Cavitation in Solar Pumps: The Failure You Can Subtract Away

The short answer: cavitation is the fifth way a pump dies — after sand scours it, water corrodes it, scale clogs it, and iron slime smothers it — and it is the only one you can eliminate with a two-line subtraction before you buy anything. When local pressure at the impeller eye falls below the water’s vapor pressure, the water boils cold, and the vapor bubbles that form there collapse against the vanes with micro-jets hitting metal at 100-500 m/s and impact pressures of 1,000 to 10,000 MPa — beyond the yield strength of every common pump metal. The defence is one inequality: available suction head must exceed required suction head plus 0.5 to 1 m of margin. Solar pumping has a structural advantage here: a submersible hangs in the water and its inlet is flooded, so the inequality passes almost automatically. Nearly all the danger concentrates in one place — the surface pump with a suction lift.

This guide gives you the physics in one page, the arithmetic in two lines, the vapor-pressure table that shows where layouts break, and the field diagnosis that catches cavitation by ear before the impeller shows it.

The fifth failure path, and the only one you can design away

This site has covered four failure paths, each attacking a different part of the system. Sand is physics on the wetted surfaces — a materials problem (screw vs impeller pumps). Brackish water drives galvanic corrosion — an electrochemistry problem (brackish water guide). Hardness lays mineral scale wherever pressure drops or heat concentrates (the hard water guide). Iron feeds slime biofilm that strangles drip emitters (the iron-rich water guide).

Cavitation is different in one decisive way: it needs no bad water at all. The cleanest, softest, coldest spring water will destroy an impeller in weeks if the suction side of the installation is wrong. It is a geometry-and-arithmetic problem, which is exactly why it is solvable on paper — the four paths above need materials and maintenance; this one needs a subtraction.

The cost of ignoring it is documented. Australian irrigation research on cotton farms found that pumps running with even slight cavitation suffer an immediate and permanent efficiency drop of 5 to 10 percent, and that prolonged light cavitation erodes the impeller enough to cut efficiency by 40 percent or more — more fuel or more sun for the same water, every day, plus eventual impeller replacement.

How bubbles demolish metal

Three steps, all of them inside the pump:

1. Form. Where velocity is highest — the eye of the impeller — pressure is lowest. If it falls to the vapor pressure of the water, the water flashes to vapor right there, in bubbles. This is boiling by decompression, not by heat.
2. Travel. The bubbles ride the flow into the impeller passages, where the pressure climbs steeply toward the discharge.
3. Collapse. Back above vapor pressure, each bubble implodes. A collapsing bubble next to a solid surface does not shrink symmetrically — it spikes a micro-jet of liquid straight at the metal. Measured jet velocities run 100 to 500 m/s; the instantaneous impact pressures reach 1,000 to 10,000 MPa for nanoseconds. Carbon steel yields at about 250 MPa, 316 stainless at about 290. Each collapse is a hammer blow that exceeds the metal’s strength many times over, and a cavitating pump delivers millions of them per hour.

The signature damage matches: spongy, honeycomb pits clustered just downstream of the eye, on the leading edges of the vanes — sometimes eating clean holes through a vane. That is also how you tell it apart from sand: sand polishes everything evenly; cavitation digs locally.

One clarification that matters for solar users: the bubbles themselves are harmless. The damage is done by collapse, which is why the pits sit slightly past the lowest-pressure point, and why a pump that only cavitates at the far end of its curve (high flow) eats its vanes while still delivering water right up until it doesn’t.

NPSH: the two-line subtraction

Everything above is controlled by one accounting identity. The head available at the pump inlet, measured above the water’s vapor pressure:

NPSHa = (atmospheric pressure − vapor pressure) + liquid level relative to pump − suction-line friction

against the head the pump demands, NPSHr, printed on its curve. The design rule:

NPSHa ≥ NPSHr + 0.5 to 1 m of margin

Two honest footnotes. First, the published NPSHr is conventionally the point where head has already dropped by 3 percent — the pump is already cavitating mildly there. That is what the margin is for. Second, both sides of the inequality move with the weather: atmospheric pressure falls with altitude (the high-altitude guide does that version), vapor pressure climbs with temperature, and friction grows with flow — which is why a layout that passes at dawn can fail at noon.

