Solar Pump Inverters and VFDs: What the Controller Actually Does, and How to Choose One

In every solar pumping system there is one component doing almost all of the thinking, and it is the one buyers spend the least time on. The panels are commodities. The pump is chosen from a curve. The controller, the box that sits between them, is what decides whether the system delivers water for fifteen years or burns out a motor in the first dry season.

The reason the box exists at all is a mismatch of temperament. A solar array produces power that varies minute by minute with irradiance, cloud and temperature. An electric motor wants a particular voltage at a particular frequency and punishes you for supplying anything else. Something has to sit between those two and translate, continuously, all day, every day.

That something is the solar pump inverter, also called a solar pump VFD (variable frequency drive) or, generically, a solar pump controller. This article explains what it actually does, where DC and AC systems each make sense, which protections matter in the field, why the physics of pump speed is so forgiving of cloudy weather, and which commissioning parameters get botched most often. It is deliberately brand-neutral. The principles here apply to every manufacturer's box, and a controller that cannot do these things is the wrong controller regardless of whose name is on it.

Why You Cannot Wire a Pump Straight to the Panels

Start with what happens if you try. Take a standard three-phase borehole pump, rated 400 V at 50 Hz, and connect it through a plain inverter that outputs fixed voltage and fixed frequency, sized off the motor nameplate.

At 09:00 with good sun, it might run. At 11:30 a cloud crosses the site and array output drops by more than half within a minute. The motor is still being asked to run at full speed and full load. The DC bus voltage collapses, the inverter trips or, worse, the motor stalls while still energised, torque drops with the square of the voltage, slip rises, current rises with it, and the winding starts cooking. Repeat that cycle twenty times a day, every day, and you have designed a motor-disposal machine.

There is a second, subtler problem. A solar array is not a stiff supply. At any given irradiance and cell temperature it has exactly one operating point, one combination of voltage and current, at which it delivers maximum power. Load it too hard and the voltage collapses; load it too lightly and you leave power on the table. That point moves all day. A fixed load never sits on it except by accident.

So the job description writes itself. The controller must track the array's maximum power point as it moves, and it must adjust the pump's speed so that the pump absorbs exactly the power the array can give at that moment, no more and no less. Everything a solar pump inverter does follows from those two requirements. The system-level view of how the controller fits alongside the array, pump, borehole and storage is covered in our solar water pumping systems guide.

What a Solar Pump Inverter Actually Does

Inside the box, four functions matter.

MPPT on the DC side. The controller continuously perturbs its operating point and measures whether power rose or fell, holding the array at its maximum power point as irradiance and temperature change. This is the same tracking function a grid-tie inverter performs, but the destination for the power is different: instead of exporting to a grid that can absorb anything, the controller must dispose of the power into a pump, which brings us to the second function. If you are weighing up grid-tie or hybrid units for a building rather than a pump, that is a different selection problem, and we cover it separately in how to choose a solar inverter.

Variable frequency output. The controller synthesises a three-phase AC waveform whose frequency, and voltage in proportion, it can vary continuously. This is the VFD function, the same power-conversion architecture that IEC 61800-2 specifies for adjustable speed AC drive systems generally. When the sun is strong the pump runs at 50 Hz. When a cloud passes, the drive winds the frequency down smoothly, the pump slows, its power draw falls steeply (more on exactly how steeply below), and the system rides through on reduced flow instead of stalling. Morning start-up is the same movie in reverse: the pump starts turning slowly as soon as the array can support it and accelerates as the sun climbs.

Soft start. A direct-on-line motor start draws several times rated current for a second or more. A solar array simply cannot supply that; it has no thermal mass and no short-term overload capacity to speak of. The drive sidesteps the problem entirely by ramping frequency from zero, so the motor accelerates gently and inrush never happens. As a side benefit, the water column in the riser pipe is accelerated gradually too, which spares the non-return valve and the pipework the daily water-hammer that direct starting inflicts.

Dry-run protection. A submersible pump uses the water it moves to cool its motor and lubricate its bearings. Run it dry and it can destroy itself in minutes. Boreholes in the regions where solar pumping makes most sense are exactly the boreholes prone to seasonal drawdown, so dry run protection for a solar pump is not an optional extra, it is a core function. Good controllers detect a dry well two ways: from a level probe or float switch in the well, and electronically, by noticing the load signature of a pump that has lost its water (power drawn at a given frequency drops sharply when the impellers are churning air). The electronic method costs nothing to install but should be treated as the backup, not the primary; a probe in the well is cheap insurance for a motor that is decidedly not cheap to replace, still less to pull from 120 metres down.

