Solar Pump Storage Tank Sizing: The Tank Is the Battery

The short answer: the storage tank is the battery of a solar pumping system — it is where the five-hour pumping day gets stretched across a twenty-four-hour demand day. But almost every tank in the field is sized by a folklore multiplier (“one and a half times daily use”) instead of by the three volumes that actually decide performance: the water you draw while the array sleeps, the deadband between the float-switch levels that quietly sets how many times the pump starts every day for the next decade, and the turnover discipline that separates a buffer from a swamp. Size those three and the multiplier takes care of itself — and the tank, not the pump, becomes the component your system can least afford to guess.

A solar pump is a daylight-only machine. It lifts water when the array feeds it, and it stops when the sky dims — but the household, the herd and the crop drink on their own schedule, not the sun’s. Everything that separates a solar system from an annoying one lives in the volume between those two schedules. Undersize the tank and the evening routine drains it dry while the pump sits idle waiting for sunrise. Oversize it and you have bought stagnant water, mosquito habitat and a tank that never turns over. The arithmetic is not hard. It is just almost never done.

This guide works through the whole chain: what the tank actually does hour by hour, the night-shift demand that sets the first layer of volume, the float-switch deadband that nobody budgets for, the three-layer sizing method with a worked household example and a worked livestock example, the irrigation case where the tank plays a different role entirely, the cloudy-day autonomy ladder by use case, and the model ladder that a tank-first design points to. Every model cited is verified against our live catalogue.

What the tank actually does in a direct-drive system

In a grid or generator system, the pump stores nothing — electricity is always available, so the pump simply runs when someone opens a tap. In a direct-drive solar system (no batteries — the default for good reason), the energy supply is a pulse: roughly five peak sun hours a day, rising from nothing at dawn to a midday peak and back to nothing by dusk.

The tank converts that pulse into a flat supply. Water in a tank is energy stored as elevation and volume, and it is the cheapest energy storage in the entire project — an order of magnitude cheaper per useful unit than batteries, with none of the maintenance. Our direct-drive guide covers the system-level logic; this guide is about the number that makes it work: litres.

Three jobs, in the order the day demands them:

1. Serve the night. From sunset to the next morning’s sun, the tank is the only source. Its volume above the outlet must cover every litre drawn in that window.
2. Absorb the clouds. A passing cloud throttles the pump for minutes; a monsoon week can throttle it for days. Buffer depth decides whether a grey Tuesday is a non-event or an empty tap.
3. Discipline the pump. The float switch turns the pump on and off, and the vertical distance between its levels is the working volume. Too small a working volume and the pump starts and stops all day — every start a small tax on the controller and connectors, compounded over years.

Each job maps to a volume. Add them honestly and the tank sizes itself.

The night shift: what you draw while the array sleeps

The first question is not “how big should the tank be” but “how much of the day’s water is consumed when the pump cannot run”. The honest answer depends on who — or what — is drinking.

Households cluster their use. A typical pattern puts a block of use in the morning and a larger block in the evening, with a trickle overnight. As a planning assumption — not a measurement — sizing for 40-50 percent of daily demand falling between sunset and the next pumping window is a defensible, slightly conservative starting point. A family using 1 m3 per day therefore needs roughly 400-500 litres simply to survive the night without the tap going quiet.

Livestock are different in a way that matters. Cattle drink throughout the day, with peaks around feeding and a long, steady overnight tail — a trough system with a float valve refills continuously, so the tank’s night job is smaller as a share of demand, but the absolute reliability bar is higher because missing water hits the animals directly, not a schedule.

Irrigation is different again, and Section 7 handles it separately, because for crops the tank often plays the role of a flow-shifter rather than a multi-day reservoir.

The night-shift number is the first layer of the stack. It is also the one layer that pure multipliers get wrong most often, because it scales with behaviour, not with total demand: a household that cooks and bathes in the evening has a night share near 50 percent; a farm cottage that draws most water at midday for equipment washing may sit near 25.

The deadband is a volume, not a switch

Here is the layer almost nobody budgets. A float switch closes at a low level and opens at a high level, and the volume between those two levels is the working volume — the only part of the tank the pump actually cycles. Everything below float-on is reserve; everything above float-off is untouched air.

