Salt Cell Inspection and Cleaning
Purpose
A salt cell gets replaced more often than it fails. The controller says low salt, somebody adds salt, the pool still will not hold chlorine, and a new cell goes in on a reading that was never a salt reading in the first place. This procedure guarantees that the cell is judged on an independent water test and on its own electrical behaviour rather than on the panel display, that it is cleaned in an order that does not strip its plate coating, and that it never energizes without proved flow. The last one is the safety half: a cell electrolysing water in a line with no flow accumulates hydrogen in the plumbing, and hydrogen in a closed pipe is an explosion, not a nuisance.
Scope
Covers inspection, cleaning, electrical judgement and end-of-life decision for an inline salt chlorine generator cell and its flow switch on a residential or light commercial system, plus the salt addition arithmetic that follows from an independent test.
It does not cover controller commissioning, relay assignment or remote setup, which the automation controller SOP owns. It does not set acid dilution ratios, which belong to the muriatic acid SOP, and it does not cover shocking a pool: a chlorinator's boost or superchlorinate mode raises output over days and cannot reach breakpoint, which is the shock dosing SOP's job.
Roles and responsibilities
| Role | Owns | Hands off |
|---|---|---|
| Route technician | Independent salt test, cell inspection and cleaning, flow switch verification, hours reading | Records the independent salt value beside the controller's displayed value, so the next visit can see the gap |
| Shop lead | Cell and flow switch stock, the replace-or-clean call, warranty claim on a cell inside term | Approves a cell replacement only against a cleaned-cell reading, never against a panel message |
| Customer | Runs the pump for the programmed hours, reports panel alarms rather than clearing them | Hears the chemistry cause behind a scaled cell, not just the cleaning charge |
What the cell is actually telling you
Output percent is a duty cycle, not a chlorine level. A cell at 100 percent is on all the time the pump runs; if the pump runs four hours, 100 percent of four hours may still be less chlorine than the pool consumes in a day. Runtime and output are two separate levers and the panel only shows one.
The displayed salinity is a derived number. Most controllers infer salt from the conductivity of the water between the plates, corrected for temperature. That means three different faults produce the same "low salt" message: genuinely low salt, cold water, and a scaled cell whose plates are insulated by calcium carbonate. Below the cell's stated cold-water cutoff, commonly in the 50 to 60 F range depending on make, the unit stops producing altogether and will report it as a fault.
So the rule that governs this whole procedure: no salt is added on a panel reading. Confirm with an independent test, a drop titration or a handheld meter, and compare against the salinity band on the cell's own plate or manual, which differs by make. Where salt genuinely is short, the arithmetic is straightforward: raising 1,000 gallons by 1,000 ppm takes about 8.3 lb of salt, because 1,000 gallons of water weighs about 8,345 lb and a part per million is a millionth of it.
Plate condition is the end-of-life test, not age. A cell whose plates are still evenly dark and which returns to its expected amperage after cleaning has life in it. A cell with pale, thinned or flaking plates is finished no matter what the hour counter says, and cleaning one more time buys weeks.
Procedure
Kill power to the chlorinator at its breaker or disconnect and confirm the cell is de-energized before you touch a union. Accept when the controller display is dark or the cell circuit reads no output on the panel diagnostic. Wrong is trusting an off button on the front of a controller, which is a control input rather than an isolating means. A cell that stays energized while flow stops is generating hydrogen inside the plumbing; that gas is what makes an opened union at a live cell dangerous rather than merely wet.
Stop the pump, open the filter air relief until it stops discharging, and only then break the cell unions. Accept when the filter gauge reads zero with the relief still open and the cell body is at atmospheric pressure. Wrong is cracking a union under head, which sprays chlorinated water at face height and can carry accumulated gas with it. Stand to the side of the union as it comes apart and let the line drain before lifting the cell out.
Remove the cell and inspect the plates in daylight before cleaning anything. Look down the length of the cell at the plate faces and the gaps between them. Accept when plates are evenly dark, flat, and separated by clear gaps. Wrong is white crust bridging the plates, which is scale, or pale patches and flaking, which is coating loss. Photograph what you find: the difference between "scaled" and "worn" is the difference between a cleaning and a replacement, and the customer is entitled to see it.
Test salinity independently before you look at the controller number again. Drop titration or a calibrated handheld meter, on a sample taken away from the returns, with water temperature noted. Accept when the independent value sits inside the band the cell's plate or manual states for that make. Wrong is adding salt because the panel said low: if the independent test is in band and the panel is not, the fault is scale or temperature, and salt added on top of an in-band pool is only removable by replacing water.
Flush and mechanically remove what will come off before any acid touches the cell. Hose the cell out end to end and work loose deposits with a plastic tool or a wooden stick. Accept when the water runs clear and the remaining deposit is a hard film rather than loose flakes. Wrong is a screwdriver, a wire brush or a pressure washer between the plates: the plate coating is thin and every scratch is permanent capacity loss. Nothing metal ever goes between those plates.
