Radiators, fuel tanks, mufflers, oil filters, air conditioners-the list of automotive assemblies that must be checked for leaks is endless. So, too, are the technologies for performing leak tests. Given the wide range of materials, test volumes and leak-rate specifications, there’s no one-size-fits-all technology for leak-testing auto parts.
Nevertheless, all automotive parts-large or small, metal or plastic-share some common characteristics in terms of leak testing. For one, there’s the production rate. Some 10.8 million vehicles were assembled in the United States in 2007. That means at least 11 million coolant bottles, 43.2 million wheel rims and 60 million fuel injectors had to be tested for leaks before they were put into vehicles.
“Rims have the highest throughput rate, with an average of one wheel in less than 10 seconds,” says Chris Goebel, director of sales and marketing at ULVAC Technologies Inc. (Methuen, MA). “Oil coolers, radiators, throttle body valve assemblies, brake lines, fuel lines and torque converters all need to be tested in less than 20 to 30 seconds per part.”
Though challenging, such rates are more reasonable than, say, the rate for disposable medical devices, which might have to be tested as quickly as one every 2 seconds, notes Jacques E. Hoffman, president of InterTech Development Co. (Skokie, IL). Better still, automotive assemblies are typically more expensive than medical disposables, so manufacturers can justify the cost of turnkey systems that can meet demanding requirements.
Another challenge common to auto parts is that leak rate specifications have become much tighter during the past decade. State and federal standards restricting emissions of chlorofluorocarbons, exhaust and fuel vapors have forced automotive assemblers to radically tighten leak rate specifications for parts such as fuel tanks, fuel rails, brake lines, torque converters and air-conditioning compressors.
“Before the Clean Air Act [of 1990], fuel tanks were never tested with helium,” says Chuck Wilkinson, corporate marketing director at VIC Leak Detection (Ronkonkoma, NY). A dunk test was sufficient.
This need to “tighten up” certain automotive assemblies has created some confusion among suppliers, and that leads to another aspect of leak testing that auto parts have in common. According to Andrew Tapper, engineering manager at Forward Technology (Cokato, MN), engineers are often unsure about what to set for a leak rate specification. “Mostly, we get the ‘it can’t leak’ response as a leak rate specification,” he says.
The key to establishing a realistic test specification is to understand where and how a part truly fails. “Many people get caught up by numbers, when they really need to focus on failure modes caused by production variances and how those modes can be detected by the testing equipment,” Tapper explains. “In the end, leak testers are process monitors, not rulers.”
Since every auto part poses a unique set of challenges, we asked seven leading suppliers of leak testing equipment to discuss how they would test various assemblies. Here’s what they told us.
Transmission CastingsTransmission castings are usually tested with air using a mass flow instrument. The leak rate specification is approximately 5 standard cubic centimeters per minute (sccm). Depending on the assembler, the test system might be a manually loaded, semiautomatic setup with quick-change tooling, or a dedicated, fully automatic system with robotic parts handling.
“The challenge with a transmission casting is whether it’s machined or not,” says Hoffman. “If it’s a rough casting, sealing the part will be more difficult. If it’s a machined surface, sealing is much more straightforward.”
Radiators and Heater Cores
Although both the heater core and the radiator carry coolant, the leak rate specification for the former (0.5 to 1 sccm) is usually tighter than for the latter (4 to 5 sccm). That’s because a heater core is located in the passenger compartment, while the radiator is not.
“The heater core is right on the edge of what you can do with leak testing equipment that uses air,” says Wilkinson. “For years, it was thought to be overkill to use a helium leak detector for a heater core. Now, it can go either way.”
A flow test is often done in conjunction with the leak test. “To verify continuity through the cooling circuits, you can introduce gas on the inlet side and monitor flow from the outlet side,” explains Wilkinson.
Fuel injectors are among the most difficult automotive assemblies to test for leaks. The test volume is less than 0.1 cc, and the cycle time is less than 2 seconds.
“We’ve tested them successfully using a mass flow system and automated parts handling,” says Hoffman. “The fixture is highly engineered, and the instrumentation is designed specifically for that application.”
Temperature compensation is necessary when testing fuel injectors, which can fluctuate in temperature during testing. This can happen for a couple of reasons. For one, they are usually welded with a laser, so if they’re tested soon after assembly, they may still have some residual heat. Injectors can also become warm due to the high pressures at which they are tested. Some injectors are tested at pressures of 100 psi or more.
“Compressing air into a very small space heats it up,” explains Gordon Splete, account manager for Cincinnati Test Systems (Cleves, OH). “When air temperature rises, the pressure increases. When it cools, the pressure drops, and that can mask a leak if you’re doing a straight pressure decay leak test.”
Fuel RailsFuel rails are usually tested with a tracer gas, such as hydrogen. Like many auto parts, the leak rate specification for fuel rails has become tighter during the past few years. Today, the leak rate limit for a fuel rail is less than 0.5 sccm at a test pressure of 100 psi.
“It’s possible to meet that specification with a pressure decay leak test instrument, but if the part is a little warm or if there are changes in the ambient conditions, testing will be difficult,” says Carl Hardt, applications engineer with ATEQ Corp. (Canton, MI).
