Lift Station and Pump Service

Purpose

This SOP takes a technician into a wet well full of effluent with a 120 or 240 volt motor hanging in it, and gets him back out having proved the pump is right rather than having watched it run.

Two things go wrong in this work and both are avoidable. The first is electrical: a submersible effluent pump sits in a grounded, conductive liquid, its cord is dragged over a concrete edge every service, and the control panel it feeds from is a yard-mounted box that has been rained on for fifteen years. The second is quieter and shows up as a drainfield replacement three years early, because "the pump runs" was accepted as a result. A pump that runs can still be dosing the field at half again the volume the design called for, and nothing on the panel will say so.

Scope

Covers service of submersible effluent and sewage pumps in residential dosing chambers, pump tanks and small commercial lift stations, performed entirely from outside the opening.

Does NOT cover entry: the moment any part of a person breaks the plane of the opening, this SOP has ended and the shop's permit-required confined space program under 29 CFR 1910.146 takes over. It also does not cover control panel replacement or new branch circuit work, which are licensed electrical scope; atmospheric testing is owned by the gas testing SOP; field acceptance is owned by the drainfield failure assessment SOP.

Roles and handoffs

Procedure

1. Isolate, lock, tag, and prove dead before anything wet is touched. Open the disconnect ahead of the panel, apply your own lock and tag, then prove the tester on a known live source, test every conductor and every conductor to ground, and prove the tester again. Acceptance: 0 V phase to phase, phase to neutral and phase to ground, with the tester proven live before and after. Wrong is turning the hand-off-auto switch to OFF and calling it isolated, because that switch opens a control circuit and leaves the line side energized. The proving sequence is NFPA 70E-2021, 120.5, and the de-energizing practice is 29 CFR 1910.333(b)(2); 1910.147 expressly excludes this electrical work at (a)(1)(ii)(C).

2. Prove the control circuit separately, because it is often fed from somewhere else. Many panels carry a separate control transformer or a second breaker for the alarm circuit, so a panel can read dead on the load side and still have a live alarm bus. Acceptance: every terminal you will touch, including float and alarm terminals, read 0 V with the lock on. Wrong is a float terminal that comes up live after the main is locked; the stop rule is to close the panel, leave the lock on, and call a licensed electrician, because an unidentified second source is not something to trace with your fingers over an open wet well.

3. Test the atmosphere before the lid moves, per the gas testing SOP. Acceptance: opening-plane and depth readings inside their gates, values written down, meter left clipped at the opening in continuous mode. Wrong is any failed gate, and the stop rule is mechanical ventilation and a re-test with the blower off. A wet well holds a deeper headspace than a septic tank and a pump lifting out of it drags the bottom layer to the opening, so the reading that matters is taken again at the moment the pump breaks the surface.

4. Record every level against a fixed datum BEFORE anything is lifted. From a mark on the rim, measure to the static liquid surface, the off float trip point, the on float trip point, the high-water alarm float trip point, and the inlet invert. Acceptance: five figures written in inches from one datum. Wrong is pulling the float tree first, which destroys the only evidence of how the system was actually set; a float geometry cannot be reconstructed once the tether lengths have moved. Reach with a marked probe from the side of the opening, never by leaning across it.

5. Relieve the stored head before the union is broken. The discharge column above a check valve is stored energy under 29 CFR 1910.147, and it will dump back through a broken union at the tech's hands and face. Crack the union slowly with the fitting turned away from you and let the column drain back into the well. Acceptance: flow from the union stops before the fitting is fully separated. Wrong is a fast break that sprays effluent into the breathing zone and across the panel; face shield and chemical-resistant gauntlet gloves go on before the union is touched, not after.

6. Lift the pump on its lifting rope or eye, never on the cord or the discharge pipe. Take the weight with two hands or a tripod and hoist, keeping the cord slack and out of the load path. Acceptance: pump clear of the opening with the cord jacket unloaded and undamaged. Wrong is lifting by the power cord, which stretches the cord entry seal and puts water into the motor on the next cycle, and is also how a pump gets dropped back into the well onto a tech's forearm. A wet pump body is heavier than it looks and the rim is slick, so it lands on a tarp beside the opening, not balanced on the rim.

7. Inspect the pump, then verify the check valve holds. With the circuit still locked out, turn the impeller by hand, inspect the volute and cutter for rags, check the cord jacket where it crosses the rim, and look at seal chamber oil clarity if the pump has a sight plug. Then refit and fill the discharge, and watch the well level. Acceptance: impeller free, no jacket damage, oil clear, and the well level static rather than rising after the pump stops. Wrong is a level that creeps up after shutdown, which is a check valve passing back; that pump will re-lift the same effluent and log cycles the household never generated, so the valve is replaced before any float conclusion is drawn.

