Test the part the way it will be loaded in use. A coolant reservoir or a pressurised water tank sees pressure from the inside, so a pressure test on the inside is the natural check. A part that is already filled with liquid, or one that has to be checked without anything entering it, is usually better tested under vacuum from the outside. Either method will find a bad weld. What separates a useful test from a frustrating one is how well temperature, part stretch and the port seals are handled.
What each test is actually measuring
In a pressure decay test the part is sealed, filled with air to a set pressure, isolated, and watched. If the pressure falls by more than an allowed amount over the test time, the part fails. The instrument does not see the leak itself. It sees a pressure change, and it assumes that change came from air escaping.
A vacuum test turns this round. The part, or the space around it, is evacuated and the change in pressure is watched. A differential pressure vacuum tester compares the test side against a reference, which helps it pick up small leaks in liquid-filled containers, sealed products and other parts where a pinhole or a crack in the weld would be a problem. It is non-destructive and nothing has to be pushed into the part.
Both are indirect. A basic pressure or vacuum reading tells you that the part leaks under those conditions, not where. Finding the faulty stretch of weld takes extra features in the tester or a separate check, and getting a trustworthy pass or fail in the first place depends heavily on the fixture and test setup.
Five things that make a leak test lie
The part is still warm. A hot plate or vibration welded tank comes off the machine warmer than the room. Air inside it cools during the test, the pressure drops, and a good part fails. Either give the part a consistent cooling time before testing, or build that time into the line so every part is tested at the same point after welding.
The walls stretch. Thin-walled plastic tanks are not rigid. Under pressure they bulge slightly, the internal volume grows and the pressure falls for reasons that have nothing to do with a leak. A stabilisation period before the measurement starts, and fixture support on large flat walls, keep this under control.
The seals leak, not the part. Every hose port, filler neck and sensor boss has to be sealed by the fixture. A worn seal or a port plug that sits slightly off-center shows up as a leaking part. A sealing test fixture built around the product's actual air ports, with clamping that does not distort the part, is the difference between rejecting bad welds and rejecting good ones.
The volume is large. The bigger the air volume in the test, the smaller the pressure change a given leak produces. Large tanks need longer test times or a different approach to reach the same sensitivity, and that has to be weighed against cycle time.
The test does not match the risk. A tank that only ever holds liquid at atmospheric pressure does not need the same test as a pressurised reservoir. Setting the limits from what the part has to survive in service, rather than as tight as the instrument allows, keeps the reject rate honest.
Matching the test to the part
Pressurised automotive reservoirs and tanks. A pressure test from the inside reflects how the part is loaded on the vehicle. The fixture has to seal every hose spigot and the filler neck.
Appliance water tanks. Usually a pressure test, set to what the tank sees in use. Large flat walls on these tanks are where wall stretch causes false fails.
Washing machine balance rings filled with brine. Once the ring is filled and sealed, pushing air into it is no longer practical. Vacuum leak testers are built for liquid-filled parts like this. On one balance ring line we delivered as a single project, the servo hot plate welder was followed by brine dosing, a leak test station and ultrasonic sealing.
Sealed housings with electronics inside. A vacuum test avoids pressurising the inside of a part that may contain components or membranes that do not like it. The choice depends on the design, so check it on sample parts.
Where testing sits in the welding line
The leak test is the check on the weld, so the two should be planned together. Weld depth control, cooling time and part handling all influence what the tester sees. When the welder and the test station come from different suppliers, a drifting weld and a drifting test can be hard to tell apart.
At SHNEHP we build both sides. On the welding side there are servo hot plate, vibration and ultrasonic machines. On the testing side there are air tightness testers for plastic tanks, welded containers and sealed housings, designed to simulate real operating conditions and to help locate where a part leaks, vacuum leak testers that work on differential pressure, and custom sealing test fixtures in CNC-machined aluminium that seal all of a part's air ports. The pressure decay and vacuum principles described earlier are general test methods, not a description of how any one of these machines measures. Specifications for the test equipment are set per project around the part.
If you have welded tanks that need a leak test, or a test that keeps rejecting parts you think are good, send us the drawing, the material and what the part has to hold in service. If you can send parts, we will weld samples free of charge, and we can look at the test method for the same part at the same time.
SHNEHP works with service partners in the USA, Brazil, India, Thailand, Egypt and Qatar, and we respond remotely within 24 hours.
Request a Technical Consultation
Send your part drawing, sample information, plastic material and target cycle time. Our engineers will recommend a practical hot plate welding machine configuration.
Information in this article is taken from: /customized-machines/, /welding-and-testing-fixture/