One more parameter belongs in this list because it protects the pump from the drive's own cleverness: the minimum frequency limit. A centrifugal pump's delivered head falls with the square of speed, so below some speed it generates too little pressure to lift water to the surface at all, and the pump churns uselessly while the motor's cooling flow past the can drops away. Submersible motor makers also specify a minimum flow velocity past the motor for cooling. For these reasons the drive is programmed never to dwell below a minimum frequency, commonly in the region of 25 to 35 Hz for borehole pumps, with the exact figure set from the pump curve and the motor cooling requirement, not guessed. Below that threshold the correct behaviour is to stop, wait, and retry when more power is available.

The performance and qualification framework for all of this at system level, array, converter and motor-pump set together, is defined in IEC 62253, which is the standard worth citing in a procurement specification if you want bidders' claims to be testable.

DC or AC: Where Each System Wins

Solar pumping systems come in two families, and the split is mostly a question of size.

DC systems connect the array, through a simpler controller, to a pump driven by a brushless DC or similar electronically commutated motor. There is one power conversion fewer, so efficiency at small scale is excellent, and the pump and controller are typically sold as a matched pair. This family dominates the small end: livestock watering, drip irrigation on smallholdings, household and community supply, roughly the range up to a few kilowatts. Within its band it is hard to beat on cost and simplicity. Its weaknesses are the flip side of the matched pair: you are locked to the manufacturer's combinations, replacement motors are proprietary rather than commodity items, and above a few kilowatts the options thin out quickly.

AC systems use a standard three-phase asynchronous motor driven by a solar pump VFD. The motor is a commodity, repairable in any competent workshop, and available in every size from fractional kilowatts to hundreds. The drive can in principle run any suitably rated motor pump set, which keeps procurement competitive and spares available locally. The extra conversion stage costs a few points of efficiency at the small end, which is why AC rarely makes sense below a kilowatt or two, but from roughly 4 kW upward the AC architecture wins on flexibility, serviceability and sheer availability of hardware, and above about 10 kW there is no real contest. The World Bank's Solar Pumping: The Basics reaches the same practical conclusion: small systems favour integrated DC packages, larger systems favour AC motors on variable frequency drives.

The boundary band, roughly 2 to 4 kW, is genuinely contested and comes down to local support. The right question is not which technology is better but which failure you can fix on a Tuesday: a proprietary DC pump end that must come from the importer, or a standard AC motor the rewinder in the nearest town has seen a hundred times.

Hybrid input controllers deserve their own mention because they answer the question every client asks: what happens when it rains for a week and the tank is empty? A hybrid solar pump inverter accepts DC from the array and AC from a grid connection or a generator, and switches or blends between them, solar first, backup when solar cannot meet demand. For clinics, hotels and irrigation with hard daily quotas this is usually the right architecture, because it sizes the solar array for the typical day rather than the worst day, which is a much cheaper array. Two cautions from the field. First, confirm the changeover logic is automatic and tested, not a manual switch that requires someone to be present at 05:00. Second, if the backup is a generator, check the controller's tolerance of generator-grade voltage and frequency wobble, and size the generator for the drive's input characteristics rather than the motor nameplate alone.

Planning a pumping system and not sure which architecture fits your site? Our renewable energy team designs and delivers solar pumping across exactly this range and will size the array, drive and pump as one system rather than three purchases.

The three solar pumping architectures and where each wins. The hybrid controller answers the question every client asks: what happens when it rains for a week and the tank is empty. Source: MIMAH engineering practice.

The Protection Set That Matters in the Field

A controller specification sheet lists a dozen protections. In hot-climate field service, five of them decide the system's lifespan.

Dry-run and low-level protection, covered above, comes first because the failure it prevents is the most expensive routine failure in borehole pumping. Fit the probe. Set the electronic detection as well. Configure a sensible auto-restart delay, long enough for the well to recover, so the system does not hammer a weak borehole with retry cycles.

Over- and under-voltage protection on the DC input. Array open-circuit voltage rises in cold conditions and falls as cells heat up, and a string specified carelessly can exceed the drive's maximum DC input on a cool clear morning. The drive must trip cleanly on over-voltage and ride through or stop gracefully on under-voltage, and the string design must keep the array's voltage window inside the drive's MPPT range across the whole temperature span the site will actually see.

Lightning and surge protection. A solar pumping site is a well-earthed metal structure connected to long conductors in open country, which is to say, an antenna. Type 2 surge protective devices on the DC side, proper equipotential bonding of the array structure, and a real earth electrode are not refinements; in storm-belt regions they are the difference between a tripped device and a dead one. The controller's internal protection is the last line, not the plan.

