Troubleshooting Single-Phase Well Pump VFD Efficiency Losses

The pump in question is a 1.1 kW single-phase submersible pump set at 25 m. Its capacitor sits at the wellhead, and 40 m of cable run back to a single-phase-in, three-phase-out drive that feeds it from one output pair. A VFD output has no neutral, so the motor is connected across two output terminals.

The measured behavior is as follows:

• At 50 Hz the installation draws up to 1.5 kW from the grid.
• The pump will not start below 35 Hz.
• It works usefully only between 40 and 50 Hz, at 850 to 1350 W.

Four decisions follow from that data:

1. Whether the capacitor can be switched out while running.
2. Where the drive should be located.
3. Whether a VFD-to-mains bypass is sound.
4. Whether the drive is worth keeping at all.

Two measurements settle them. Liters delivered per kilowatt-hour settles the drive question. Main-winding current at full delivery settles the capacitor question.

Power balance at 50 Hz: 1.1 kW rating versus 1.5 kW grid draw

A grid-side kilowatt reading includes every loss between the meter and the water:

• the drive rectifier and DC link
• the IGBT output stage
• 40 m of cable
• the capacitor branch
• the motor copper and iron
• the hydraulic end

On submersible pumps the headline power is normally the rated shaft output (P2), not the electrical input (P1). Read the nameplate or the pump manual to confirm which one 1.1 kW refers to.

• If 1.1 kW is P2: a 1.5 kW grid draw at rated load is 1.1 / 1.5 = 73% wire-to-shaft. That is plausible for a small single-phase motor behind a drive.
• If 1.1 kW is P1: the installation is drawing 36% over rated input, and the current check below becomes urgent.

The current limit bounds the problem. Assume the drive delivers mains voltage at 50 Hz (230 V; the text also quotes 220 V). The 7 A maximum then corresponds to a single-phase apparent power of 230 × 7 = 1610 VA, or 1540 VA at 220 V. If all 1.5 kW reached the motor at 7 A, the power factor would have to be 0.93. Drive losses come off first, so motor input is below 1.5 kW. Only a current reading shows how close the motor sits to 7 A.

Measure it in one of two ways:

• Read the drive's output-current display.
• Take a true-RMS clamp reading on the pump feed during a direct mains test.

An averaging clamp on a PWM output reads wrong.

The 1350 W and 1.5 kW figures may come from different points. If 1350 W is the drive's displayed output power and 1.5 kW is the utility meter, the roughly 150 W gap is drive loss plus the error in the drive's power estimate. Compare only figures taken at the same point.

Also check the drive's voltage/frequency settings:

• Base (maximum) output voltage: should match the motor's 230 V. A higher setting over-fluxes the motor and raises current and heat at every speed.
• Base frequency: should be 50 Hz.
• Torque boost: excessive boost raises low-frequency current.

Both errors show up as heat, not as a logic fault. The drive keeps running while the winding cooks.

Quantity	Value / limit	Where to read it
Rated power	1.1 kW (P1 or P2 to be confirmed)	Pump nameplate / manual
Maximum current	7 A	Nameplate; compare with drive output current display
Apparent power at 7 A	1610 VA at 230 V; 1540 VA at 220 V	Derived: V × I
Grid real power at 50 Hz	Up to 1.5 kW	Energy meter upstream of the drive
Usable range	40–50 Hz, 850–1350 W	Drive display vs meter (confirm which point)
Drive base voltage / frequency	Match motor: 230 V / 50 Hz	Drive V/f parameter group

Head–speed physics behind the 35 Hz limit and the 40–50 Hz window

Centrifugal pump performance scales with speed n as follows:

• head scales with n²
• flow scales with n
• shaft power scales with n³

Relative to 50 Hz, the head available is:

• 81% at 45 Hz
• 64% at 40 Hz
• 49% at 35 Hz

The cube law predicts 40 Hz power of 1350 × (40/50)³ = 691 W. The measured figure is 850 W, which is 63% of the 50 Hz value instead of 51%.

The gap is the static head. The affinity laws hold only along a system curve that passes through zero head at zero flow, which is a friction-only system. A borehole lifting water through 25 m plus the rise to the tank has a large fixed head. At reduced speed the operating point slides left along the pump curve toward shutoff. Flow collapses faster than speed, pump efficiency drops, and motor and drive fixed losses become a larger share. Power falls more slowly than the cube, and water per kilowatt-hour falls with it.