Worked at sea level: how many metres survive

Start from the atmosphere’s 10.33 m of theoretical lift at sea level and subtract the real world. Water at 20 C: vapor pressure 2.34 kPa ≈ 0.24 m. A reasonable suction line with strainer: about 0.8 to 1.0 m of friction at design flow. A small centrifugal with a requirement around 2.5 m. Margin 0.5 to 1.0 m.

10.33 − 0.24 − 1.0 − 2.5 − 1.0 ≈ 5.6 m of honest design lift — which is why practice quotes roughly 6 m, and why a nameplate claiming 8 m is a laboratory figure, not an installation instruction (river and canal pumping works through the same budget from the water side).

The number most people forget is the one that moves on its own: vapor pressure.

Hot water: the layout killer nobody budgets for

Run the same installation — 3 m of static lift, 0.8 m of friction, a pump needing 2.5 m, sea level — with the water 15 degrees warmer and the arithmetic turns hostile:

• 20 C water: NPSHa = 10.33 − 0.24 − 3 − 0.8 = 6.29 m → margin over a 2.5 m requirement: +3.79 m. Fine.
• 60 C water: NPSHa = 10.33 − 2.03 − 3 − 0.8 = 4.50 m → margin: +2.00 m. Still fine.
• The same pipework with 6 m of lift instead of 3: at 60 C, NPSHa = 1.50 m → −1.00 m. Cavitating, permanently, at every stroke of the sun.

Vapor pressure roughly doubles every 10 to 12 degrees C. Any layout that moves warm water — shallow aquifers that run hot in the dry season, storage tanks standing in the sun, heat-recovery loops — must be checked at the warmest realistic temperature, not at 20 C. A layout that passes cold can destroy its impeller warm, on an otherwise identical installation.

The three places solar pumps actually meet cavitation

1. The surface pump with a suction lift — the main event. Every metre between the water surface and the pump centreline is subtracted from your budget before friction even starts. River banks are the classic trap: the pump that sits conveniently on the bank in February is 1.5 m higher above the water in April. If your computed lift at the lowest seasonal level exceeds about 5.5 m, stop designing and move the pump down or the intake into a pit. No panel, controller, or bigger pump will rescue it — this is the same verdict the river-pumping guide reaches from the water side, for the same physics.

2. Vortexing — air sucked in from above. As the water surface over an intake drops low enough, it stops drawing down as a flat surface and starts forming a rotating funnel that carries air down into the pipe. The pump is now compressing a mixture, losing head and priming in strokes. The fix is submergence over the inlet — a rule of thumb keeps at least a few inlet diameters of water over the strainer — or an anti-vortex plate.

3. The well level chasing the submersible. A submersible’s inlet is flooded, but only while the pumping water level stays above it. Drawdown that pulls the level to the intake produces vortexing and then air entrainment — and running even briefly in air damages a submersible motor in minutes. This is why commissioning practice keeps several metres of water above the intake at all times, and why the controller’s dry-run protection is not optional equipment (the commissioning test).

Why the submersible is almost immune

Put the pump in the water and the arithmetic flips in your favour. Atmospheric pressure still acts — on the water surface — and the pump’s inlet sits below that surface by the submergence depth. Available head at the inlet: 10.33 m plus submergence minus vapor pressure. At 2 m of submergence and 30 C water, that is about 12.1 m on offer against any reasonable requirement — a margin no honest surface installation can match.