Instead, the part is pressurized with hydrogen and placed inside a vacuum chamber. Sensors inside the chamber-one for each injector port-sample the atmosphere above the assembly, looking for increases in hydrogen concentration.
Because they carry a gas, evaporators, condensers and other air-conditioning components are tested to a higher leak rate specification-0.00008 to 0.00002 atmospheric cubic centimeters per second (accs)-than auto parts that carry liquids. These components are usually tested using helium or hydrogen, says David Morris, marketing manager at Alcatel Vacuum Products (Hingham, MA).
“The advantage of using a tracer gas is that you don’t have to worry about the part flexing or changing temperature,” he says. “Another advantage is the ability to locate where the leak is. Obviously, if you have a very expensive component, you want to be able to repair it if it leaks.”
The leak can be isolated using a sniffer probe that can be handheld or mounted to a robot. “If it’s a large leak, the probe can be as much as 1 inch away from the part,” says Morris. “If it’s a very small leak, it has to be within 0.25 inch.”
With an automated rotisserie-style system, up to eight compressors can be tested is less than 1.5 minutes, depending on how many plugs have to be installed and how many hose connections need to be made, adds Goebel. To save time, plugs can be installed prior to testing.
To save on helium, the system can perform a quick pressure decay test using air. “There’s no need to charge the part with helium if it has a gross leak,” says Goebel. “Next, a pressure-rise test is done. The benefit of this test is that the compressor is now void of air, so helium charging is quick and the concentration of helium can be carefully controlled. Approximately 98 percent of the helium can be recovered after the test is completed.”
Maintenance is critical in any leak testing system for compressors, advises Goebel. Residual oil from the machining process can contaminate leak testing equipment and adversely affect the test results.
Every engine compartment contains several plastic reservoirs for various liquids, such as power steering fluid, coolant, and windshield washer fluid. Detecting leaks in these assemblies can be challenging because the plastic expands and contracts with positive or negative pressure.
Depending on cycle times, bottles are tested with air using a mass flow or pressure decay system. The pressure decay method works better with reservoirs that have thick, rigid walls. The mass flow technique is better for reservoirs with thin, bendable walls.
“In either case, the system automatically compensates for volume changes in the part. If the reservoir expands when pressurized, it could mask a leak,” says Hoffman.
Another challenge is to integrate function tests into the system, says Tapper. For example, some reservoirs are equipped with switches that alert the driver when fluid levels are low. To check if the switch is working correctly, Forward Technologies created a fixture that rotates the bottle 180 degrees during the leak test.
Wheel RimsThe leak rate specification for wheel rims is not terribly stringent-perhaps 0.0003 accs at a test pressure of 40 psi. This application requires a high-throughput system, which can be loaded manually or automatically.
A manual system is feasible if the rims aren’t too large or heavy and the distance from the conveyor to the test stand is relatively short. “In many cases, a rotisserie is used, in which the rim follows a circular test path. When the rim has rotated 360 degrees, it is fully tested,” says Goebel.
If automation is an option, a multi-chamber system can be set up. In this case, the rims are transported on a conveyor. Two robots can shuttle rims in and out of the chambers: One can be loading while the other is unloading. “The test involves sealing the top and bottom surfaces of the rim, which is laid on its side,” says Goebel. “Helium is introduced on one side and sensed on the opposite side to determine if there is a leak.”
System calibration is important. Before each shift start or after any break period, the test system should be recalibrated to ensure that pass-fail limits are correct and up to date. “We’ve worked with several rim manufacturers where the working environment changes dramatically from 7 a.m. to noon,” warns Goebel. “Temperature, humidity and background levels of helium need to be considered and compensated for.”
Air filters, especially carbon-filled models, pose a unique challenge for leak testing equipment. “The carbon acts like a sponge, creating a part that takes an extremely long time to stabilize for pressure decay monitoring,” says Tapper. “To solve this, we enclose the filter in a chamber, pressurize the filter internally to a specified pressure, and pull a small vacuum externally. Then we monitor that vacuum for decay.”
Like the heater core, the gas cap is “on the border” between needing a pressure decay system or a helium mass spectrometer system, says Wilkinson. With the latter, the cap would be fixtured, then pressurized from the tank side of the assembly. The spectrometer would then look for helium escaping into a leaktight chamber enclosing the outside of the cap.
“We offer a handheld sniffer for helium, so you know where the gas is escaping from,” adds Wilkinson. “However, most high-volume suppliers want to avoid sniffing and use an operator-independent test. If there is a leak, they can always go to the identification mode to locate it.”
With a leak rate specification of 0.00001 to 0.0001 accs, fuel tanks are tested using a helium mass spectrometer system. Depending on the size of the tank, test time ranges from 45 to 65 seconds.
“You could use the helium accumulation method, but the part is so large, it would take a long time for helium to accumulate to a measurable amount,” says Splete. “You can test it much more quickly with mass spectrometry.”
The challenge with testing fuel tanks is that they can only withstand a positive or negative pressure differential of 1 or 2 psi without bursting or buckling. To prevent that, the tank and the chamber must be evacuated simultaneously. “Then, we isolate the tank and backfill it with helium,” says Wilkinson. “That way, the tank is at an elevated pressure relative to the vacuum chamber, but within the allowable pressure differential of 1 to 2 psi.”