8. Restore power in a stated sequence and read running current against the nameplate. Remove the lock, close the disconnect, and take the clamp reading on the pump cord outside the enclosure with the cover closed. Acceptance: running amps at or under the nameplate full-load amps, read alongside the measured supply voltage. Wrong is a current above nameplate, and before you condemn the pump, read the voltage: at a constant shaft load a motor supplied below its rated voltage draws MORE current, which on a long buried run is a wiring fault, not a pump fault. If the cover has to come off to reach a test point, that is energized work requiring a permit and arc-rated PPE under NFPA 70E-2021, so the reading waits for the electrician.

9. Measure drawdown and compute the dose per cycle. Run one full cycle on automatic and measure the depth between the on and off trip points. Compute the volume as chamber cross-sectional area times drawdown depth, using 3.1416 times the radius squared for a round chamber and 7.48 gallons per cubic foot. Acceptance: dose per cycle within 10 percent of the design dose on the permit. Wrong is a dose over that band, which over-fills the trench each cycle and ponds a field that is otherwise sound; reset the tether lengths, re-run, re-measure, and where the design dose is unknown, stop and get it from the designer rather than guessing. Then raise the alarm float by hand and confirm the alarm sounds and the light latches.

When the site does not match

A disconnect that cannot accept a lock is a stop: the pump does not get serviced on that panel until a lockable disconnect is fitted, and the customer gets that in writing. A pump hung on a corroded rope or a seized guide rail is a two-person recovery job with proper rigging, not something to be freed by leaning in, and it gets rebooked. Where no design dose exists because the system predates the record, measure and log the drawdown anyway and mark the acceptance "no design dose on file", so the next visit has a baseline even without a target. A duplex station where only one pump will start is a control fault: leave the working pump in service, red-tag the failed side, and hand it to the electrician with the readings attached.

The record this produces

One lift station service record: date, both technicians, pump make, model and nameplate full-load amps, the five datum measurements in inches, atmospheric readings, check valve back-flow result, impeller and cord condition, running current with the measured supply voltage beside it, drawdown depth, computed dose per cycle, design dose from the permit, the percentage difference, alarm test result, and any float reset with the before and after tether lengths.

The service manager compares dose against design across the route and finds over-dosed fields before they surface. Whoever quotes the next drainfield reads the dose history and knows whether the field was ever given a fair chance. The electrician handed a repeat overload gets current and voltage together instead of a phone description. And the reason both the current AND the voltage are recorded is that neither one means anything alone.

Worked pass: residential dosing chamber, 48 in diameter, pump reported running frequently

Two techs, ticket carried the pump model, a nameplate full-load current of 9.4 A and a design dose of 60 gallons. Disconnect locked and tagged, tester proven both ways, all conductors and the alarm terminals read 0 V. Atmosphere passed at all depths.

Datum readings from the rim: static liquid 41 in down, off float 44 in, on float 32 in, alarm float 26 in, inlet invert 22 in. Union cracked slowly and the column drained back; pump lifted on its rope onto a tarp. Impeller free, cord jacket sound, oil clear, and after refitting the discharge the well level held static, so the check valve passed. Power restored, clamp reading 8.8 A at 231 V measured, inside the nameplate 9.4 A.

Step 9 failed. Drawdown between the on and off trip points measured 12 in. A 48 in chamber is 2 ft in radius, so 3.1416 times 2 squared is 12.57 sq ft, and one inch of depth holds 12.57 divided by 12, times 7.48, which is 7.83 gallons. Twelve inches is therefore 94 gallons per cycle against a 60 gallon design dose, 57 percent over and far outside the 10 percent band. Under the step 9 stop rule the tethers were shortened to give 8 in of drawdown, which is 63 gallons, 5 percent over design and inside the band. Re-run confirmed 8 in. Alarm float raised by hand, alarm sounded and latched.

Checking the run against the rules above: the 8.8 A is compared with the nameplate 9.4 A at the measured 231 V, and both were recorded because a current alone would not have distinguished a tired pump from a sagging supply. The 94 and 63 gallon figures both come from the same 7.83 gallons per inch and the same measured chamber, and the pump's frequent running was explained by the float spacing rather than by the check valve, which was tested and held.

References

• NFPA 70E-2021, 120.5, in the edition your employer's electrical safety program has adopted, for establishing an electrically safe work condition and the proving sequence.
• 29 CFR 1910.333(b)(2) for electrical safe work practices; 29 CFR 1910.147 for the stored hydraulic energy in the discharge column, noting its electrical carve-out at (a)(1)(ii)(C).
• 29 CFR 1910.146 for the wet well as a permit-required confined space, which this SOP stays outside of.
• Pump and panel manufacturer documentation for nameplate current, seal chamber oil condition and float tether specification.
• See related: the gas testing SOP, the drainfield failure assessment SOP, and the alarm call response standard SOP.