Solar Pump Commissioning Test: Prove the Curve Before You Trust It

The short answer: a solar pump purchase is a contract with a curve, and the commissioning test is the only hour that proves the contract was kept. The catalogue says 6 m3/h at 60 m — but that is a corner point, your farm is a point in the middle, and between the two sits every honest tolerance in the pump industry. One bucket of known volume, one stopwatch, one weighted tape, and one afternoon will tell you whether your system delivers what you paid for: measure the static level, pump it down, time the buckets, compute your real total dynamic head, and judge the result against the ISO 9906 acceptance grades — ±5% flow for grade 1B, ±8% for 2B, ±9% for 3B. Solar adds one test no other pump gets: the midday-to-afternoon bracket, because your power supply moves across the sky.

Everything in this guide needs equipment that fits in a shopping bag. No flow meter, no data logger, no technician on retainer. What it demands instead is the discipline most installs skip: writing down the numbers before you judge them, and knowing the tolerance band you are allowed to judge inside. For well preparation, pump placement and wiring before this test, see the solar pump installation guide.

Why the catalogue curve is a promise, not a fact

Open any pump datasheet and you get two headline numbers: maximum flow and maximum head. The uncomfortable truth is that the pump never produces both at once. Maximum flow happens against zero lift; maximum head happens at zero flow. Your farm lives on the curve between them, at a point determined by your well depth, your tank height and your pipe friction — the total dynamic head you calculated when sizing (if you have not done that yet, start with how to size a solar water pump).

Three separate things can therefore make an installed system diverge from the brochure:

1. The curve itself is an estimate. Manufacturing tolerance means one pump of the same model does not sit exactly where the printed curve sits. The industry quantifies this with acceptance grades — more on those below.
2. Your TDH estimate was an estimate. Static level drifts with the season, drawdown deepens as the aquifer works, and friction losses get underestimated. The commissioning test replaces the estimate with a measurement.
3. Solar input varies. A grid pump gets factory test conditions. Yours gets whatever the sky is doing — which is why the solar-specific tests later in this guide exist.

The yield-test guide covers the other half of this handshake — how much water the well itself can give. Here we verify the pump’s side of the contract.

The kit: what a commissioning test actually needs

Cost: near zero. The bucket-and-stopwatch method is accurate to roughly ±10-15% — comfortably inside every tolerance band you will judge against. That is the whole reason this test is honest: the instrument error is smaller than the decisions it informs.

The test, step by step: level, buckets, drawdown, pressure

Step 1 — Rest the well. No pumping for 12-24 hours before the test. A static level measured an hour after the pump last ran is not static; it is mid-recovery, and every number downstream inherits the error.

Step 2 — Record the static water level. Lower the weighted tape until it touches water, note the depth from the wellhead. Write down the date and time. This number is your baseline for every future test.

Step 3 — Start the pump and hold the discharge steady. Open the discharge to its working configuration — the tank fill, the same pipe the farm will use every day. Testing against a wide-open short hose measures a system you do not own.

Step 4 — Time buckets every 10-15 minutes for at least an hour. Each reading: three timed fills, averaged, then converted:

• Flow (L/min) = 60 × bucket litres ÷ fill seconds
• Flow (m3/h) = L/min × 0.06
• Flow (GPM) = US gallons × 60 ÷ fill seconds

A rising-then-flat series is a healthy well reaching equilibrium. A steadily falling series is the aquifer losing the race — see the diagnosis table below.

Step 5 — Measure the pumping level at the end of the test. While the pump still runs, lower the tape again. The difference from static is your drawdown, and static minus… no: pumping level is what feeds the TDH arithmetic. Record it at the same time as your final bucket, because the pair only means something together.

Step 6 — Compute the measured TDH. Add up what the pump actually fought:

• Lift: pumping level (from wellhead) + height from wellhead to the discharge point
• Pressure head: gauge pressure at discharge × 10.2 m per bar (or × 2.31 ft per psi) — zero for an open tank
• Friction: measured, or estimated at 10-15% of the vertical lift for a reasonable pipe run (the pipe sizing guide has the full tables)

Safety note that outranks every number here: never pump the well dry to see what it can do. Keep several metres of water above the pump intake throughout the test. Pulling the intake into air causes cavitation and can damage a submersible motor in minutes.

ISO 9906: how much tolerance is honest

Pump performance acceptance is standardized in ISO 9906:2012 (rotodynamic pumps — hydraulic performance acceptance tests, grades 1, 2 and 3; China’s GB/T 3216-2016 adopts it identically). The grades you will meet in contracts and manufacturer manuals are the bidirectional “B” grades, and the tolerance table looks like this:

Three reading rules that matter in the field:

• Efficiency is one-sided in practice. Measured efficiency may fall below the guarantee by the tolerance, never above it matters — a pump that beats its promised efficiency is a gift, not a violation.
• Small pumps get relief. For shaft power up to about 10 kW, the standard relaxes the bands to roughly ±10% on flow and ±8% on head. Every solar pump in this catalogue qualifies; judge your 750 W rung by the relaxed bands, not the headline ones.
• The grade must be agreed before purchase. The same measured pump can pass 3B and fail 1B on the same afternoon, and both verdicts are honest. For household and farm duty, 2B is the normal commercial expectation; grade 1B is for contract water where a third party pays per cubic metre.

