Your Water Heater Didn't Fail — Your Specification Did

The equipment is usually the last thing that breaks

I'm a quality and brand compliance manager at a plumbing and heating wholesaler. I review every submittal and warranty package before it leaves the building — roughly 200 of them a quarter — and I rejected about 31% of first-pass submittals in 2024. Fewer than one in ten of those rejections was for an actual defect. The rest were spec problems: wrong circuit, wrong water, wrong model for the application, or a document nobody read until something stopped working.

So here's my position, and I'll take the heat for it: most of what this industry calls an equipment failure is a decision that was made wrong months before the box was opened. The model number is usually the least useful clue on site.

I've been at this long enough to notice the pattern repeats at every scale — from a 36 kW tankless heater down to a twelve-dollar furnace filter. Same mistake, very different price tag.

Argument one: "bigger is safer" is the most expensive sentence in the trades

It's tempting to think the biggest model in the line is the safe choice. But oversizing is the most common specification error I see, and it fails in the most expensive place — not at the unit, at the service.

Take the Stiebel Eltron Tempra 36 Plus tankless water heater. Thirty-six kilowatts at 240 volts works out to roughly 150 amps of connected load. Split across the two circuits the unit actually uses, that's still about 75 amps per circuit. The NEC treats a fixed electric water heater as a continuous load, so the branch circuits get sized at 125% of the nameplate — which means the panel has to carry a load a lot of houses have never seen. I want to say it lands on a pair of two-pole breakers with conductors you don't buy at the hardware store, but don't quote me on the wire gauge. That's the electrician's calculation, not mine.

The amp math isn't the point. The point is that a 100-amp service — still common in housing built before the 1990s — can't carry it, no matter how well the unit is engineered. Good engineering doesn't fix a 100-amp panel. Nothing fixes a 100-amp panel except a service upgrade.

In my first year on this desk, I made the classic submittal error: I approved a package because the model number matched the order and nothing else looked obviously off. It matched. It was also headed into a house with a 100-amp service and a 60-amp feeder to the garage. The contractor ate the return freight, I ate the credibility, and we've required a load calculation with every tankless submittal since.

The flip side is the version I see most weeks. A homeowner saved about $700 by staying with a smaller unit to dodge an electrical upgrade, then paid something like $2,400 eighteen months later when they added an EV charger and a heat pump and had to upgrade the service anyway. Doing the service calculation once, at the start, is basically free. Doing it three times is not.

Argument two: search "hot water repairs Adelaide" and you'll usually find the wrong problem

When someone in Adelaide searches for Stiebel Eltron hot water repairs, the assumption baked into that search is that the unit failed. In the claim packages I actually see, that's rarely the whole story. The failures cluster around site conditions, not the heater:

• Scale and water chemistry. Any electric water heater that heats water directly is a scale surface. Adelaide's supply runs on the harder side most years, and hard water plus a heating element plus no service interval is a maintenance problem that eventually presents as a hardware problem.
• Delivery temperature. AS/NZS 3500.4 governs hot water supply and delivery temperature, and in a domestic setting that generally means no more than 50 °C at the sanitary outlets. A tempering valve is doing part of that work. A stuck or undersized valve changes how hard the unit runs and how long it runs before it satisfies.
• Voltage and terminations. An instantaneous unit is an enormous resistive load. Loose terminations and undersized cable show up as "the unit keeps tripping," which gets recorded as a unit fault.

Then there's a naming problem that costs people real money. The Tempra line is a North American product — different model numbering, different components, different spares channel. The Australian range carries its own model numbers and its own assumptions. So a search for a Tempra repair in Adelaide tends to end the same way: someone orders a part that doesn't fit a unit they don't own. Same company. Different hemisphere. Different parts bin.

I want to say we logged fourteen of those mis-ordered-parts cases in a single quarter, though I might be misremembering the exact number. It was enough that we started asking for a photo of the rating plate before quoting anything.

To be fair, some of these are genuine defects — control boards, flow sensors, elements. Every brand has them; nobody's immune. But when I sort our claim log by root cause, site conditions and misdiagnosis outnumber manufacturing defects by a wide margin. Which is also why I don't get defensive when someone tells me German-engineered equipment failed on them. It didn't fail. It did exactly what the spec told it to do, in conditions the spec never accounted for.

Argument three: the filter arrow and the leaf blower are the same mistake at one-hundredth the scale

Which way to put air filter in furnace — that's one of the most-asked questions in the trades, and the answer is genuinely simple. The arrow points in the direction of airflow, which is toward the blower. On a standard upflow furnace, the arrow points at the cabinet, away from the return duct. If the filter has an arrow, that arrow is not decorative.

But here's the blind spot. Most people ask which way the arrow goes and never ask whether the filter actually seals the cabinet. A one-inch filter sitting loose in a four-inch rack with a half-inch gap around the frame isn't filtering much — air takes the path of least resistance and goes around the media, not through it. The arrow is the easy part. The bypass gap is the part that decides whether your blower motor and coil breathe grit for the next five years.

Same story outside. An outdoor fan that stops moving air gets replaced as a motor when it's more often a run capacitor or a blade that went back on upside down after a motor swap. Fan blades aren't symmetric. Flip one and it still spins — just with noticeably less airflow, and nobody notices until the compressor is working harder on the hottest day of the year.

Then there's the DeWalt blower. Honestly, it's the right tool for dry leaves and grass clippings on a condenser, better than a hose for loose debris. But blow through the coil from the inside out, and don't hold the nozzle six inches from the fins. Fold those fins over and you've turned a ten-minute job into a coil replacement — a much bigger bill than the blower cost. Saved a service call, spent a coil. That's the version I see most.

The pushback I get

"I'm not a contractor. Why should any of this matter to me?" Because you're the one paying for it, and the decision that gets made in fifteen minutes at the quote stage is the one that costs five figures to unwind. Informed customers ask better questions and get to a decision faster. I'd rather spend ten minutes explaining options upfront than spend fourteen months arguing about a warranty claim neither of us can win.

"Isn't this over-engineering a house?" Granted, some of it is. If you change a filter every three months and nothing has ever failed, you don't need to audit your ductwork. But over-thinking a load calculation is cheap. Under-sizing a service entrance is not.

"Everybody just wants the best price." I get it — budgets are real and I'm not going to pretend otherwise. But the lowest quote and the lowest total cost are rarely the same number. Total cost includes the return freight, the redo, the service upgrade you paid for twice, and the weekend you spent waiting on a part. The cheapest quote is usually just the one that hasn't finished being paid for.

So here's where I land

I don't think this industry has an equipment quality problem. I think it has a specification problem — and the equipment takes the blame because the equipment is what's visible when the water goes cold.

Read the spec sheet. Do the load calculation before you buy, not after. Figure out which way the arrow goes, then check whether the filter actually seals. Blow the condenser coil out from the inside. And if you're in Adelaide searching for a Tempra, look at the rating plate first — you're probably chasing a model number that was never sold in this country.

Fifteen minutes of reading beats fourteen months of arguing. Every time.