At 35 Hz the pump's shutoff head is about half its 50 Hz shutoff head. If that is below the static lift from dynamic water level to tank inlet, the pump spins against a closed check valve and delivers nothing. A second mechanism looks the same from the surface: the motor fails to break away because torque is too low. Separate them with two observations:

Observation at 35 Hz	Motor turning, no delivery	Motor stalled
Drive output current	Settles at a steady running value	Sits at the drive's current limit
Drive frequency	Holds 35 Hz	Held back or trips on overcurrent
Water at the outlet	None	None
Cause	Hydraulic: head below static lift	Electrical: capacitor branch or V/f boost

A stall at low frequency traces to the capacitor. Its reactance rises as frequency falls, so auxiliary current and starting torque drop (see the capacitor section below).

Low-speed running is also a thermal problem for submersibles. The motor sheds heat into water flowing past its jacket. At or near zero delivery it sits in stagnant water, and winding temperature climbs with no current signature a general-purpose drive will trip on. Read the minimum cooling flow and any minimum operating frequency in the pump manual, and set the drive's lower frequency limit above that point.

Liters per kilowatt-hour: the test that decides VFD versus mains

Preliminary readings from the meter and tank level already favor 50 Hz. That is the expected result for a lift dominated by static head, where the energy needed per cubic meter rises as speed falls. Lower speed saves energy only when pipe friction dominates, as with a long or undersized delivery line.

The theoretical minimum energy to lift water is E = ρ·g·H·V. Assume the 25 m setting depth is the total lift. Then 1 m³ × 1000 kg/m³ × 9.81 m/s² × 25 m = 245 kJ, which is 0.0681 kWh/m³. For the real system, replace 25 m with the measured dynamic water level plus the height from wellhead to tank inlet.

1. Record the tank's internal cross-section area A in m².
2. Start every run from the same well rest condition. The dynamic level depends on how long the well has recovered.
3. Run each case to a fixed tank-level rise Δh:
   • drive at 40 Hz
   • drive at 45 Hz
   • drive at 50 Hz
   • direct mains at 50 Hz, with the pump fed through its capacitor without the drive
4. Read kWh on a meter upstream of whatever feeds the pump. For the drive cases this is upstream of the drive.
5. Compute volume V = A × Δh and specific energy E_meas/V in kWh/m³.
6. Compute wire-to-water efficiency: η = 0.0681 × (H/25) × V / E_meas.
7. Repeat each case at least twice. Discard runs where the well level drifted.

A PV installation can change the answer. The tank is filled only on photovoltaic surplus and discharged by gravity. In that case, grid kWh per m³ is the figure of merit, not total kWh per m³. The drive's 850–1350 W window lets the pump track surplus as clouds pass. A fixed 50 Hz pump drawing 1.5 kW imports from the grid whenever surplus dips below that. Log grid import during both modes on a variable day before concluding the drive is useless.

Run capacitor behavior under PWM and off-nominal frequency

Single-phase induction motors at this rating come in three topologies:

• PSC (permanent split capacitor): the auxiliary winding and a film run capacitor stay connected continuously.
• Capacitor-start: a start capacitor, and often the auxiliary winding itself, drops out after starting via a current relay or centrifugal switch.
• Dual-capacitor: one capacitor stays in circuit and a second drops out after the start.

Pump load needs little breakaway torque, so pump motors are commonly built as PSC.

On this Lowara unit, the power winding measures less than half the resistance of the capacitor-fed winding. That ratio does not tell the topologies apart. A PSC auxiliary winding is also wound with more turns of finer wire, sized so that together with the capacitor it produces the design phase shift and voltage. Its resistance is higher in both families. The decisive evidence is:

Evidence	PSC (keep capacitor in)	Start-only auxiliary winding
Control box contents	One capacitor, permanently wired	Capacitor plus current relay or centrifugal switch
Capacitor marking	Continuous-duty run rating (film)	Intermittent-duty start rating (electrolytic)
Manual wiring diagram	Auxiliary winding always energized	Auxiliary circuit opened after start

Capacitive reactance is X_C = 1/(2πfC). At 40 Hz it is 1.25 times its 50 Hz value. Auxiliary current falls and the phase shift drifts from the design point, and three effects follow:

• The rotating field turns elliptical, with a backward-rotating component.
• Torque drops.
• Losses rise.

Removing the capacitor from a PSC motor at low frequency reduces torque further. Adding capacitance in parallel at low speed moves the other way. Holding X_C constant requires C_f = C₅₀ × 50/f, which gives 1.11 × C₅₀ at 45 Hz and 1.25 × C₅₀ at 40 Hz. Winding reactances also fall with frequency, so the true balance point can need more than that. Tune on the scope: measure auxiliary and main winding currents and adjust toward 90° between them. A single capacitor centered on 45 Hz is a compromise across the 40–50 Hz band.