This is worth stating plainly because it reverses a common instinct: the deep-well pump is not the risky one. It is the safest pump in the catalogue on this particular failure mode. The surface pump on the river bank is where cavitation lives. When a surface application is unavoidable and the suction budget is tight, the modern answer is often to stop fighting the suction side and put a small submersible in the water instead (submersible vs surface).

Diagnosis: your ear is the instrument

Cavitation announces itself before the impeller shows damage:

• The sound. Gravel or marbles rattling through the pump — harsh crackling that rises and falls with flow. The single most reliable field sign, reported identically from irrigation stations to drilling rigs.
• Fluctuating discharge pressure. Gauges swinging 10 to 30 percent at steady speed, because vapor in the passages makes head collapse and recover stroke by stroke.
• Flow below the curve. The pump delivers less than the curve promises at the same sun, especially at the high-flow end of the day.
• Vibration. Rough and random rather than smooth.

Solar adds one confirmation trick: cavitation is flow-driven, and solar flow peaks at noon. If the gravel sound is worst at midday and fades toward evening, you are watching the required head climb with flow while the available head fell with the morning’s level drop — the signature is unmistakable once heard. Do not throttle the discharge valve closed to quiet a cavitating pump as a fix; it masks the symptom while the suction arithmetic stays broken. Fix the suction side.

The defence list, in order of what actually works

1. Lower the pump or raise the water. Every 1 m of lift you remove adds a full metre of margin. The cheapest pump house is a pit.
2. Short, fat, straight suction pipe. Large diameter, minimal length, few fittings; strainer clean and generously sized. Suction-side friction is subtracted one-for-one.
3. Check the worst case, not the average. Lowest seasonal water level, warmest realistic water, highest honest flow. A margin that survives that triple is safe.
4. Keep the margin. Available must exceed required by 0.5 to 1 m at minimum; critical or unattended systems deserve the full metre.
5. If the budget still fails, change the architecture. Flood the suction — dig the pit, or put a submersible in the water. No pump choice rescues a failed suction layout, but a flooded inlet removes the problem class entirely.
6. Buy toughness last, not first. Stainless impellers last 5 to 10 times longer under cavitation erosion than bronze — the reason the stainless rungs of this catalogue (3DSC / 4DSC and up) are the honest answer for abrasive or cavitating water. But stainless changes how long the pump survives cavitation, not whether it cavitates; the subtraction still rules.

Worked example: the river bank, twice

A vegetable plot draws 20 m³ per day from a canal, lift to the pump about 3 m at the April low. A 750 W surface pump — the DCPM21-14-72-750, 21 m³/h and 14 m of head at 750 W — fits the duty from the discharge side. From the suction side:

• 20 C water, 3.0 m lift, 0.9 m friction, requirement 2.5 m: NPSHa = 10.33 − 0.24 − 3.0 − 0.9 = 6.19 m → margin +3.69 m. Safe.

Now the same pump in the worse version of the same job: someone has it on the bank 5.5 m above the April low.

• 20 C: NPSHa = 10.33 − 0.24 − 5.5 − 0.9 = 3.69 m → margin +1.19 m. Passing, thin.
• 40 C water (shallow, sun-warmed): NPSHa = 10.33 − 0.75 − 5.5 − 0.9 = 3.18 m → margin +0.68 m. Inside the minimum band.
• Add one clogged strainer worth 0.8 m more friction: margin −0.12 m. Cavitating.

Same pump, same water source, same everything except numbers a pencil can check — the installation lives or dies on the subtraction. The remedy here is not a bigger pump; it is moving the pump 1.5 m down the bank, which buys back more margin than any upgrade on the discharge side ever could.

The model ladder, verified

Ladder from the catalogue, every entry checked against the live product data — submersibles first (structurally immune), then the surface pumps that demand the subtraction:

Read it as a design rule, not a shopping list: the first ten rows keep the inlet flooded and let the subtraction pass itself; the last three are honest pumps that will serve you for years — provided you run their suction arithmetic before you pour the slab.