Suppose lowering a pump's inlet pressure on a test rig changes its sound before the head gauge shows a clear change. Does the absence of the agreed percentage drop establish that there is no cavitation? No. First bubbles, audible noise and a repeatable loss of head are different observations. A head-drop criterion is useful for comparing performance degradation; it does not locate the first vapour bubbles.
Identify what the curve actually reports
Available net positive suction head represents the inlet head above the liquid's vapour-pressure head. Local pressure can fall further as liquid accelerates at the impeller inlet. The first bubbles may have little effect on overall head. Noise or vibration can change as bubbles collapse where pressure recovers. Air entrainment, mechanical faults and pipe resonance can also change the sound, so listening does not establish cavitation.
A common test result is the suction head at which measured head falls three percent at a fixed flow, often called NPSH3. Confirm the definition, speed, flow, liquid temperature and measuring locations used for the report. Different pump arrangements or published requirements can use different conventions. If the manufacturer specifies a required value for selection, ask whether it already includes an operating margin. Running with available suction head just equal to a three-percent drop point does not establish acceptable service or freedom from cavitation.
Record distinct events in the same test
Suppose reference head with ample inlet supply is 10.0 metres at a specified flow and speed. After a change at the inlet, it reads 9.98 metres and sound changes. At a further reduction in available suction head, it reads 9.70 metres, meeting the illustrative three-percent criterion. The first small difference may be within instrument and operating scatter, and the sound may not come from bubbles. The second observation records a defined drop in head. Keep them separate. Where a rig supports suitable observation, record actual bubble evidence without substituting noise for it.
Have qualified staff follow an approved rig procedure. Keep temperature, flow and speed comparable and check for air entry, pressure datum errors and changes in supply conditions. Record inlet conditions, stabilized head and any sound or vibration at each point. For selection, determine available suction head at the least favourable relevant source level, temperature and inlet losses. Apply the manufacturer's and designer's margin guidance to that condition. Do not obstruct the inlet of a field pump or hold it in damaging operation to obtain a number. The example is not a universal acceptance limit.
A report should also say how reference head was established and how stable the reading was. Comparing a steady reading at one inlet condition with a transient reading after a valve move can invent a head drop. Repeat suspicious points under the same observation window. That produces a curve a buyer can interpret and a record that separates a performance criterion from the question of how the installation will sound or wear.
At a site, distinguish increasing submergence from adding discharge resistance. The former may improve available inlet head; the latter changes operating flow. A quieter result after throttling does not validate the original suction-head calculation. In a professional review, establish whether pressure is absolute or gauge and identify temperature and pressure-tap elevation. Mixing those quantities can create a margin on paper that is absent in the installation. Investigate air entry independently, including a surface vortex or a return jet on the test loop. Controlled clean-water observations should not be transferred without review to a gas-laden or different liquid. Keep the actual site condition separate from the point printed on a laboratory curve.
Frequently asked questions
If head is unchanged, must the new sound be cavitation?
No. Look for entrained air, speed changes and mechanical vibration, and see whether the symptom repeats under controlled inlet conditions.
Why log water temperature?
Vapour pressure changes with temperature. Cold and warm runs may have different inlet margins and may not belong on a single comparison curve.
Do equal three-percent points mean equal sound?
No. Compare noise, mounting and suction margin at the actual duties. Equal values at one test criterion do not predict those outcomes.
