Dig deeper to find root cause of leaks

No part is 100% sealed. Whether it is meant to or not, every fabricated product leaks to some degree, even if it’s a tiny amount.

By Bruce Takasaki, Sciemetric Instruments May 11, 2011

No part is 100% sealed. Whether it is meant to or not, every fabricated product leaks to some degree, even if it’s a tiny amount. The key is to determine whether the leak rate—the volume of gas escaping per minute—is acceptable within the parameters set by internal product-quality standards or by industry and governmental regulations. To avoid field failures and optimize critical manufacturing processes, it’s also essential that manufacturers uncover all the root causes of an unacceptable leak, something that the leak rate alone will not provide.

Only through comprehensive data collection and analysis can the root causes of leaks be determined and rectified. Many leak-testing methodologies simply do not provide manufacturers with the amount of insight required to determine root cause.

Manufacturers that employ systems that provide extensive data about leaks can critically analyze that data to gain a comprehensive view of the leak characteristics and can discover what is causing the leak to fall outside of acceptable parameters. This helps improve yield, as leak testing is often a bottleneck on the production line.

There are different methods of leak testing, but for the purposes of this article, a best-practices approach to pressure decay leak testing will be explored.

Stages of the pressure decay leak test

A typical pressure decay test comprises four stages. Following a system check, the fill cycle begins, where the part under test is filled with a gas such as air. This portion of the test may be completed using a fast fill or a regular fill cycle. Then, the stabilization stage takes place, to allow the pressurized part to reach a thermal equilibrium.

After that, the inlet valve is closed, isolating the part from the pressurized gas source, and the pressure within the part is allowed to decay naturally as the gas leaks from the part. The leak rate is derived from a measurement of the pressure decay rate during this portion of the test. In the final stage, the remaining gas is exhausted from the part.

While the pressure decay portion provides the actual leak rate measurement, insight into the part characteristics and the test itself can be gained by careful examination of each of the test phases.

Signature analysis

Traditional pressure decay leak tests derive the leak rate from the difference in pressure between two points in time. These simple two-point measurement systems are prone to error. By relying on limited data from just two moments during the test, this approach is far more susceptible to the effects of noise generated within the pressure transducers themselves.

A more comprehensive and accurate approach to pressure decay leak testing is to store the entire pressure decay curve and apply signature analysis. In this approach, the pressure is monitored and analyzed throughout all four phases of the leak test, to create a “process signature” or “waveform.” Signature analysis is the application of mathematical algorithms to these process signatures to extract key values or indicators, which can then be used to capture and identify defects. These values can be compared against acceptable limits to determine pass or fail status.

In this case, the leak rate is calculated from an entire section of the leak test waveform, not just two points. Using all the relevant data points available in the curve leads to increased accuracy. The impact of measurement noise is dramatically reduced, resulting in significant improvements in gage RR (gage repeatability and reproducibility).

An additional benefit to this approach is that it catches the hard-to-find defects and supports the discovery of root causes of quality issues. For instance, the shape of the fill curve can indicate where the problem is occurring; the seals between the test station and the part being tested might have become worn and started to leak in spurts causing sudden, quick changes in pressure. By 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. A complete waveform, properly analyzed, will isolate these specific causes.

Optimizing the leak test

At a high level, the most important elements in determining the efficacy of a leak test setup are to understand the characteristics of the part being tested, understand the variables of the test process, and then carefully calibrate that process to deliver a desired outcome. Then, it’s essential that the pressure decay waveform be studied and analyzed.

A solid understanding of the characteristics of the part being tested is imperative. There are many external variables that impact part quality and impact the integrity and results of a leak test. First and foremost, though, the characteristics of the part itself must be well understood.

Are the walls of the cavity flexible? Will they deform under pressurization? If so, will the volume change as the pressure changes? This will affect the pressure decay characteristics since the pressure will decrease not only because the amount of gas is decreasing (leaking) but also because the volume and pressure are changing together.

