Dr.Vakuum
All applications

Leak testing and brake-booster end-of-line test in automotive plants

6 min read

Thailand builds two million vehicles a year and most of the parts in them: brake boosters and master cylinders, fuel tanks, radiators and air-conditioning components, castings, battery housings. Every one of those leaves its line through a leak test, and most leak tests are vacuum tests — the part or a chamber around it is evacuated, isolated, and the pressure watched for a rise. The pump at the test station is sized by one number, the takt time, and judged by another, the repeatability of the result.

1. The methods

Vacuum decay (pressure rise). The part is evacuated to a test pressure, isolated from the pump, and the pressure rise over a hold time is measured. Leak rate Q = ΔP × V / Δt. For a 5 L part held 10 seconds with an allowed rise of 0.5 mbar: Q = 0.5 × 5 / 10 = 0.25 mbar·L/s, equivalent to about 15 cm³/min — a typical limit for a coolant or fuel component. Simple, cheap, and good to about 10⁻³ mbar·L/s with care.

Helium accumulation / sniffing. The part is filled with helium (or a helium-air mix) and placed in a chamber that is evacuated; a mass spectrometer leak detector on the chamber measures the helium that leaks out. Sensitive to 10⁻⁶ mbar·L/s and below — needed for AC components, fuel systems under evaporative-emission rules and battery packs. The chamber is the vacuum duty here, and it is larger than the part.

Brake booster functional test. The vacuum servo of a brake booster is tested for function and tightness by applying the vacuum it will see in service — typically −500 to −700 mbar gauge — and measuring pedal effort and hold. The booster's own volume (3–8 L) is the chamber.

Vacuum filling and test of radiators, AC loops and brake lines: evacuate to remove air and moisture, hold to prove tightness, then fill under vacuum. The test is a by-product of the filling process.

Why does the pump's knee set the takt time?

Because the test pressure is usually close to the pump's ultimate. A vacuum decay test at 50 mbar on a 90 mbar-ultimate pump is impossible; at 50 mbar on a 0.5 mbar pump it is easy. A helium chamber test at 1 mbar on a single-stage pump with a 0.5 mbar ultimate spends most of its cycle in the last decade, where the pump has lost half its speed. The time to reach test pressure is

t = (V / S_eff) × ln[(p₀ − p_ult) / (p_test − p_ult)]

and as p_test approaches p_ult the logarithm grows without limit. A pump whose ultimate is a decade below the test pressure keeps the evacuation on the plateau of its curve, and the takt time becomes predictable. Choosing the test pressure and the pump together is the whole art.

2. Sizing to takt

A booster test station: booster volume 6 L plus 2 L of test fixture and hose, test vacuum −650 mbar gauge (363 mbar absolute), evacuation budget 3 seconds within a 20-second takt.

S_eff = (0.008 m³ / 3 s) × ln(1,013 / 363) × 3,600 = 0.00267 × 1.03 × 3,600 ≈ 10 m³/h

A small pump — a DRV10 at 12 m³/h with a 90 mbar ultimate is on its plateau at 363 mbar and does it with margin. But the station runs 180 cycles an hour, so the fixture's valves open and close 180 times, and a vacuum receiver between pump and fixture smooths the duty: a 50 L receiver held at 200 mbar by the pump dumps the booster to test pressure in under a second through a large valve, and the pump recovers the receiver in the remaining 19 seconds. The pump then runs continuously at a steady load instead of cycling.

A helium chamber test: chamber 60 L around a 10 L fuel tank, test pressure 1 mbar, evacuation budget 15 seconds.

S_eff = (0.06 / 15 s) × ln(1,013 / 1) × 3,600 = 0.004 × 6.92 × 3,600 ≈ 100 m³/h

— on paper. At 1 mbar a single-stage V pump is at its knee; use a two-stage DV90 (90 m³/h, 5 × 10⁻³ mbar ultimate) and the 100 m³/h is real all the way down, or a DV175 for margin. The chamber's roughing pump is separate from the leak detector's own small pump, and the roughing pump should be dry or well-trapped — oil back-streaming into a helium test chamber raises the background and slows the test.

3. Repeatability: the enemies

A leak test station that passes a part at 08:00 and fails the same part at 14:00 has a measurement problem, not a part problem. The usual causes:

  • Temperature. A vacuum decay test measures pressure rise; gas in a part warming by 1 °C rises 0.3 % in pressure — at 363 mbar that is 1.2 mbar, more than the leak limit. Parts from a washer or a welding cell must stabilise; the fixture must be at a steady temperature; and a reference volume (differential decay) cancels most of it.
  • Virtual leaks. Trapped volumes behind seals and in threads outgas into the test volume and look like leaks. Design the fixture with vented fasteners and smooth bores.
  • Adsorbed water. A humid part outgasses water for minutes at low pressure. Deep tests (below 10 mbar) on fresh castings need a dry nitrogen purge or a longer stabilisation, or the limit must be set with the outgassing in mind.
  • Pump and valve timing. A pump that reaches test pressure with a second to spare one cycle and not the next produces a random stabilisation time. Oversize the pump 30 % against the arithmetic.
  • Seal wear on the fixture. The fixture seal leaks before the part does; a daily master-part check (a known good and a known calibrated-leak part) catches it.

4. Dry pumps on the test floor

Test stations stand on the assembly line, not in a plant room, so the pump is beside operators and near painted or clean components.

  • Dry rotary vane (DRV Series, 90 mbar ultimate) for every test above about 150 mbar absolute: brake boosters, radiator and coolant decay tests, vacuum filling of brake and clutch hydraulics. Oil-free, quiet, no mist on the line.
  • Oil-sealed two-stage (DV Series) for helium chamber roughing and any test below 10 mbar, in a cabinet with the exhaust ducted, and with a foreline trap if it also backs the leak detector.
  • Dry screw (DSI Series) where a large helium chamber (battery pack enclosures, whole fuel-system assemblies) needs 150–300 m³/h to 1 mbar and the plant wants no oil at all.

Whichever pump: a non-return valve at the pump (a stopped pump must not let the receiver repressurise), a vacuum receiver sized for the dump volume, and the pump controlled by a pressure switch with hysteresis so it is not starting 180 times an hour.

5. Worked stations

Station Volume Test pressure Evacuation Pump Receiver
Brake booster EOL 8 L 363 mbar abs 3 s DRV10 50 L
Radiator decay test 12 L 200 mbar abs 5 s DRV25 100 L
Fuel tank, helium chamber 60 L chamber 1 mbar 15 s DV90 —
AC evaporator, vacuum fill 3 L 2 mbar 20 s + hold DV16 —
Battery pack enclosure, helium 800 L chamber 5 mbar 40 s DSI300 —
Multi-station bench, 6 fixtures 6 × 10 L 300 mbar abs 4 s each, staggered DRV60 300 L

Recommended Dr.Vakuum machines

Dry rotary vane (DRV Series) — line-side decay and functional tests, oil-free: DRV10 (12 m³/h), DRV25 (30 m³/h), DRV40 (48 m³/h), DRV60.

Two-stage oil-sealed vane (DV Series) — helium chambers and tests below 10 mbar: DV16, DV60, DV90, DV175.

Dry screw (DSI Series) — large chambers, oil-free: DSI150, DSI300.

The physics of the last decade that governs the deep tests is set out in How to size a vacuum pump.

Station sizing. Send us the part and fixture volume, the test pressure, the evacuation budget and the takt, and we will size the pump and receiver and comment on the stabilisation time. Talk to an engineer.