The count that follows is simple and unforgiving:

Pump starts per day ≈ daily pumped volume ÷ working volume

Run the arithmetic for a system moving 15 m3 a day and watch the working volume do the work:

Every start is an inrush event. Solar pump controllers soft-start the motor and the catalogue’s DC models ramp rather than slam, but contactors, relays, connectors and the well itself all keep score across tens of thousands of cycles. A pump that starts 15 times a day lives a different life than one that starts 3 times, and the difference costs nothing but float travel: a taller tank, or a float arm with a longer sweep, or simply choosing a tank whose usable height is measured in metres rather than centimetres.

Two practical notes sit next to the table. First, the deadband volume also defines the minimum refill time — a 6 m3/h pump refills a 1,000 L working band in ten minutes; that is also the shortest rhythm your plumbing turbulence and float chatter must tolerate. Second, the reserve below float-on is not waste: it is your cloudy-day autonomy, which is the next layer.

The three-layer arithmetic: a worked household example

The method stacks three volumes. The scenario: an off-grid homestead, family of six, basic-plus use of about 100 litres per person per day — 0.6 m3/day total. The borehole: static water level 20 m, pumping level 28 m, tank on a 6 m stand, friction about 2 m — total dynamic head roughly 36 m.

1. Night layer. 40-50 percent of 600 L drawn between sunset and the next pumping window → plan 300 L.
2. Cloud layer. One day of autonomy for a household → 600 L.
3. Deadband layer. Working volume above float-on. For a small pump this is less about start counts (a 300 W DC pump soft-ramps and tolerates cycling far better than a big AC unit) and more about refill rhythm → 150-200 L.

Stack: 300 + 600 + 175 ≈ 1,075 L working. Nominal tank: 1,100-1,200 litres — one standard 1,100 L poly tank, opaque, on the stand.

The pump: 600 L over five peak sun hours needs an average of 0.12 m3/h at 36 m — a light duty point. The 2DPC1.7-45-24-300 (300 W, 24 V, 1.7 m3/h maximum flow, 45 m maximum head) sits on this rung comfortably: at 36 m the curve delivers on the order of a cubic metre per hour, which refills the working band in about an hour and the whole day’s demand in the first half of the sun window. The array follows the catalogue rule — solar panel at least 1.3 times pump power — so one 400-450 W module covers a 300 W pump with margin.

Note what the tank did in this example: it let a small, cheap, 300 W pump serve a 24-hour household. The alternative — batteries sized for the evening block — would have multiplied the budget for the same glass of water.

Livestock: a different asymmetry

The second worked example changes the demand shape. The scenario: 60 head of cattle at roughly 45 litres per animal per day — 2.7 m3/day, drawn at troughs through the day with a modest overnight refill. The borehole: pumping level 35 m, tank stand 5 m, friction 2 m — total dynamic head roughly 42 m.

1. Night layer. Overnight trough refill is a smaller share here — plan 30 percent → about 800 L.
2. Cloud layer. Livestock autonomy is conventionally one day: the herd drinks tomorrow whether the sun shines or not → 2,700 L.
3. Deadband layer. A float-valve trough line cycles continuously, so the tank working band matters for the well pump’s rhythm → 200-300 L.

Stack: 800 + 2,700 + 250 ≈ 3,750 L working → two linked 2,000 L tanks, or one 4,000 L.

The pump: 2.7 m3 over five sun hours is 0.54 m3/h at 42 m. The 3DSC4.8-95-48-750 (750 W, 48 V, stainless-steel wetted parts, 4.8 m3/h maximum flow, 95 m maximum head) clears this duty with a wide margin — at 42 m the curve delivers several times the requirement — and the stainless build is the right call for a well that will run hard every day for years. Array: 1.3 × 750 = about 975 W, two 550 W modules in series within the controller’s VOC limit.

Read the two examples side by side and the lesson is the shape, not the litres: the household’s tank is sized from behaviour (evening cluster), the herd’s from biology (daily drinking volume). The multiplier folklore cannot tell those apart. The three layers can.

Irrigation: when the tank shifts flow instead of storing days

For crops, the honest framing flips. Soil is itself a storage medium — a loam holds 100-200 mm of available water per metre of root zone — so irrigation usually does not need multi-day autonomy in a tank; it needs the pump’s midday output moved to the morning and evening windows when evaporation is low and scheduling is humane.