This is also the answer to “the dealer promised 6 m3/h and I got 5.6” — at grade 2B, 5.6 m3/h against a 6 m3/h guarantee is a −6.7% shortfall: inside tolerance, the pump owes you nothing.

Reading a wrong number: the diagnosis table

The test’s real value is what a failure points at. Each signature has a shortlist of causes:

One more reading rule: the bucket test verifies a system, not a pump. A 30% shortfall can be 10% array, 10% cable and 10% expectation error stacked. The test sheet tells you which layer to open first.

The solar-specific tests: the midday bracket and the worst month

A grid pump’s commissioning verdict is one number. Solar needs two extra checks, because your power plant moves and dims.

The midday-to-afternoon bracket. Run identical bucket timings twice on the same clear day: once in the 10:00-11:00 window, once at 13:00-14:00 — symmetric about solar noon, so sun angle is comparable and only array health differs. Then compare:

• Afternoon flow ≥ 80% of morning — array is adequately sized; wiring and soiling are not stealing power.
• 60-79% — marginal. Expect weak shoulders and bad weeks; clean the panels, check cable sizing, and consider adding a module up to the controller’s limit.
• Below 60% — undersized array or a component losing power as it heats. The catalogue’s 1.3× array rule is the floor, not the target.

The worst-month re-test. A system that passes in March can fail its farm in December. The honest verdict is the one delivered in your lowest-sun month, at your season’s peak demand. If you live where the dry season and the weak-sun season coincide, this is not optional — schedule the annual re-test there deliberately.

The noon power cross-check. Read the controller’s display input power at solar noon. It should be in the neighbourhood of the pump nameplate — an array delivering 1.3× the nameplate should show the controller comfortably fed. A noon reading far below nameplate with a clear sky is the fastest single evidence of array or wiring trouble.

Worked example: a 1B-versus-2B verdict on a real farm

The scenario, assembled the way a real commissioning day unfolds: a household plus a half-hectare vegetable block needs 5 m3/day. Static water level measured at 40 m after the well rested overnight. The design stage had already chosen the 3DSC4.8-95-48-750 — 750 W, stainless wetted parts, 4.8 m3/h at 95 m maximum — against a calculated TDH around 74 m (pumping level 58 m, tank 12 m above the wellhead, friction 4 m). On the catalogue curve that duty point sits at roughly 1.9 m3/h guaranteed.

1. Pump and wait. The pump runs; bucket timings settle after 40 minutes. The 60-minute pumping level reads 58 m — drawdown 18 m, stable over the last three readings.
2. The buckets. Three timed fills of the 19 L bucket: 38 s, 38 s, 39 s. Average 38.3 s → 18.9 L × 60 ÷ 38.3 = 29.6 L/min = 1.78 m3/h.
3. The verdict arithmetic. Deviation from guarantee: (1.78 − 1.9) ÷ 1.9 = −6.3%. Against grade 1B (±5%) this pump fails; against grade 2B (±8%) it passes. Small-pump relaxation (±10%) passes it with room to spare. The same bucket, the same hour, three honest verdicts — which is why the grade belongs in the purchase contract, not the argument afterwards.
4. Does the farm care? Demand was 1.0 m3/h (5 m3 ÷ 5 sun hours); measured 1.78 m3/h clears it with a 78% margin. The system serves the farm; the tolerance argument was about the contract, not the crops.
5. The bracket. Morning fill 38 s, afternoon fill 41 s → ratio 89%. Array healthy, no action.
6. The sheet. Everything above lands on one dated page — the baseline the whole system will be judged against next year.

The verification ladder: models that show their work

Every model below is verified against the live catalogue. Use the ladder to sanity-check your duty point before the test — if your measured TDH sits at more than about 80% of a pump’s maximum head, you are testing near the shut-off corner where flow collapses, and you should be one rung higher:

Reading the ladder: the 3DSS screw rung is the exception to the shut-off warning — a progressive-cavity rotor keeps nearly all its flow right up to its maximum head, which is also why screw pumps verify almost flat: measure one at 90 m and it still gives close to its 1.7 m3/h. Everything centrifugal lives by the 80% rule.

Your one-page test sheet

Copy this into the pump-house notebook. Filled honestly once a year, it is the entire maintenance history the system will ever need:

The commissioning hour is the cheapest insurance in the whole project — an afternoon with a bucket that either buys you the confidence the brochure promised, or hands you the evidence while the installer is still on site. Pumps do not fail loudly in week one; they fail politely, at 5% a year, in numbers only a test sheet remembers.