Variable Speed Pump Replacement and Programming

Purpose

This procedure guarantees that a variable speed pump replacing a single speed unit is commissioned against the flow every connected device actually needs, and leaves the pad with written speeds a second technician can restore. The retrofit fails two predictable ways. The first is factory default speeds left in place, which either misses turnover or throws away the reason the customer bought the pump. The second is quieter: a drive wired downstream of the mechanical time clock that used to switch the old motor, so the drive loses power whenever the clock opens and its schedule never runs, while everybody blames the pump.

Scope

Covers replacing a single speed filter pump with a variable speed unit on a residential or light commercial pad: supply and control survey, flow-target arithmetic, commissioning, the minimum-flow survey, speed programming and handover.

It does not cover sizing to the plumbing, bonding, or the mechanical set and union work, which the pump and motor replacement SOP owns in full and which apply here unchanged; the branch circuit, disconnect or clock removal, which are the electrician's; or controller-side scheduling, which the automation controller SOP owns.

Roles and responsibilities

The speed floor is set by the neediest device, not by the energy math

Three relationships and one rule decide every speed on the card.

The affinity relationships. At constant impeller diameter, flow varies with speed, head with the square of speed, and shaft power with the cube. Half speed is roughly half the flow, a quarter of the head and about an eighth of the power, and because the same turnover then takes twice as long, the energy per turnover falls to about a quarter rather than an eighth. That cube behavior approximately survives on a pool filtration loop because such a loop is nearly all friction head with little static lift; a system that genuinely lifts water (a raised water feature, roof-mounted solar) saves less than the cube implies.

Turnover. Pool volume divided by design turnover hours divided by 60 gives the gpm the filtration speed has to deliver. The pipe ceiling. No speed may push velocity past what the plumbing tolerates; the pump and motor replacement SOP carries that arithmetic, so use its figures rather than re-deriving them.

And the rule this article owns: the lowest usable speed in any mode is the highest minimum flow among the devices that must operate in that mode. A heater's flow switch, a salt cell's flow sensor, a suction cleaner, an in-floor system and a solar array each publish their own minimum, and the lowest is irrelevant. Programming to the pump's own minimum is what produces a heater that lights and drops out every four minutes, or a cell reporting no flow while water is visibly returning. The floor is discovered by running the pump up and finding where each device holds in, then programming above it with margin.

Procedure

1. Open the disconnect, lock and tag it, and prove the motor terminals dead on a meter proved on a known live source immediately before and after. Accept when every conductor combination reads zero and the meter re-reads correctly on the known source. Wrong is a dead reading from a dead meter. Work practices are 29 CFR 1910.333(b)(2), not 1910.147, which excludes electrical utilization work at its own (a)(1)(ii)(C); the proving sequence is NFPA 70E-2021, 120.5.

2. With the circuit still locked out, survey the supply and the control path before anything comes apart. Confirm the measured supply voltage against the new drive's input rating, whether the supply is continuous or switched by a time clock or relay, GFCI protection on the branch circuit, and how the pump will be commanded (onboard schedule, dry-contact speed inputs, or a communication cable to a controller). Accept when the ticket states measured voltage, continuous or switched, and the control method chosen. Wrong is a clock-switched supply: stop, do not commission, and book the electrician, because a drive that loses power keeps no schedule. Any measurement taken inside an energized enclosure is qualified-person work under 29 CFR 1910.332 and 1910.399; if nobody on site is qualified, that reading goes to the electrician rather than being taken anyway.

3. Write the three flow targets down before choosing a single speed. Required turnover flow, the plumbing's ceiling flow, and each connected device's published minimum flow taken from that device's own manual with the model noted. Accept when all three are on the ticket with sources. Wrong is carrying a remembered minimum from another make; heater flow switches differ by model and fuel. Stop and call the shop lead if a device's published minimum exceeds the plumbing ceiling, because that is a plumbing conversation. Hazard: none at this step, it is arithmetic done at the tailgate.

4. Set the pump, make up the unions, and land the bonding conductor per the pump and motor replacement SOP. Accept when the pump sits flat, both unions seat squarely with no pipe weight on the ports, and an 8 AWG solid copper bonding conductor is landed on a clean lug with a verified continuous path, per NEC Article 680 in your jurisdiction's adopted edition. Wrong is a wrench on a plastic union nut or a lug with the copper corroded green. Stop and do not energize if the bonding path is open: that repair is the electrician's.

5. Land the power conductors and any communication cable, set the drive's input configuration, and close the enclosure. Keep low-voltage communication conductors out of the line conductors' raceway unless the drive's manual permits it. Accept when the configuration matches the measured supply from step 2, terminal covers and conduit seals are closed, and no conductor is pinched. Wrong is landing a dual-voltage unit on the wrong configuration, which destroys it in seconds and is not a warranty claim. Hazard: an open terminal cover on a wet pad is a shock path, so close it before the lock comes off, not after.

