The Well Water That Explained the Failing Heater
Why this matters
When the same appliance fails twice at the same address, the appliance is usually not the problem. A plumbing shop replacing a water heater under warranty for the second time is about to install the third heater into the same conditions that killed the first two, and the warranty clock will run out before the water does. This case walks through four explanations for repeated water heater failures in the order the shop tested them, cheapest first, and shows how the answer that finally held was sitting in the customer's well the whole time, in a water supply nobody regulates and nobody had tested.
Before the tank was opened
Every test below that touched the heater started the same way. The breaker feeding the 240 V heater was opened, locked and tagged, and absence of voltage was tested at both element terminals with a meter proven on a known live source before and after, the live-dead-live sequence in NFPA 70E Article 120 (120.5 in the 2021 edition, renumbered in 2024), with work practices under 29 CFR 1910.333(b)(2) for the shop's employees. Before any drain valve was opened, hot water was run at a nearby tub until the outlet water was lukewarm, because a freshly heated tank holds water at scalding temperature, and the drain hose was run to a floor drain and secured so it could not whip. Gloves and eye protection went on before the anode was broken loose, with the tank held against twisting so the breaker bar did not pull the heater off its connections.
The record
The house was rural, on a private well with a submersible pump, a pressure tank and a pressure switch, three people living there. The history the shop put together from its own tickets and the homeowner's memory:
- Heater 1, a 50-gallon electric unit installed by another contractor when the house was bought, leaked from the bottom at about 48 months.
- Heater 2, installed by this shop, lost its lower element at 14 months; the element was replaced. The tank itself leaked at 34 months.
- Heater 3 was about to go in under the tank warranty, from a different manufacturer than heater 2.
A senior tech read that list on the morning of the heater 3 install and asked for an hour before anything went in.
Hypothesis one: a bad part
Two different heaters from two different manufacturers, bought years apart through different supply channels, both failing early, makes a defective batch unlikely. The tech still checked: heater 2's serial against the manufacturer's notices, and the failed element for a manufacturing defect such as a cracked sheath near the flange. The element sheath was pitted along its length and had split near the middle. That is corrosion from the water side, not a fault from the factory. Eliminated, and it pointed somewhere: something in the tank was eating metal.
Hypothesis two: the installation
Installation faults that shorten tank life are well known, and the tech went through each one with the heater out.
Pressure. A gauge with a drag needle on a hose bib read the pressure switch cycling at 40 and 60 psi, normal for a well system. The tech then left the gauge on through a full heating cycle with no water drawn. Because the check valve at the pump keeps heated water from expanding back into the well, the system is closed, and the needle climbed to 76 psi before a faucet was opened. That is under the 150 psi setting of the temperature and pressure relief valve, and repeated swings like it stress a tank, so the shop added a thermal expansion tank to the job, sized to the instructions. But an expansion stress failure shows as a seam or fitting failure and does not explain a pitted element or a tank eaten from inside. Partly confirmed as a contributor, eliminated as the cause.
Connections. Heater 2 had dielectric nipples at both connections, steel tank to copper piping, so the tank was not part of an unprotected galvanic couple at the fittings. The piping was bonded as the electrical code requires and the tech read no voltage between the cold-water pipe and the equipment grounding conductor. Eliminated.
Hypothesis three: how the heater was used
High temperature settings and very heavy demand speed up corrosion and sediment. The thermostats were set to 120 F, the household was three people with ordinary use, and there was no recirculation loop. The tank had very little sediment when drained, which was itself a clue: hard water leaves scale, and this water was leaving almost none. Eliminated.
Hypothesis four: the water
That left the water itself, and the tech had two pieces of field evidence before any lab result.
The anode. Heater 2's magnesium anode, pulled from the drained tank, was consumed down to its steel core wire over almost its full length. An anode is a sacrificial rod: a more reactive metal that corrodes in place of the steel tank, as long as there is anode left to corrode. When the anode is gone, the tank's glass lining protects it only as well as the lining's smallest flaw, and the steel corrodes at every flaw. An anode spent in under three years says the water is unusually aggressive toward metal.
The fixtures. The homeowner mentioned, when asked, that the tub and the toilet tank had blue-green stains under the faucet drip and the fill valve. Blue-green staining is copper being dissolved out of the household piping, a common sign of low-pH water.