Enclosure rating and cooling. Ingress protection classes are defined by IEC 60529: the first digit for solids, the second for water. An IP54 drive in a ventilated, shaded kiosk, or an IP65 unit mounted in open shade, are both defensible; a drive in a sealed steel box in full sun is not, whatever its rating, because ratings address ingress, not heat. Power electronics are derated above a threshold ambient, typically 40 to 50 °C depending on the model, and site ambient in the regions we work in sits at or above that threshold for months at a time, before adding solar gain on the enclosure. Check the derating curve at your real ambient, mount the drive in shade with free airflow, keep the heatsink fins vertical and clean, and treat dust filters as a maintenance item. Heat does not usually kill a drive dramatically. It shortens capacitor life quietly, and the drive dies at four years instead of ten.

Cable runs and voltage drop. Boreholes put long cables between drive and motor, and the drop along that cable is lost head at the impeller. Conventional practice keeps voltage drop in the low single digits of percent, which on a 100-metre-plus drop cable frequently forces a cable one or two sizes larger than the current rating alone would suggest. Long motor cables on fast-switching drives also stress motor insulation through voltage reflection; past roughly 50 to 100 metres, depending on the drive, an output filter or reactor is cheap protection for a motor you cannot inspect. On the DC side the same arithmetic applies between array and controller, which is one more argument for placing the controller near the array and running AC down the hole.

The Affinity Laws: Why Slowing Down Costs So Little

Here is the piece of physics that makes the whole variable-speed approach work, and it is worth internalising because it drives both the economics and the control logic.

For a centrifugal pump, flow varies in direct proportion to speed. Head varies with the square of speed. Power absorbed varies with the cube. These are the affinity laws, and the cube is the interesting part.

Run the pump at 90 % speed and it draws roughly 73 % of full power. At 80 % speed, roughly half. At 70 % speed, about a third. So when a cloud takes away half the array's output, the drive does not need to halve the flow; it slows the pump by about 20 % and the power balance closes. The system sheds a painful amount of power for a modest loss of water delivered. This is why a variable-speed solar pumping system harvests usefully on hazy and partly cloudy days that would leave a fixed-speed machine tripping in and out, and why daily water volume is far more stable than the irradiance trace it rides on.

The cube law is also why the controller and pump must be chosen against the real duty. The flow-to-the-cube relationship holds cleanly where the load is mostly friction; in a borehole with high static lift the head requirement does not fall away as the pump slows, which squeezes the usable speed range and raises the minimum workable frequency. That interaction between static head, pump curve and speed range is precisely the calculation covered in our solar pump sizing guide, and it is the reason sizing a solar pumping system from a nameplate alone goes wrong so reliably.

The cube law is why a solar pump copes gracefully with passing cloud: losing half the array output costs only about a fifth of the speed. Source: pump affinity laws; MIMAH engineering analysis.

Sizing the Drive Against the Motor

A grid-fed VFD is normally sized at or one step above the motor's rated current. A solar pump drive needs slightly different thinking, in both directions.

On the drive side, current is the number that matters, not kilowatts. Match the drive's continuous output current to the motor's full-load current with margin, and then apply the temperature derating honestly: a drive that must deliver full motor current at a 45 °C enclosure ambient frequently needs to be one frame size larger than the catalogue kilowatt match. Manufacturers publish the derating curves; the failure mode for ignoring them is the quiet capacitor death described above.

On the array side, the array is commonly sized at roughly 1.2 to 1.4 times the motor's rated power. That looks like waste and is not. Array nameplate is a laboratory number; real output is trimmed by cell temperature, dust, and every hour that is not solar noon. The oversize buys a longer pumping day, earlier starts, later finishes and better cloud ride-through, and since panels are now the cheap part of the system, extending the pumping day with modules is usually the lowest-cost water you can buy.

On the motor side, check the drive's output matches what the motor actually needs across the speed range, and resist the urge to solve marginal delivery by raising the maximum frequency above 50 Hz. The cube law works against you just as powerfully in that direction: 55 Hz costs roughly a third more power than 50 Hz, and unless the motor and pump were selected for it, that margin comes out of the motor's thermal life.

Commissioning: The Parameters That Get Botched

A solar pump inverter leaves the factory with defaults, and defaults are written for a laboratory, not your borehole. In our experience, when a correctly sized system underperforms, the fault list is short and repetitive.

Motor data entered from imagination. Rated current, voltage, frequency and speed must come off the motor nameplate, not from the drive's defaults. The drive's thermal model protects the motor using these numbers; feed it fiction and it protects a fictional motor.

