Why does the helium background stay high after pumping down?

Four causes, and they are distinguishable by how the signal behaves over time rather than by how high it is: helium dissolved in elastomer seals, helium trapped in the test object from a previous test, an internal leak in the instrument's own valve block, or a saturated detector that simply needs time and a clean pump-down.

Start with the behaviour, because it splits the four cleanly. A background that falls slowly and steadily over minutes to tens of minutes is desorption — helium coming back out of elastomers or out of the part you just tested. A background that sits flat and will not move is an internal leak or a contaminated analyser cell. That one distinction tells you whether to wait or to open the instrument.

Why elastomers do this: helium is a small, inert atom that dissolves into and diffuses through Viton and similar polymers. After a test with a generous helium spray, the O-rings in your test port hold a reservoir that bleeds back out for a long time. It is not a fault, and spraying more helium next to the seals during the test makes the next test worse.

The practical fix for that class: spray helium sparingly and from below the part where possible, ventilate the test area between tests, and allow a clean pump-down between runs rather than starting the next test as soon as the reading looks acceptable. On high-throughput lines, a metal-sealed test port removes the reservoir entirely.

If the background is flat, check the instrument itself. Calibrate against the internal or external standard leak — a detector that no longer reproduces its calibrated value has a real fault, not a background problem. Then look at the valve block: an internal leak between the helium-rich side and the analyser will hold the background up indefinitely regardless of what you do to the test part.

Two things that mask themselves as background: a leak in the test fixture rather than the test part (test the empty fixture as a control — a step every busy line skips and then loses hours to), and an aged ion source filament, which shifts sensitivity and makes the same physical leak read differently from last week. We supply filaments, electron multipliers, analyser cell parts and calibrated leaks for the common Pfeiffer/Adixen ASM and Leybold/INFICON instruments — send the model and the symptom.