The capacitor also sees PWM. Here it sits in series with the auxiliary winding, and the winding inductance limits the current at each switching edge. The switching ripple still flows through the capacitor, which is rated for a 50 Hz sine. After a full-length run, measure the capacitor case temperature against ambient.

A hot or swelling run capacitor is heading for an open auxiliary circuit and a motor that cannot start. Do not connect a capacitor directly across the drive output terminals with no winding in series. The edge currents can destroy the output stage.

Capacitor cut-out relay: gain, risk and interlock logic

In the field test the capacitor was opened with a miniature breaker while the pump ran. The motor kept running, sounded smoother at the low 50 Hz harmonics when listened to through the pipe, and every measured current fell slightly.

At off-nominal drive frequency this is plausible. A mis-tuned capacitor branch draws current that builds backward field instead of torque, and removing it removes that current. At 50 Hz and full delivery the picture changes for a PSC motor:

• The main winding carries the whole load alone.
• The field pulsates, producing torque ripple at twice supply frequency (100 Hz on a 50 Hz supply).
• Main-winding current rises relative to balanced operation.

A small drop in total line current can hide a rise in main-winding current.

Decide on the numbers. At 50 Hz and steady full delivery, log three quantities with the capacitor in and then with it out:

• main-winding current, clamped on the main-winding lead only
• total real power in watts
• flow

Keep the capacitor if either of these happens:

• Main-winding current approaches or exceeds 7 A.
• Watts per liter does not fall.

A few tens of watts are not worth rewinding a submersible motor pulled from 25 m.

If the relay goes in anyway, wire it fail-safe:

1. Put a normally-closed contact in the capacitor path, so a de-energized relay leaves the capacitor connected for every start.
2. Energize (open) the contact only after the drive reports running frequency for a set time and the pressure sensor confirms delivery.
3. Drop the relay before any stop command, on any drive fault, and on loss of controller supply.
4. Feed a relay auxiliary contact back to the controller, and block any start unless that contact confirms the capacitor is connected.
5. Rate the contact for at least the capacitor's marked voltage. In a PSC circuit the capacitor voltage exceeds line voltage.

Do not allow a restart with the capacitor open. With the auxiliary branch open, a single-phase motor has zero starting torque. It sits at locked-rotor current until its protection trips, and a submerged winding overheats fast in that state.

Cable I²R losses and drive location

The 40 m house-to-wellhead run is a two-conductor loop, 80 m long. Loop resistance is R = ρ·2L/A, with ρ_Cu ≈ 0.0175 Ω·mm²/m at 20 °C. The conductor size is not given, so the table assumes two common sizes at the 7 A maximum.

Assumed cross-section	Loop R (80 m)	I²R loss at 7 A	Voltage drop at 7 A
1.5 mm²	0.93 Ω	46 W	6.5 V (2.8% of 230 V)
2.5 mm²	0.56 Ω	27 W	3.9 V (1.7% of 230 V)

Loss scales with current squared, so it is lower at the 850 W operating point.

Moving the drive to the wellhead does not remove this cable. The drive's input current flows through it instead. A single-phase diode-bridge front end draws current in narrow pulses near the voltage peaks. Its RMS current is well above the fundamental needed for the same real power. On the motor side, the PSC motor's power factor also inflates current, so the two cases are comparable. Measure both with a true-RMS clamp: drive input current versus drive output current as reported by the drive.

Relocation buys PWM cable length, not energy. The PWM run is currently at least 65 m (40 m surface run plus 25 m drop cable). A shorter run means:

• less capacitive charging current
• lower voltage overshoot at the motor terminals
• less stress on the output stage and on the motor insulation

Check the drive manual's maximum motor cable length for unshielded cable at the switching frequency in use. A wellhead drive also needs an enclosure suited to condensation and outdoor temperature.

Start and stop methods compared for a 25 m borehole pump

A soft start here prevents mechanical damage in the shared borehole:

• Direct starts produce a torque reaction that twists the riser and wears the rope.
• The reaction can back off threaded fittings.
• It can wrap one pump's pipe around the neighbor's.

If that happens, both sets may have to come out together, a combined 120 kg. A pump-specific inverter and a general-purpose drive produce the same waveform on a scope, with the same switching frequency. What differs is protection logic: pressure switch or flow switch inputs, dry-run detection, and restart suppression for small leaks. A general-purpose drive needs that logic supplied externally, here by the pressure sensor and ESP32 controller.

