Why the Same Water Scales One System and Attacks Another

Why this matters

One building, one meter, one supply. The water heater fills with deposit while a branch two rooms away pits from the inside. Both systems are drinking the same water, so "the water is bad" cannot be the finding, and a treatment plan written from the report alone will fix at most one of them.

Deposition and attack are not properties of water. They are surface processes that happen in the thin layer of fluid against the metal, and four operating conditions decide which one runs there: skin temperature, velocity, whether oxygen is being replenished, and how long the water sits. The report is one input. These four are the other four, they belong to the system rather than to the water, and three of them are usually cheaper to change than the chemistry.

Before you go looking

A water-based fire protection system is not a piping system you may open. Taking one out of service is an impairment, and NFPA 25 requires an impairment procedure - notification, tagging, and a plan to restore - before the work starts. Draining or flushing one on your own initiative during a diagnosis is the wrong call regardless of what the water is doing inside it.

Flushing any line that has been standing for years releases what has been growing in it. The route is inhalation for the aerosol and for hydrogen sulfide from sulfate-reducing bacteria in an oxygen-depleted line, and the rotten-egg odor deadens at higher concentrations so your nose is not a detector. Ventilate, and where the space is confined or poorly ventilated use atmospheric monitoring and respiratory protection issued under a written program (29 CFR 1910.134). In a building running a water management program, whether that line is flushed at all is that program's decision (ASHRAE Standard 188).

Do not lower a domestic hot water storage temperature as a scale-control measure on your own judgment. Storage temperature is a Legionella control, and changing it belongs to the building's water management program (ASHRAE Standard 188). Scald risk at the fixture is controlled with thermostatic mixing at or near the outlet, which is what the plumbing code addresses, not by cooling the storage.

Opening any heat-transfer surface to inspect a deposit means isolating it, relieving pressure to a gauge reading zero, and letting it cool before a cover comes off; hot water flashing from a loosened joint scalds through clothing.

The four system variables

Skin temperature, not bulk temperature. Calcium carbonate becomes less soluble as temperature rises, so it comes out of solution at the hottest metal it can find, and that metal is hotter than the water leaving the appliance. A companion article shows the same water moving 1.18 index units between a cold main and a boiler surface; the practical consequence is that a bulk-temperature calculation systematically understates what is happening at the surface. Then the feedback closes: deposit insulates, so the surface behind it runs hotter, so it deposits faster. That is why deposition rates accelerate rather than staying linear.

State the other end, because it is a real system and not a hypothetical: on a surface below bulk temperature, a chilled water coil for instance, carbonate is more soluble than in the bulk, so carbonate scale is not the failure mode there at all. What that surface gets instead is higher dissolved oxygen in colder water, condensation on the outside, and biological growth in the temperature range that favors it.

Velocity, with a ceiling and a floor. Too slow and solids settle, films form unevenly, and stagnant conditions let biology establish under whatever has settled. Too fast and the flow strips the protective film faster than it re-forms, which is a mechanical process rather than a chemical one and is covered by a sibling article on telling erosion from corrosion. Commonly cited working ceilings for copper tube run around 8 feet per second in cold service and 2 to 3 feet per second in hot recirculating service, with the applicable figure set by the tube standard and the design guide your jurisdiction uses rather than by a general rule. Note the hot figure is the lower one: heat and velocity compound.

Oxygen replenishment. In a genuinely closed system the oxygen in the fill is consumed within weeks and never replaced, so general corrosion effectively stops. In anything continuously refreshed - a domestic system, an evaporative system, a closed loop with a leak - the supply never ends. Same water, same metal, two completely different service lives, and the variable is a plumbing condition rather than a chemical one.

Residence time. A dead leg, an idle seasonal branch, an unused fixture riser: the water sits, disinfectant residual decays, biology establishes, and the deposit that forms creates its own local environment underneath. Turnover is a design characteristic. So is surface area per unit volume, which is why a compact high-flux exchanger fouls on a schedule a storage tank on the same water never approaches.

One building, two failures

Municipal supply, hard and well buffered: hardness 224 mg/L as CaCO3, alkalinity 180, chloride 45.

The domestic hot water system is scaling. Heater set at 140 F, saturation index at that surface is positive by a comfortable margin, the water is hard, and throughput is high because it is a commercial kitchen. Every one of the four variables points the same way: high skin temperature, ample throughput delivering mineral mass, and constant replenishment. The deposit is not a surprise; it is the arithmetic.

