Booster Pump System Service

Purpose

A starved booster pump reproduces every symptom of a worn one. Falling discharge under load, a lag pump running all day, cavitation noise: it all looks like a tired impeller and none of it is, when the real cause is a throttled suction valve, a loaded strainer or a supply that cannot hold residual at the morning peak. This procedure guarantees the suction side is measured and proven before anybody opens a pump, that the low-suction protection is verified rather than assumed, and that the set is put back with its rotation and staging checked. Skip it and a shop sells an impeller into a supply problem.

Scope

Covers routine and complaint-driven service of packaged domestic water booster sets in small commercial buildings and low to mid-rise multifamily: suction verification, performance logging, strainer and check valve service, mechanical seal inspection, hydropneumatic tank pre-charge, and return to service.

Does not cover pump or motor rebuild, sizing or replacement selection. Does not cover any fire pump, which is a different machine under a different standard and is not serviced under this procedure. Does not cover floor pressure reducing valves, owned by the pressure reducing valve replacement SOP. Electrical work beyond isolation and rotation verification goes to an electrician.

Roles and responsibilities

Procedure

1. Pull the controller history and the complaint pattern before you open anything. Read logged suction and discharge, the discharge setpoint, run hours per pump, fault history and which pump is leading. Acceptance: setpoint, run hour split and any recorded low-suction events written down. Wrong looks like starting at the pump because the pump is what the complaint named. Stop rule: a run hour split badly skewed toward one pump means the alternation is not working, and that is a control fault to resolve before any mechanical conclusion. Hazard: the controller enclosure stays closed unless a qualified person opens it, and nothing inside it is touched to read a display.

2. Measure suction pressure during the actual complaint window and prove the suction path is open. Read suction at the building's peak draw, confirm the suction isolation valve is fully open, and pull the strainer. Acceptance: a suction reading in psi at a stated time of day, valve position confirmed at the stem or indicator, clean strainer basket. Wrong looks like a suction reading taken at midnight on a building whose problem is at 7 am. Stop rule: suction below the minimum residual your water purveyor states for their main, commonly around 20 psi and a number that is theirs rather than the plumbing code's, is a supply conversation for the lead before any pump work. Hazard: pulling a strainer opens a pressurized potable line, so isolate, relieve to a drain and confirm zero at a gauge before the cover moves, and stand to the side as it releases.

3. Log a performance snapshot as a pair, at a known draw. Record suction and discharge together, with the number of pumps running and the drive frequency where fitted. Acceptance: paired readings at both a light draw and a peak draw, with the time of each. Wrong looks like a single discharge figure, which cannot distinguish a pump that will not make head from a supply that will not deliver. Stop rule: where discharge falls but the difference between discharge and suction holds up, the pump is making its head and the problem is on the suction side; that finding stops the mechanical work. Hazard: none at this step beyond reading gauges, and taking readings while the set stages is worth doing from outside the guard rather than leaning over a coupling.

4. Isolate fully before any cover comes off, electrical and mechanical and pressure. De-energize at the disconnect, close suction and discharge isolation on the pump being worked, and relieve casing pressure to a drain. Acceptance: zero at the casing gauge, valves closed and tagged, and the disconnect locked and tagged. Wrong looks like a casing drain cracked on a hot pump, where the water flashes at the drain. Stop rule: no cover moves until the casing gauge reads zero, whatever the isolation valves suggest. Hazard: two separate energies here, so mechanical isolation and stored pressure follow 29 CFR 1910.147 while the electrical work follows 29 CFR 1910.333(b)(2) by a qualified person under 1910.332, proved dead on a known live source before and after per NFPA 70E-2021, 120.5, and a drive is left for the manufacturer's stated bus discharge time before its cover comes off.

5. Inspect the check valves and the mechanical seal while the pump is open. Acceptance: each check valve seating with no backflow when the adjacent pump runs, and a seal with no weep and no scoring on the sleeve. Wrong looks like a failed check valve left in, which lets the running pump push back through the idle one so the set never makes setpoint and both pumps run. Stop rule: a check valve that will not seat is replaced now rather than noted, because the symptom it produces is the one the customer called about. Hazard: seal faces and impeller edges cut, so gloves stay on and nobody spins an open pump by hand.

6. Check the hydropneumatic tank pre-charge against the control setting, with the water side at zero. Relieve the tank's water side to zero, then read the air valve with a low-range gauge. Acceptance: pre-charge at the cut-in pressure minus 2 psi on a staged set, or at the manufacturer's stated value for a variable speed set, measured with the water side relieved and recorded. Wrong looks like a pre-charge read against a pressurized water side, which reads the system and not the tank. Stop rule: a tank that gives water at the air valve has a failed bladder and is replaced, not re-charged. Hazard: relieve the water side to a drain and not to the floor, and use a hand pump rather than shop air on the charge, since an uncontrolled shop line will overshoot and stress the bladder.

