How Pressure Affects Leak Rate and Jeopardizes Product Passing Resulys
A unit can pass at 150 psi and still leak at 350 psi.
This isn’t about whether the sensor is sensitive enough to find a leak. It’s about testing at a pressure that doesn’t match how the unit actually runs. The same physical opening can behave completely differently depending on how hard the system is pushing refrigerant through it — and most HVAC and Refrigeration Systems test far below their real operating pressure.
The gap that matters
Say a unit is leak-tested during production at 150 psi, then installed and run in the field at 350 psi. Same unit, same leak path — but two very different pressures pushing refrigerant through it.
Leak produces a small amount of tracer gas. Falls inside the pass/fail threshold. Unit ships.
Same leak, more than double the pressure behind it. Refrigerant escapes at a meaningfully higher rate than the test ever measured.
Why the physics makes this worse, not better
How much gas escapes through a leak isn’t fixed — it depends on the pressure difference (ΔP) pushing gas through it, and how that dependence works depends on the size and shape of the leak path itself. There are two regimes:
Molecular (Knudsen) flow
In very small, hairline leak paths, gas molecules pass through mostly independently of one another. Leak rate rises proportionally with pressure — roughly a straight-line relationship.
Viscous (laminar) flow
In larger, more open leak paths, gas behaves more like a fluid stream. Leak rate rises with roughly the square of pressure — a small pressure increase produces a much larger jump in leak rate.
Viscous flow: Q ≈ Cv · ΔP² Q — leak rate. C — a constant set by the leak’s size and shape. ΔP — pressure difference across the leak.
Going from a 150 psi test to 350 psi field operation is more than a 2x increase in pressure. For a molecular-type leak, that’s roughly a 2x increase in leak rate. For a viscous-type leak — the kind more likely to cause a noticeable refrigerant loss — it’s roughly a 5x increase, because the relationship is squared.
| Leak type | ΔP ratio (350 ÷ 150 psi) | Leak rate at field pressure vs. test |
|---|---|---|
| Molecular (fine, hairline) | 2.3× | ~2.3× higher |
| Viscous (coarser, more open) | 2.3× | ~5.4× higher |
In plain terms: a leak that comfortably passed the 150 psi test with room to spare can still cross the acceptable threshold once the unit is actually running at 350 psi — especially if it’s the larger, viscous-flow type of leak that a customer would actually notice as lost refrigerant.
Most leak paths in HVAC and Refrigeration Systems equipment — flare and braze joint imperfections, service valve seats, gasket seams — are on the larger, more open end of the spectrum, not hairline pinholes. That means most of the leaks these units actually develop are the viscous-flow type, the kind where leak rate scales with the square of pressure. The pressure gap between test and field isn’t a minor factor here — it’s the dominant one.
Say the acceptance spec is 0.10 oz/yr of refrigerant loss, measured at a 150 psi test. A part right at that limit isn’t failing anything — it ships. But once installed and running at a higher field pressure, that same leak path passes more refrigerant, because ΔP has gone up and the leak is viscous-flow.
| Field pressure | ΔP ratio vs. 150 psi test | Viscous leak rate factor (ΔP²) | Actual field leak rate |
|---|---|---|---|
| 250 psi | 1.7× | ~2.8× | ~0.28 oz/yr |
| 350 psi | 2.3× | ~5.4× | ~0.54 oz/yr |
Either way, a part that looked well within spec at test pressure is leaking two to five times the specified rate once it’s actually running in the field — without the part, the leak, or the test procedure ever doing anything “wrong.” The spec was simply never verified at the pressure that matters.
Why this matters to the business
Refrigerant that leaks in the field doesn’t just mean a service call. It means reduced system performance, a warranty claim, and — depending on the refrigerant type and jurisdiction — a regulatory reporting obligation, since refrigerant leak rates are increasingly tracked under environmental rules. A leak test that only proves a unit is sound at test pressure isn’t proving what customers actually need: that it’s sound at the pressure it will really run at.
Questions worth asking your test team
What pressure do our units actually run at in the field, and how does that compare to our test pressure?
Was our pass/fail threshold set with that gap in mind, or just against the test pressure alone?
Do we know whether our typical leak paths behave more like fine (molecular) or coarse (viscous) leaks — since that changes how much the pressure gap matters?
For any field refrigerant-loss complaints, were those units tested at full operating pressure, or below it?