Method	Start torque kick	Losses during 50 Hz run	PV tracking	Direct mains fallback	Main risk
VFD continuous (present)	Gentle ramp	Drive losses plus PWM ripple in capacitor	Yes, 850–1350 W	Rewire needed	Capacitor heating; low-speed cooling
VFD start, contactor transfer to mains	Gentle ramp	None from drive once transferred	No while on mains	Yes	Unsynchronized transfer transient; backfeed if interlock fails
Single-phase soft starter, mains run	Reduced (torque ∝ V²)	Near direct-online with bypass relay	No	Inherent	Starter must support capacitor-run motors
Direct-online on mains	Full breakaway torque	Lowest	No	Inherent	Torque reaction on rope, pipe and fittings
230 V three-phase pump on existing drive	Gentle ramp	Drive losses, no capacitor	Yes	Lost on single-phase supply	New pump and a pull from the shared bore
Pump maker's inverter-and-motor package	Gentle ramp	Designed as a set	Depends on model	Depends on model	Cost

Recommended arrangement: mains run through a single-phase soft starter

For irrigation duty at a fixed 50 Hz, run on mains through a single-phase soft starter with the capacitor left connected. This arrangement:

• removes drive losses and PWM ripple from the capacitor during long runs
• keeps a reduced-torque start to protect rope and fittings
• never switches between two unsynchronized sources while the motor turns

Centrifugal pump torque starts near zero and rises with speed squared. Reduced-voltage starting therefore delivers enough torque through the whole ramp.

Keep the VFD only for PV-tracked tank filling, and only if the liters-per-kWh test shows lower grid import. Install it as a second path selected at standstill.

1. Confirm from the control box and manual whether the motor is PSC or start-capacitor type.
2. Choose a single-phase soft starter that lists capacitor-run motors as supported. It should be rated at or above 7 A continuous and have an internal bypass relay that closes at the end of the ramp.
3. Wire it as starter output to capacitor box to motor, so the auxiliary branch sees the ramped voltage exactly as it sees mains.
4. Set the initial voltage just high enough that the motor breaks away early in the ramp. On the first starts, confirm that current falls to its running value before the ramp ends. A ramp that ends with current still high means the motor stalled and heated.
5. Set a soft-stop ramp to reduce check-valve slam from the 25 m water column.
6. For the VFD/soft-starter changeover, use mechanically and electrically interlocked contactors, break-before-make. The controller should allow the changeover only when both paths report stopped and the pump current is zero.
7. Set motor overload protection at the 7 A nameplate current.

VFD-to-mains transfer: interlocks, dead time and flying restart

Starting on a drive and transferring to mains has long been done on large three-phase pumps. The reverse transfer was the hard part until drives gained the catch on fly (flying start) function, which synchronizes the drive onto a spinning motor.

The transient comes from phase. A drive running at 50.0 Hz is not phase-locked to the grid. When the mains contactor closes, the motor's residual voltage and the supply can be anywhere up to 180° apart, and the resulting current and torque pulse can exceed a direct start. During the open-transition dead time the residual voltage decays, but the rotor also slows against the 25 m head. In practice the transfer behaves like a direct restart from reduced speed, which is the torque event the drive was installed to avoid.

If the bypass is built anyway, sequence it strictly:

1. Hold the drive at 50 Hz with steady current.
2. Remove drive run-enable so the output coasts. Do not open the output contactor while the drive is modulating.
3. Open the drive output contactor and confirm it open through its auxiliary contact.
4. Close the mains contactor.
5. To stop softly, open the mains contactor and confirm it open, close the drive contactor, start the drive with flying start at 50 Hz, then ramp down.

Mains and drive output must never be connected together. Mains voltage on the drive output terminals destroys the output stage, so use a mechanical interlock between the two contactors in addition to the controller logic.

Without flying start, a drive starting into a motor spinning near 50 Hz trips on overcurrent or brakes it hard. Check the drive manual for a speed-search or flying-start function before committing to soft stop through the drive. Otherwise simply stop on mains.

Post-change checks on current, capacitor heat and delivery

Check	Pass condition	Where to read it
Running current, 50 Hz full delivery	≤ 7 A	True-RMS clamp on mains feed or drive output display
Main-winding current (if capacitor switching is tried)	No rise versus capacitor-in baseline	Clamp on main-winding lead only
Real power	Below the 1.5 kW baseline for the same delivery	Energy meter at the same point as baseline
Specific energy	Lowest kWh/m³ of all cases tested	Tank Δh × area vs meter kWh
Capacitor temperature after a full run	Small rise over ambient, no swelling	Contact or IR thermometer on the case
Start ramp	Current falls to running value before ramp end	Clamp trend or starter/drive display
Dry-run / underload trip	Trips between shutoff current and normal running current	Measure both, set threshold between them
Restart suppression	No restart on small leaks; minimum off-time enforced	Pressure sensor log on the controller
Motor and cable insulation	At or above the pump manual's minimum, stable trend	Insulation tester, conductors to earth, pump isolated

FAQ