7. Restore power, verify rotation, and prove the low-suction protection still works before the set carries the building. Acceptance: rotation correct against the casing arrow on a brief bump, the low-suction cutoff setpoint confirmed at or above the purveyor's stated residual, and the cutoff seen to operate on a controlled test per the manufacturer's method. Wrong looks like a set restored after electrical work with rotation unverified, which runs backwards, makes about half its head and sounds almost normal. Stop rule: a low-suction cutoff that does not operate on test is a hard stop; the set does not carry the building until it does, because that device is what stops the pumps pulling the main down and running dry. Hazard: this is the step that puts energy and pressure back with the guard off and people nearby, so guards go on first, everyone stands clear of the coupling for the bump test, and the panel work stays with the qualified person from step 4.

8. Prove the set under a real draw before you leave the room. Open enough fixtures or a hose bib to load it, watch the pumps stage up and down, and read suction and discharge again as a pair. Acceptance: discharge holding setpoint within the controller's stated band through staging, suction holding above the cutoff, no cavitation noise, and pumps sequencing without hunting. Wrong looks like a set tested at no flow, which will hold setpoint even when badly worn. Stop rule: hunting between stages after service is a control setting, not a mechanical fault, and it gets resolved before handover rather than left for the building to live with. Hazard: hot water can sit in a casing that ran at no flow, so open a draw before you crack anything, and keep hands off the casing until it has carried flow.

9. Disinfect what you opened and close out the record. Where the potable system was opened, disinfect and flush as your adopted code requires, which the IPC covers at Section 610 in the edition your jurisdiction enforces. Acceptance: disinfection performed and flushed to the required condition, and every reading from steps 2, 3, 6, 7 and 8 written down. Wrong looks like a potable system returned to service with no disinfection because the opening was brief. Stop rule: where the building cannot be taken off line for it, that is scheduled work with the property, not a step to skip quietly. Hazard: the disinfecting chemical is handled to its own SDS with the protection it names, never mixed with anything else, and any respirator only under a written program per 29 CFR 1910.134.

The record this produces

Per visit: controller setpoint, run hours per pump, fault history, suction and discharge pairs at light and peak draw with times, suction valve position, strainer condition, check valve results, seal condition, tank pre-charge with the cut-in it was set against, rotation verified, low-suction cutoff setpoint and test result, staging under load, and disinfection if performed.

These land on the equipment record. The next tech reads last visit's peak suction pair before touching a pump, which is the point of the procedure. The lead reads the suction trend across visits, because a suction that falls year on year is a purveyor conversation that no amount of pump work fixes. The property reads the cutoff setpoint, since that number stands between their pumps and a dry run.

One worked pass, including a failure

Five-story multifamily, duplex booster set, complaint of weak pressure on the top floor between 6:30 and 8 am. Height to the top floor fixture is 52 ft, and at 2.31 ft of water per psi that is 52 divided by 2.31, or 22.5 psi of static loss to the top.

Step 1: discharge setpoint 65 psi, run hours split heavily toward pump 1, three recorded low-suction faults in the last quarter.

Step 2 fails. Suction at 7:10 am reads 18 psi. The suction isolation valve is found roughly two thirds closed, left that way after a shutdown, and the strainer basket is loaded. The purveyor's stated minimum residual on that main is 20 psi, so 18 psi at the pump suction is under it, and the low-suction cutoff is set at 10 psi, which means it was never going to protect anything. Stop rule taken: no pump work proceeds on the starved suction. The valve is opened fully, the strainer cleaned, and the cutoff setpoint is raised to the purveyor's stated residual after confirming that number with them.

Step 3 before the correction: at 7:10 am discharge was holding only 48 psi. At the top floor that is 48 minus 22.5 static, or 25.5 psi at the riser top, less about 5 psi of friction to the fixture, so about 20.5 psi at the tap, which is the complaint. After the correction, suction at the next morning peak reads 46 psi and discharge holds the 65 psi setpoint, so the top floor sees 65 minus 22.5 minus 5, about 37.5 psi.

Step 5 finds pump 2's check valve not seating, which explains the run hour skew: pump 1 was pushing back through pump 2 whenever it ran alone. Replaced. Step 6 reads the tank pre-charge with the water side relieved to zero, finds it 4 psi below cut-in minus 2, and corrects it. Step 7 verifies rotation on a bump and confirms the raised cutoff operates on the manufacturer's test. Step 8 loads the set at three hose bibs: discharge holds inside the controller band through a stage up and a stage down, no hunting, no cavitation. No potable opening beyond the strainer and check valve, so step 9 disinfects that section and flushes it.

References

• 29 CFR 1910.147 for mechanical isolation and stored energy, 29 CFR 1910.333(b)(2) with 1910.332 for electrical work by a qualified person, and NFPA 70E-2021, 120.5 for the proving sequence.
• Your water purveyor's stated minimum residual pressure and their rules on booster connections, which are theirs and not the plumbing code's, and which govern the low-suction cutoff setting.
• IPC Section 610 for disinfection of a potable water system after it is opened, in the edition your authority having jurisdiction enforces.
• Booster set and drive manufacturer documentation for staging bands, pre-charge, bus discharge time and the low-suction cutoff test method.
• See related: the Plumbing pressure reducing valve replacement SOP and the shared riser shutdown and tenant notification SOP.