6. Fill the pot, remove the lock and tag, energize, and prime - then verify the protections you disturbed. Accept when the pot runs full and clear at prime speed, the filter relief passes solid water, the branch circuit GFCI trips on its test button and resets, and the bonding conductor is still landed. Wrong is a pump that surges or will not hold prime, which is an air leak, not a programming problem. This is the step where energy, pressure and motion all go back at once: stand to the side of the tank and pot lid, keep hands clear of the motor fan end, and stay out of the run-out of a fresh union.

7. Survey the actual floor: run the speed up until each device makes, then back down until each drops out, and record both numbers per device. Accept when every device's make-and-drop pair is on the ticket and every drop-out rpm sits below the speed you intend to program for that mode. Wrong is testing only on the way up: a device that makes at one speed and drops out just under it will cycle all season. Stop rule: if a device's holding rpm demands more flow than the plumbing ceiling allows, do not program around it - find the restriction (filter differential against clean baseline, valve position, water level), re-run the survey, and escalate if the restriction is not on the pad.

8. Program the modes, confirm turnover, and hand over on measured numbers. Set prime speed and duration, filtration speed and schedule, heat speed, cleaning speed and freeze protection, each at or above its surveyed floor with margin. Accept when measured flow at filtration speed completes the pool's volume inside the scheduled run hours, the clamp meter reads at or below nameplate full-load amps at the highest programmed speed, and the heater holds in fifteen minutes at heat speed without cycling its flow switch. Wrong is a schedule that meets turnover only on paper. Leave the written speed card at the pad, show the customer the disconnect and the prime procedure, and register the warranty.

The record this produces

One commissioning block on the work order plus a physical speed card at the pad.

• Supply survey: measured voltage, continuous or clock-switched, GFCI present, control method chosen.
• Flow targets: volume, design turnover hours, required gpm, plumbing ceiling gpm, each device's published minimum with its model.
• Minimum-flow survey: per device, the rpm it made at and the rpm it dropped out at.
• Programmed values: prime speed and duration, filtration speed and hours, heat speed, clean speed, freeze protection.
• Commissioning readings: flow at filtration speed, computed turnover hours, amps at the highest programmed speed against nameplate, heater hold test result.
• Bonding and protection: conductor size, lug condition, continuity result, GFCI test result.

The drop-out rpm is the field nobody writes and everybody needs: the next call about a heater short-cycling is answered in one look against it. The speed card is what survives a customer who "changed something on the panel" - a five minute fix against a written setting, an afternoon without one.

One retrofit, filled in

A 15,000 gallon pool, two inch plumbing, gas heater, salt cell, suction cleaner. Design turnover 8 hours, so required flow is 15,000 divided by 8 divided by 60, about 31 gpm. The plumbing ceiling from the pump SOP's arithmetic for two inch pipe is about 63 gpm, so 31 gpm sits well inside it. Filter clean baseline on the tag is 10 psi.

Steps 1 through 6 pass: 240 V measured, supply continuous from the disconnect with the old clock removed at a prior visit, GFCI confirmed, onboard scheduling chosen, each device's published minimum copied off its own manual, bonding conductor landed and continuous, GFCI tripping and resetting on test, prime held first attempt.

Step 7 fails. The heater's flow switch will not hold in below 2,450 rpm, and at 2,450 rpm measured flow is 39 gpm - inside the plumbing ceiling, but far above the 31 gpm the pool needs. The stop rule says do not program around it. The filter gauge reads 18 psi against its 10 psi clean baseline, 8 psi over, so the filter is cleaned per its own SOP and the survey re-run. On the second survey the heater makes at 1,850 rpm and drops out at 1,780, the cell holds to 1,700, and the cleaner needs 2,300.

Step 8: filtration programmed at 2,000 rpm, measured 32 gpm, turning the pool over in 15,000 divided by 32 divided by 60, about 7.8 hours inside an 8 hour schedule, and sitting 220 rpm above the heater's 1,780 drop-out. Heat speed 2,100, clean speed 2,400, prime 3,000 rpm for two minutes, freeze protection 2,000 rpm.

Against the old single speed unit, which ran near the pipe ceiling at about 63 gpm and turned the pool over in 15,000 divided by 63 divided by 60, about 4.0 hours: 2,000 against a 3,450 rpm rated speed is a ratio of 0.58, and 0.58 cubed is about 0.19, while run time goes from 4.0 to 7.8 hours, a factor of 1.95. Energy per turnover is therefore roughly 0.19 times 1.95, about 0.38, a little under two fifths of the old pump's. That is the cube relationship on a friction-dominated loop, so it is an estimate; what goes on the ticket is the measured 6.2 A at 2,400 rpm against a 9.4 A nameplate, both at 240 V.

References

• 29 CFR 1910.333(b)(2) for electrical work practices, with qualified-person scope at 1910.332 and 1910.399; 1910.147 excludes electrical utilization work at its own (a)(1)(ii)(C).
• NFPA 70E-2021, 120.5, for the live-dead-live proving sequence.
• NEC Article 680, in the edition your jurisdiction has adopted, for pool equipment bonding and GFCI protection.
• The pump, heater, salt cell and cleaner manufacturers' manuals for input configuration, rated speed and each device's published minimum flow.
• See related: the pump and motor replacement SOP for isolation, sizing and bonding; the automation controller SOP for controller-side scheduling; the filter cleaning SOP for the clean baseline.