the right limits to ensure an accurate and reliable test, with the shortest possible cycle time.
Generally, the early part of the decay curve will show more variability and, therefore, will result in a less repeatable measurement, but waiting until later means a longer test. Balancing the two to come up with the best trade-off between cycle time and repeatability is up to the manufacturer. Whether this test optimization is an iterative, time-consuming process, or one that can be achieved in a software environment in a matter of a few days, is determined by the system itself, which will be explained in more detail.
Insight at every stepOne effective methodology that captures the complete characteristics of the leak is process signature verification. Every manufacturing process produces a signature, a waveform that represents the characteristics of that process from start to finish. Any deviations from the norm, regardless of whether or not the part met the desired endpoint of the test, are readily identified and root causes easily determined by examining the process signatures. This approach is effective in most testing situations, with applications for myriad other tests such as weld, press fitting, crimping, torque, sound and vibration, and functional testing.
Consider again the example of the incomplete fill cycle. This would only show up as a change in the shape of the fill pressure curve, which is not normally monitored. Furthermore, the shape of the fill curve can indicate where the problem is occurring.
For example, the seals between the test station and the part being tested can become worn and start to leak in spurts, causing sudden, quick changes in pressure. In contrast, a blockage or poorly drilled opening in the part will change the slope of the fill curve but without the sudden changes described above.
The complete manufacturing process is captured using both hardware and software components. Hardware on the production floor collects the process signature, then analyzes it to provide real-time pass or fail judgement, while simultaneously sending data to a database that stores the information.
This capture and storage of comprehensive process data supports regulatory compliance, enables selective recalls (if required), and supports the optimization of test parameters in a modeled environment, among other benefits.
Being able to store and re-analyze historical leak test data in software eliminates the need to optimize the test parameters using a live, iterative approach. The key to this approach is having a software package that uses the entire waveform, and not just single point values. The end result is a leak test that provides accurate and repeatable defect detection, while minimizing test time.
Whether to optimize existing test systems or adopt a new one must be a well-examined decision.
Manufacturers can be challenged by the need to meet increasing throughput requirements, particularly if their leak test stations are creating bottlenecks due to lengthy cycle times. Many feel that in order to accommodate their increasing volume, they need to either invest in additional test stations or improve the throughput of their existing leak test method.
Optimizing an existing conventional system means that the test provides the best compromise between accurate defect detection (including avoiding false fails) and test time. The downside is that no matter how optimized the test cycle is the test still provides limited or no information for identifying root causes for leakages and may still miss some failure modes altogether. Also, it will take much longer to optimize the system without signature storage and analysis.
By implementing or upgrading to a system that decreases leak test cycle time, the manufacturer does not have to invest in additional stations or absorb the cost of maintaining and manning additional test stands. Or, for the same amount of test stands, enjoy a significant increase in productivity. An attractive return on investment for making the switch to a better method is quickly realized.
Faulty products cost more than time and moneyIn today’s market, where regulatory bodies and consumers alike demand more insight into quality issues, effective and comprehensive testing modalities are integral to avoiding irreparable brand damage and millions of dollars in penalties and legal settlements.
Manufacturers are wise to use test systems that provide them with the information necessary to rectify quality control issues before hundreds or even thousands of faulty products are produced. Effective systems can provide competitive advantage and support regulatory compliance.
Manufacturers should be aware that without a system that provides comprehensive information about manufacturing processes that can readily determine what caused a problem, they can be faced with lengthy and expensive shut downs until the root cause is determined and rectified. In the case of industries with stringent regulatory considerations, such as the medical field, corrective actions required by regulatory bodies could result in significant losses.
The are several benefits that arise from having more insight into a leak testing and having more leak test data at one’s disposal. With the complete leak test waveform available for analysis, it is possible to identify and diagnose failure modes that might otherwise go undetected, enabling the development of additional feature checks to minimize downstream failures. These can then be applied in real-time on the manufacturing floor, instantly improving downstream yields.
The data also can be uploaded to analytical software, where the impact of test parameter changes can be simulated in minutes. This allows for the rapid optimization of pass/fail criteria, minimizing commissioning times and ensuring maximum yields.
Use of process signature technology that provides comprehensive data about the leak test offers manufacturers significant benefits that impact the bottom line. Investment in modern leak test technology delivers a quick ROI, supports compliance initiatives, improves productivity and ultimately, improves product quality.
For more information on leak testing, visit www.qualitymag.com to read the following:
“Choose the Right Seal”
Quality 101: “Leak Testing: Calibration Matters”
“Trace Leaks with Hydrogen Gas”