Acid soak only if scale remains, in the manufacturer's ratio, in a cell stand, timed. Mix in the marked acid vessel with water in first, then acid, at the ratio the cell manual states, commonly around four or five parts water to one part acid. Accept when the fizzing stops and the plates are clear, within the soak time the manual allows. Wrong is a longer soak "to be sure": acid attacks the plate coating once the scale is gone, so an overnight soak is how a good cell is destroyed by a technician trying to help. Goggles and acid gloves stay on, and the soak happens outdoors, never in a closed pump house.
Replace both cell union o-rings and reinstall with the flow direction and orientation the manual specifies. Accept when o-rings are new or undamaged, silicone lubricated, and the cell sits in the orientation the manual shows with its arrow following flow. Wrong is a reversed cell or a cell mounted where gas can pocket in the housing. Hand tighten the union collars only; a wrench on a plastic collar cracks it and the crack shows up as a leak two weeks later.
Restore flow first, then power, and confirm the cell energizes only when the flow switch has closed. Accept when the cell shows no output with the pump off and comes up to output within seconds of flow being established. Wrong is a cell that produces with the pump stopped, or a flow switch that will not close with flow verified good. A flow switch with proved flow and a physically broken or stuck paddle is genuinely failed and gets replaced with the correct part; it never gets jumpered, taped closed or wired out, because it is the only thing preventing hydrogen accumulation in a dead line.
Read the cell's diagnostic, set output to the pool's demand, and fix the chemistry that scaled it. Compare the controller's cell voltage and amperage, or its equivalent diagnostic, against the range in the manual, and log the cell hours. Accept when the diagnostic sits in the manual's range and displayed salinity now agrees with the independent test within the tolerance the manual gives. Wrong is leaving a cleaned cell on the same water that scaled it: high pH and high calcium hardness will rebuild the deposit in weeks, so correct pH per the acid SOP and quote hardness dilution where it is the driver.
The record this produces
One cell block on the ticket, filed to the equipment record for that chlorinator.
- Both salinity numbers: the controller's displayed value and the independent test, with water temperature beside them.
- Plate condition in words and photographs: scaled, clean, pale, flaking.
- Cleaning performed: mechanical only, or acid at what ratio for what soak time.
- Flow switch result: closes with flow, holds open without it, or replaced with the part number.
- Cell diagnostic and hours against the manual's range, and the output percent left set.
- Chemistry cause and action: the pH, calcium hardness and saturation condition that produced the scale, and what was corrected or quoted.
The two salinity numbers side by side are the whole point of the record. One visit's gap is a scaled cell; the same gap reappearing three months later on a cleaned cell is a cell losing its plates, and that is a warranty conversation while the cell is still in term. The shop lead should refuse any cell replacement request that does not carry a post-cleaning diagnostic reading, because the panel message alone has never been sufficient evidence.
One cell service, filled in
A 20,000 gallon pool, controller displaying a low salt warning, output already at 100 percent, and the pool testing 0.5 ppm free chlorine. Water 62 F in early spring.
Steps 1 and 2 pass: chlorinator breaker off, display dark, filter relief vented to zero, unions broken from the side with the line drained. Step 3 finds heavy white bridging between the plates over roughly the middle third of the cell, plates dark and flat where visible, no flaking. Photographed.
Step 4 is where the visit turns. Independent drop test reads 3,100 ppm against the controller's displayed 2,400, and the cell's plate states a band with 3,200 as target. The stop rule is explicit: no salt is added on a panel reading, and the independent value is in band. Had the technician dosed to the displayed number, the arithmetic would have been 20,000 gallons times 800 ppm at 8.3 lb per 1,000 gallons per 1,000 ppm, about 133 lb of salt into a pool that needed none, and the only correction for that is replacing water.
Steps 5 and 6: flushed, loose deposit worked off with a plastic tool, then a timed soak at the manual's ratio in the marked acid vessel, water in first. Fizzing stopped inside the manual's stated window and the cell came out with clear gaps.
Step 8 fails. Flow restored, power on, and the cell does not come up to output. Flow verified good at the returns and at the filter gauge, and the flow switch paddle is found snapped at the hinge, lying loose in the housing. This is the opposite case from a heater limit that opens for a reason upstream: here flow is proved good and the device is physically broken, so it is genuinely failed. Replaced with the correct part. It is not jumpered for the afternoon, because a jumpered flow switch on a chlorinator lets the cell energize in a dead line and make hydrogen.
Step 9: with the new switch, the cell energizes within seconds of flow and its diagnostic reads inside the manual's range. Displayed salinity now reads 3,150 against the independent 3,100, which confirms the scale was the false reading rather than the salt. Cell hours logged. Water chemistry: pH 8.0 and calcium hardness 420 ppm, which is the scale story, so acid is dosed per the acid SOP and partial dilution for hardness is quoted rather than left for the cell to keep absorbing.
References
- Cell and controller manufacturer's manual for the salinity band, cold-water cutoff, acid ratio, soak time, orientation and diagnostic ranges. These differ between makes and none of them travel.
- 29 CFR 1910.1200 for labeling and safety data sheet access on the acid used in the soak.
- See related: the muriatic acid SOP for dilution and pour order; the shock dosing SOP for why boost mode is not a shock; the automation controller SOP for controller setup and handover.