The tech asked the homeowner to have the well water tested by a state-certified drinking water laboratory, with a sample collected using the lab's bottles and instructions, and asked for pH, alkalinity, hardness, chloride, total dissolved solids and iron. Strips and a pool kit can show direction; the lab gives numbers that a treatment installer and a health department will accept. The results:
| Test | Result | EPA secondary standard |
|---|---|---|
| pH | 6.0 | 6.5 to 8.5 |
| Alkalinity | 14 mg/L as CaCO3 | none set |
| Hardness | 22 mg/L as CaCO3 | none set |
| Chloride | 180 mg/L | 250 mg/L |
| Total dissolved solids | 420 mg/L | 500 mg/L |
| Iron | under detection | 0.3 mg/L |
Read together, these describe soft, acidic water with little buffering, which is aggressive water: low pH dissolves metal, low alkalinity means nothing in the water resists the pH, and low hardness means no protective scale forms on metal surfaces. Elevated chloride and dissolved solids raise the water's conductivity, which speeds the electrochemical reaction that consumes the anode. Only the pH sat outside the EPA's secondary range (40 CFR Part 143), and the secondary standards are non-enforceable guidelines for taste, appearance and corrosivity, not health limits; but corrosion is the result of the combination, not of any one line. The table's own columns are a warning about reading one number at a time.
What the shop said, and what it did not
The shop is not a health authority, and the conversation stayed on what the shop could stand behind. Private wells are outside the federal Safe Drinking Water Act; that law covers public water systems, so the owner of a private well is the one responsible for testing and treating it, with state and county health departments setting recommendations and, in some states, testing requirements at the sale of a house or for a new well. The tech said plainly that acidic water can dissolve lead and copper from household plumbing and fixtures, that the shop could not say whether the water was safe to drink, and that a lab test for lead and copper at the tap, with the county health department's guidance on interpreting it, was the next step for that question. The homeowner ordered it.
The fix, and how they confirmed it
The repair had three parts.
- Treat the water at the source. An acid neutralizing filter, a tank of calcium carbonate media that dissolves into the water and raises its pH and alkalinity, was installed after the pressure tank so every fixture got treated water. The installer set it up to the media maker's flow and backwash requirements for the household's peak demand, and noted that neutralizing adds some hardness, a deliberate tradeoff. The well pump's breaker was locked off and the pressure tank drained to zero on its gauge before the line was cut, so no stored pressure or pump start met an open pipe.
- Install heater 3 with the expansion tank and a plan for the anode. The new heater went in with the expansion tank, its anode was photographed and measured before installation, and an anode check was put on the shop's schedule at 12 months.
- Restart in the right order. After heater 3 was connected, the cold supply was opened and a hot faucet was run until it delivered a steady stream with no sputtering air, so the tank was full before any power reached it; energizing an electric element in an empty tank burns it out in seconds. Then the breaker was closed with the tech standing to the side of the panel, the element current was read with a clamp meter on the supply conductors with the cover in place, and the relief valve was checked for leaks after the first heating cycle.
Confirmation came in two steps. Thirty days after the neutralizer went in, a follow-up lab sample from the kitchen tap read pH 7.2, inside the secondary range, and the blue-green staining stopped spreading. At the 12-month anode check, heater 3's anode showed some wear along its upper third and most of its diameter left. That is not a like-for-like comparison with heater 2, which was pulled at 34 months on untreated water from a different maker's tank, but a rod barely touched at 12 months beside one spent at 34 is the direction the fix predicted, and the anode is now checked each year.
What would have flipped the diagnosis
Hard water would have changed everything. A tank failing early on very hard water usually fails from scale and sediment, an element that burns out under a crust, a rumbling tank, and the fix there is softening or regular flushing, which on this soft water would have made corrosion worse.
A sulfur smell from the hot water only would have pointed at the anode reacting with sulfate-reducing bacteria, a different problem with a different fix, usually an anode of a different alloy and a chlorination treatment.
Normal water and a spent anode would have sent the shop back to stray current and bonding, because something other than the water would have been consuming the rod.
References
- U.S. EPA, National Secondary Drinking Water Regulations, 40 CFR Part 143, for the secondary ranges in the table; and EPA guidance on private well owner responsibility under the Safe Drinking Water Act's scope.
- The state's list of certified drinking water laboratories, and the county or state health department's private well testing guidance.
- The water heater manufacturer's installation instructions and warranty, including any water-quality conditions they state, and the treatment media maker's sizing requirements.
- NFPA 70E Article 120 (120.5 in the 2021 edition, renumbered in 2024), and OSHA 29 CFR 1910.333(b)(2) for electrical isolation.
- See related: Private Wells and the Water Nobody Tests; Hardness, Alkalinity and pH and What Each One Does; How to Check Water Hardness on a Service Call.