The scenario: a 0.5 ha vegetable block, peak-month demand 5 mm/day → 25 m3/day (1 mm over 1 ha is 10 m3). Water source: a pond beside the field, lift about 8 m to the tank pad. Irrigation runs two windows — morning 06:00-08:00 and evening 18:00-20:00 — at a drip manifold demand of about 6.3 m3/h while flowing. The pump cannot run in either window; both windows draw from storage.

1. Window layer. The evening block draws 12.5 m3 with nothing refilling → working volume must hold 12,500 L.
2. Deadband layer. Add about 1,000 L so the big pump is not short-cycling between window drawdowns.
3. Autonomy layer. The soil buffer carries day-to-day variance; the tank adds a partial day, not two → covered by sizing the nominal tank up one standard step.

Working volume ≈ 13,500 L → nominal 15 m3 (two linked 10 m3 tanks, or a masonry reservoir).

The pump: the midday job is refilling 25 m3 in five sun hours — an average of 5 m3/h, but at a pond-side 8 m lift the right rung is a low-head, high-flow one. The 4/6DSC36-22-110-1500 (1,500 W, 110 V, 36 m3/h maximum flow, 22 m maximum head) delivers on the order of 30 m3/h at an 8 m lift — the morning window’s 12.5 m3 refills in well under an hour of full sun. Array: 1.3 × 1,500 = about 2,000 W.

This is the irrigation tank’s real job, stated plainly: the array and pump are sized for energy; the tank is sized for the window. A 15 m3 tank on a 2 kW pumping system delivers an evening irrigation block that no amount of extra panels could deliver at 19:00.

The cloudy-day ladder

Autonomy — how many days the tank carries demand with the pump contributing nothing — is where use cases genuinely differ. The honest ladder, by what happens when the water runs short:

Two cautions ride with the ladder. First, autonomy is cheapest bought in tank, not in panel: doubling cloudy-day reserve by doubling the array costs far more than a second tank. Second, the top of the ladder is where the next section’s warning begins.

When storage cannot cover the response time, follow solar pumping redundancy design to add the appropriate power or equipment backup.

Turnover: oversizing has its own failure mode

Water is perishable. A tank that cycles its full volume every one to two days stays fresh with no attention at all — each fill renews the residual chlorine, each draw replaces the oldest water. A tank that sits three or more days between full turnovers is a different chemistry set: residual chlorine decays in days, taste drifts, algae find sunlight through any translucent wall, and sediment settles into a layer the first flush of the tap stirs up.

So the sizing rule has a ceiling, not just a floor:

• Default to one to one-and-a-half days of autonomy. The village guide’s 1.5-day multiplier is a ceiling for good reason, not a minimum to exceed.
• If a bigger buffer is unavoidable, schedule the turnover. Draw irrigation from the bottom outlet, clean annually, keep the wall opaque.
• Watch the float switch’s minimum level. Reserve below float-on that never moves is the first place water goes stale — occasionally drain to that level deliberately and let the pump refill through it.

The failure mode of the oversized tank is quiet — no alarm, just water that tastes like the tank — which is exactly why the folklore multiplier overshoots it everywhere.

The model ladder for tank-first systems

Every model below is verified against the live catalogue. The rungs a tank-first design actually climbs:

Reading the ladder: at the small end, the tank is doing most of the working — the pump only needs to outrun the night layer, so the 300-500 W rungs carry households that folklore would have sized with a kilowatt of panels. In the middle, the stainless 3DSC/4DSC rungs buy daily-reliability duty for herds and villages. At the top, the A/D wide-voltage rungs exist precisely because reservoir refill at this scale is an energy-transfer problem measured in kilowatt-hours, where the AC input halves the array a pure-solar build would need.

The habit that keeps the arithmetic honest

A tank is the one component whose performance you can verify with a stick and a notebook. Once, at commissioning, and then every season: mark the float-on and float-off levels on the tank wall, measure the litres between them, and compare against your three-layer stack. Log the pump’s starts-per-day against the table above. Time one full refill at midday.

The numbers will drift — sediment raises the floor, valve wear moves the float levels, and the household’s evening cluster grows with the family. The tank that gets checked is the tank that stays a battery. The one that gets forgotten becomes a monument to the multiplier — full of water, and full of nothing else.