Run the tempting correction and see how far it gets. Dropping the setpoint from 140 F to 120 F changes only the temperature term of the index, and on this water that term moves the index from plus 0.72 to plus 0.52 - a reduction of 0.20 units that does not cross zero. So the water still deposits, just somewhat slower, and the change has been bought at the cost of a Legionella control that is not the service technician's to trade. That is the honest result: the setpoint is a real variable and on this water it is a weak one, and the correction that actually matters here is on the mineral load, not the temperature.

The recirculation return on the same system is thinning. Same water, same temperature, opposite failure. The recirculation pump was replaced with a larger one during a complaint about wait time at a far fixture, and the return leg now runs well above the 2 to 3 feet per second commonly cited for hot recirculating copper. Elbows and the downstream side of fittings show smooth, directional wear rather than deposit. Nothing about the water changed. One variable did, and it is the one nobody records after a pump swap.

The wet standpipe in the same building is pitting. Same supply, filled once, and then nothing. No heat, so no carbonate deposition. No flow, so no film erosion. The oxygen in the fill water was consumed in the first weeks and never replaced, which stopped general corrosion, and what took over in the oxygen-depleted stagnant water was biological: growth under deposit, localized cells, pinhole-type attack at the tubercles, and air pockets at high points creating a wet-dry interface where the two states meet. The corrosion morphology itself is covered by sibling articles; the point here is that this failure is caused by the absence of the two variables that caused the other two failures.

What flips each verdict. Soften the supply and the heater's problem largely goes away while the standpipe's is untouched, because softening addresses mineral content and the standpipe was never a mineral problem. Put the recirculation pump back to its original size and the return leg stops thinning while the heater keeps depositing. Restore flow to the standpipe on a scheduled basis, which is a fire-protection decision under NFPA 25 rather than a plumbing one, and the residence-time problem changes character while the water stays exactly as it was.

The failure mode this creates: a treatment program gets written for the building because the building has three water problems, and it is written from the report. It addresses the mineral load, which was one of the three. The recirculation leg keeps eroding and the standpipe keeps pitting, both now under a treatment contract, and the shop's own paperwork says the water is being managed.

Which variable to reach for first

Ordered by how often it is the cheapest real fix, not by how often it gets tried:

1. Residence time. Removing a dead leg, restoring turnover on an idle branch, or scheduling movement in a rarely used run costs no chemical and no ongoing program.
2. Velocity. Correcting an oversized recirculation pump or a throttled balancing valve is a one-time change, and both directions are correctable - a line running too slow and one running too fast are the same class of finding.
3. Oxygen replenishment. On a closed system this means finding the makeup, which a companion article covers in detail. It converts an ongoing process into a one-time event.
4. Skin temperature. Genuinely useful where a surface is running hotter than it needs to for a control reason. Almost never available on a domestic hot water system, where the temperature is set by health and code considerations rather than by preference.
5. The water itself. Last on the list because it is the only one that becomes a permanent program with a recurring obligation, and because on many buildings it addresses one of several failures.

How to verify you named the right variable

Compare two systems on the same supply in the same building before you conclude anything about the water. If one is failing and the others are not, the water is a constant and the difference is in the four variables above. That comparison costs nothing and it settles most of these arguments in one visit.

Then check that your explanation predicts the location of the damage, not just its existence. Deposition should be worst at the hottest surface; erosion should be worst at the fittings and direction changes on the high-velocity leg; stagnation damage should be worst at the ends and the high points. If the damage is not where your mechanism says it should be, the mechanism is wrong, however good the chemistry sounded.

References

• NFPA 25, Standard for the Inspection, Testing, and Maintenance of Water-Based Fire Protection Systems, for the impairment procedure required before a fire protection system is taken out of service
• ASHRAE Standard 188, Legionellosis: Risk Management for Building Water Systems, for storage temperature decisions and for whether a stagnant line is flushed
• 29 CFR 1910.134 for respiratory protection where flushing or opening a long-stagnant line generates aerosol or releases hydrogen sulfide
• Copper tube standards and mechanical design guides for the velocity limits applicable to a specific tube, temperature and service, which are not general figures
• See related: What a Scaling Index Tells You and What It Does Not; How to Tell Erosion from Corrosion; Pitting, Crevice and Uniform Corrosion; Why Makeup Water Is the Variable That Decides Everything