In addition to the malleability of the part, it’s important to know the temperature sensitivity of the part. Does it absorb heat easily? Changes in temperature cause changes in pressure and so will affect the pressure readings.

It is important to keep in mind that temperature, pressure, and volume are related through the Ideal Gas Law, which states that PV =nRT where P = pressure, V = volume, n = number of gas molecules, T = temperature, and R = is a constant.

So, if the temperature rises during the pressure decay portion of the cycle, it will cause a corresponding increase in the part pressure that will reduce the measured pressure decay rate. The more temperature sensitive your part, the more precautions you’ll have to take to ensure that the part is maintained in a stable thermal environment.

A well-understood test process

In addition to knowing the attributes of the part being tested, in particular any characteristics that could impact the leak test data, it’s also important to understand the test process well. There are different approaches to the various cycles—a fast fill rather than a regular fill cycle, for instance—that should be identified in order to fully understand the test and the results.

As noted above, one important variable that impacts the test and its results is temperature. Is the test environment’s temperature variable such that it could cause temperature changes in the part being tested? Even doors that open to other areas in the plant or perhaps the outdoors need to be considered, as a cold breeze can render a test result inaccurate.

Most testers take a little time to thermally stabilize after a cold start and so should be allowed to run through a number of cycles until the tester has reached thermal stability and can produce repeatable measurements. Then the tester should start his or her measurements.

Careful calibration is key

In pressure decay, the leak rate is not measured directly but rather is calculated from the rate of change in cavity pressure. This must be calibrated against a known leak standard. The use of a specially constructed “master” part for calibration purposes rather than production parts can introduce significant problems.

By way of illustration, consider a company that machined a perfect replica of one of its plastic cartridges out of aluminum because it wanted something durable. The aluminum master was useless as a calibration part because its characteristics, such as heat transfer and material flexibility, were completely different than the production plastic cartridge.

A good approach is to get a selection of known good parts and keep them for calibration testing. That way, the standard is based on the actual characteristics of the part in production, something that is prone to the same sensitivities, reacts in the same manner to variables, and will provide legitimate, accurate comparison of ideal parts with those in production.

It is also very important to calibrate to the exact same parameters as production, using the same test pressure, same time durations, and other actual production variables. Any deviation from these parameters will impact the accuracy of test results.

As described earlier, the typical pressure decay waveform contains four zones: fill, stabilize, test, exhaust. Each of these zones has a predictable set of characteristics, represented by the shape of the curve. Deviations from these characteristics can indicate problems with the test apparatus, or specific defects in the part.

Recognizing these characteristics can save hours in trying to figure out why all your parts are suddenly failing. Or, even worse, why they aren’t failing at leak test but are failing downstream.

A comprehensive view

A best-practices approach to leak detection requires a comprehensive view of the leak test process and a solid understanding of the characteristics of the part being tested and how those attributes may impact test results. Manufacturers should use actual production parts and real production parameters for testing purposes as specially fabricated “masters” and unrealistic production variables will result in inaccurate results.

Every process—leak testing included—generates a waveform that maps the ideal process. Engineers tasked with configuring the leak test and analyzing the data should be very familiar with the four zones of the leak test process and be able to readily identify deviations from the norm and what sorts of defects might be detected within each phase of the test.

Manufacturers that adopt signature analysis leak testing have more insight into their leak tests because they have more test data available for analysis. Failure modes that might otherwise go undetected can be identified and diagnosed. This enables the development of additional process checks to minimize downstream failures. Checks can be applied in real time on the manufacturing floor, instantly improving downstream yields.

Modern leak test equipment that uses signature analysis quickly delivers a return on investment by eliminating quality issues at the source while identifying defects that might otherwise go undetected. As the leak test process is improved, productivity is boosted, as is product quality.

– Bruce Takasaki, PhD, is product marketing manager at Sciemetric Instruments.