Air Leak Testing Equipment Pressure Decay Airtightness Tester 0.1 Pa Resolution For Valve Pipe Assemblies

Most parts that need a leak test do not need helium. They need a number, a repeatable verdict and a cost per test close to zero. This is the machine for that band of work: dry compressed air, differential measurement against a master part, two independent channels, and a controller that refuses to blame the part when the fixture is at fault.

Where the money actually goes

A helium system is not expensive because of the analyser. It is expensive because every single test consumes a gas that has to be bought, stored, recovered and purified, and because the vacuum hardware that makes the measurement possible needs maintenance.

On a part where the criterion is a bubble count or a millilitre per minute, that money buys you nothing. A pressure decay test on the same part uses shop air and electricity. For a mid-size valve body tested twice a minute, two shifts a day, the difference is not a rounding error — it is a line item that shows up in the annual cost review.

The honest framing is this: spend on helium where it changes a decision, and spend nothing extra where it does not. Many of our customers run an air tester at the end of the line as the routine gate, and keep a helium chamber in the lab for design validation and for the parts that fail.

What pressure decay can see, and what it cannot

Air testing resolves gross and medium leaks — roughly 1×10^-3 mbar·L/s and above, depending on part volume and the time you are willing to give it. That covers the overwhelming majority of production defects: a missing O-ring, a porous casting, a cracked weld, a valve seat that never lapped in.

What it cannot see is the fine band — the hairline braze void, the moulding weld line that weeps under thermal cycling, the seal that passes today and fails in eighteen months. If your customer's criterion lives in that band, no amount of clever signal processing will rescue an air test, and you should be buying helium.

The four design details that decide whether the result is real

Differential measurement against a master part

Two identical parts, one known good, are filled to the same pressure and isolated. The sensor reads the difference between them, not an absolute pressure. Ambient temperature drift, which is the single biggest source of false rejects in air testing, affects both sides and cancels. Resolution is 0.1 Pa on the difference.

Fixture seal verification

A damaged fixture seal looks exactly like a leaking part. The controller tests the fixture itself before every run and refuses to judge the part if the fixture fails. This one feature removes most of the "the machine is wrong" conversations we used to have.

Temperature compensation on the fill transient

Compressing air heats it, and the part cools back down over the first seconds of the test. That curve looks like a leak to a naive algorithm. The controller models the transient and waits for the slope to settle before it starts measuring.

Flow test as a second mode

Some failures are the opposite of a leak: a blocked passage. The same pneumatics can hold a fixed pressure and measure flow, which catches swarf, flash and a seal that has been fitted the wrong way round.

How a cycle runs

• Load and clamp — part into the fixture, barcode scan, pneumatic clamp closes the seals.
• Fixture check — the controller verifies its own seals before pressurising the part.
• Fill and stabilise — air to the recipe pressure, then a settle window while the thermal transient decays.
• Measure — differential pressure is logged over the test window and converted to a leak rate; the verdict is compared against the limit.
• Release and record — part released, result stored with serial number, SPC trend updated.

Technical parameters

What it finds

Setting the limit without guessing

• Use a real master part, not a solid block. A block cancels temperature drift but ignores the compliance of the actual part, and the reading becomes optimistic.
• Calibrate with an orifice, not an opinion. Fit a known leak standard in place of the part and confirm the machine reads what it should. Repeat it monthly; it takes ten minutes.
• Pick the test time from data. Run ten good parts and ten rejects. If the two populations overlap, lengthen the test rather than tightening the limit — a tighter limit on a noisy measurement just rejects good parts.
• Write the method on the drawing. Pressure, stabilisation time, test time, and the reference part. A limit without a method cannot be audited.

Compared with the alternatives

Delivery, installation and the first month

• Fixture concept reviewed against your part drawing before the order is confirmed.
• Master part selection and reject-limit study during commissioning, on your own parts.
• Orifice calibration and a repeatability run, documented for your quality file.
• Operator training on loading, reading the curve and what to do when the fixture check fails.
• 24-month warranty on controller and pneumatics.
• Spares in Guangzhou: seals, solenoid valves, pressure sensors, touch screens — 48-hour dispatch.
• MES handshake over Modbus TCP, with barcode login for traceability.
• Annual calibration against a certified pressure standard.

Frequently asked questions

About Guangzhou Hejin Automation Control Technology

We have built test equipment in Guangzhou since 2003, with 50–100 staff and about half of our output exported. Our range covers electrical safety testers, material test machines and leak detection — air, helium and tracer gas — which means the recommendation you get is based on what fits the part rather than on what we happen to have in stock.

Machines are assembled, pressure-tested and calibrated before shipment, and each one is recorded under its serial number so a question about a five-year-old unit can still be answered with data rather than a guess.