Minimum frequency left at zero. The drive happily dwells at 15 Hz all morning, the pump generates no useful head, delivers no water, cools nothing, and the day's report says "ran seven hours." Set the minimum from the pump curve so that below useful head the pump stops rather than stirs.

Dry-run detection never actually tested. Commissioning should include lifting the probe out of the water, or dry-running detection should be provoked with the valve arrangement, and the trip and auto-restart timed and witnessed. A protection that has never tripped in a test is a hope, not a protection.

Rotation unchecked. A three-phase pump runs backwards on a phase swap, delivers a fraction of its flow at full power draw, and sounds almost normal at the wellhead. Two minutes with a bucket and a stopwatch against the pump curve settles it.

Restart behaviour unconfigured. How long after a dry-well trip? How many retries before lockout? What happens at dawn, does the system start itself, and what happens after a grid or genset changeover on a hybrid unit? Every one of these is a parameter, and every one has a default that suits somebody else's site.

A disciplined commissioning sheet, an hour of measurements against the pump curve, and a saved parameter file cost half a day. If your pumping system was commissioned on defaults, or was never handed over with a parameter record at all, our operations and maintenance team can audit, correct and document it before the defaults choose their own failure date.

The commissioning settings that decide whether a correctly specified system survives. A disciplined commissioning sheet and a saved parameter file cost half a day. Source: MIMAH commissioning practice.

Frequently Asked Questions

Do I need a solar pump inverter, or will an ordinary VFD work? An ordinary industrial VFD provides the variable frequency output and motor protection but not the solar-side intelligence: MPPT on the DC input, sun-following frequency control, wake and sleep logic, and solar-appropriate dry-run handling. Some general-purpose drives accept firmware or option cards adding solar pump control, and configured properly they work well. What does not work is a plain drive with none of that logic, fed DC and left to its own devices; it will trip on under-voltage at every serious cloud.

What does MPPT mean on a pump drive? Maximum power point tracking. A solar array has one voltage, changing continuously with sun and temperature, at which it delivers its maximum power. An MPPT pump drive holds the array at that voltage by continuously adjusting the load the pump presents, which in practice means adjusting pump speed. Without it the array runs off its optimum and a meaningful fraction of the energy you paid for is simply never collected.

Can the same controller run from solar and the grid or a generator? Hybrid input controllers do exactly this, running solar-first and drawing from grid or genset when solar cannot meet the duty. They suit sites with a hard daily water requirement. Confirm automatic changeover, check the drive's tolerance of generator supply quality, and size the generator against the drive's input requirements rather than the motor nameplate.

Why does my pump stop and start on cloudy days? Below its minimum useful speed a centrifugal pump generates too little head to deliver water, so the controller stops it rather than let it churn, then retries when power recovers. Frequent short cycling suggests the minimum frequency, sleep delay and restart delay are poorly tuned for the site, or that the array is undersized for the season. Persistent cycling in good sun is a different animal, often a drawdown or dry-run issue, and worth investigating rather than tolerating.

Is a bigger drive always the safe choice? Within a frame size or two, oversizing the drive is a benign form of insurance, and in hot ambients it is frequently necessary once derating is applied honestly. A drive grossly oversized for the motor is a different matter: its current sensing resolution is coarse relative to the small motor's load signature, so motor protection and electronic dry-run detection both work worse. Size with margin, not abandon.

How long should a solar pump controller last? Electronics permitting, the honest answer is that heat writes the schedule. A quality drive kept shaded, ventilated and clean is reasonably expected to serve around ten years; the same drive in a sealed hot enclosure can halve that. The controller is also the most field-replaceable component in the system, which is a reason to favour architectures and suppliers with genuinely available spares over marginal efficiency claims.

The Box Is the System

A solar pumping system is bought as three line items but lives or dies as one machine, and the controller is where the machine's judgement lives. It decides how much of your array's energy becomes water, whether the motor survives the borehole's bad weeks, and whether cloudy season means reduced delivery or daily nuisance tripping. Specify it by its functions: real MPPT, smooth variable frequency control, soft start, tested dry-run protection, a minimum frequency set from the pump curve, honest thermal ratings at your real ambient, and the input flexibility your site's backup arrangements need. Every reputable manufacturer can meet that specification. The brand matters far less than whether anyone bothered to write the specification down.

We say this from delivery, not theory. MIMAH has installed roughly 1.39 MW across some 110 solar sites, and builds solar water pumping systems from 2 to 200 HP across Nigeria, Egypt, Sudan and the UK, which spans the DC package range at the bottom to serious AC drive systems at the top.

Specifying a pumping system, or living with one that underdelivers? Talk to our engineering team. We will size the array, drive and pump as one system, put the protection settings in writing, and hand you a machine that pumps water instead of excuses.