It was 7:40 on a Monday morning in January 2024, and a whole wing of our office was cold. The screen on our 11-year-old Vaillant combi boiler was flashing F22—Vaillant's code for low water pressure.

I'm the office administrator for a 65-person logistics company. I manage roughly $180,000 a year in purchasing across 17 vendors, and heating repairs fall under my job whether I like it or not. This story is about a decision I made back then. It was almost a very expensive one.

I knew F22 meant the system pressure had dropped below the minimum, so I topped up the water with the boiler's filling valve, reset it, and the heat came back.

For about 36 hours.

Then it dropped out again. So I did what any admin buyer would do: I called three HVAC companies and asked for quotes.

  • A one-man operation: $190 to replace the Vaillant water pressure sensor, same-day.
  • A manufacturer-certified installer: $380, but only if I let him run diagnostics first.
  • A third company: $450, same diagnostic-first approach.

A pressure sensor is a $40 part. The boiler is supposed to be solid German engineering. My purchase log tracks every dollar, and I answer to finance. On paper, the $190 quote was the obvious choice.

I approved it. That's where this story starts.

Why the Cheap Quote Wasn't Cheap

The tech was friendly, on time, and swapped the sensor in about 20 minutes. The system held pressure for eleven days.

Not even two weeks.

To be fair, the $190 guy was fast and available same-day. He just wasn't equipped to find the actual fault. When it dropped out again, he came back—the second trip was $150, as best I recall—frowned at the boiler, and started talking vaguely about the unit "getting old." That's when I did what I should have done first: I paid the manufacturer-certified installer $150 for a diagnostic.

He found the real fault in 40 minutes. The expansion vessel had lost its air charge. In a sealed heating system, water expands when it heats up, and the expansion vessel absorbs that extra volume. A waterlogged vessel makes pressure spike every time the boiler fires, the pressure relief valve opens, and water dribbles down the drain. When the system cools, pressure is low again. The cycle repeats.

The old sensor wasn't lying. It was reporting exactly what was happening. The problem was the relief valve, not the sensor. We replaced the messenger and ignored the message.

The Problem With Treating a Boiler Like a Lightbulb

It's tempting to think: part fails, replace part. But a boiler is a closed system. By the time you see low pressure, the question isn't "which part is cheapest?" It's "why is the pressure dropping?" The cause might be the expansion vessel, a weeping relief valve, a slow radiator leak, or simply years of nobody checking the system pressure. One of those is a $900 repair. Another is a $3,800 repair. And one is free—a scheduled top-up and a watchful eye. You don't know which you're facing until someone actually looks. No phone quote will tell you.

The Second Cause Was My Own Purchase Log

Let me be honest about the uncomfortable part. My buying system rewarded the wrong thing.

When I took over purchasing in 2020, I built a vendor log that tracks unit price, lead time, and payment terms. Nothing in it tracks what happens after the invoice is paid. There's no column for callback trips, for the days the office was cold, or for the staff who started working from home because the building was miserable.

On paper, the $190 quote was a win. In reality, it had already started costing more than the $450 quote would have.

The Kerosene-Heater Detour

While we waited, someone bought a kerosene heater. It cost $160, plus $42 each time we refilled it—or rather, each time our office manager refilled it, because I refused to touch the thing. It smelled, it was underpowered for our open-plan layout, and the safety sheet alone made me want to return it. We still spent about $300 on kerosene over nine days—money that bought comfort, not a fix.

I can't prove this with a statistic, but after five years of managing building maintenance, my sense is nearly all of us fall into the temporary-heater trap at some point. Temporary solutions feel cheap because the cost arrives one refill at a time.

Counting What the "Bargain" Actually Cost

Here are the real numbers, in order:

  • $190 — the sensor replacement that didn't fix anything
  • $150 — the second visit from the same company
  • $150 — the diagnostic I should have bought first
  • $900 — expansion vessel, pressure relief valve, labor, and annual service
  • $460 — kerosene heater and fuel

Total: $1,850, plus ten days without reliable heat in that wing. The diagnostic-first route would have been $1,050, done in three days. The "cheap" path cost 76% more and took three times as long.

But it gets worse. Because the system kept running under-filled for a week, the primary heat exchanger was damaged. The installer caught it before total failure, but his warning was blunt:

"If you'd run it like this another couple of weeks, you'd be looking at replacing the whole boiler."

On an 11-year-old Vaillant, that meant a $3,800 quote. I was one approval away from ordering a replacement from an online supply house when the installer called back and said he'd found the real fault. I canceled the order hours before it went through.

Dodged a bullet. A $3,800 bullet.

Once I started looking, I saw the same pattern in unrelated purchases. I bought a $99 backpack leaf blower for our grounds contractor to trim costs. The housing cracked on the second use. The $320 replacement has now run for two full seasons. Same math: the lowest price on the day is rarely the lowest total cost. The proof usually sits in the second purchase you have to make.

What I Do Differently Now

The actual repair was unglamorous: a new expansion vessel, a new pressure relief valve, and a full service. Total: $900. We haven't touched the boiler pressure since.

I can't give you a 47-point procurement manual. But three rules have genuinely changed how I buy heating repairs—and, honestly, most other things.

1. Pay for diagnosis before you pay for parts.

For any error code, the smart purchase is a diagnostic check, not a part. A good one costs $150-250 and tells you whether you have a $190 problem, a $900 problem, or a $3,800 problem. Guessing at parts means you might buy all three.

2. Put a hard deadline on temporary fixes.

The kerosene heater was supposed to last two days. It lasted nine. If a stopgap runs longer than a week, it's not a stopgap anymore—it's a decision you're avoiding. Temporary solutions need an expiration date, or they'll eat your budget one refill at a time.

3. Learn the equipment you already own before buying new.

While the boiler saga was unfolding, I discovered our Honeywell thermostat was still running the previous tenant's schedule, and employees were overriding it daily because they were cold. I spent ten minutes looking up "honeywell thermostat how to use" and reprogrammed it. The U.S. Department of Energy estimates that consistent thermostat setbacks can cut heating costs by about 10% (Source: energy.gov, accessed January 2025). Our gas bill dropped roughly 8% the following month. The thermostat taught me the same lesson as the boiler: the cheapest fix is sometimes the one you already paid for.

The Bottom Line

The $190 sensor replacement ended up costing $800 more than the diagnostic-first route, plus a week of cold offices, kerosene fumes, and avoidable remote-work days. A wrong purchasing decision never arrives as a single line item. It arrives as a second callout, a part replaced twice, a temporary heater, and a near-miss on a $3,800 boiler replacement.

I don't have hard data on how many boiler faults get misdiagnosed because the buyer insisted on a part quote instead of a diagnosis. Anecdotally, from five years and too many vendors to count, the lowest quote has ended up being the most expensive option roughly six times out of ten.

When your Vaillant combi boiler throws an F22—or any equipment gives you an error code—don't ask who will replace the part for the least money. Ask who can tell you why the part failed, and what else is about to fail.

The part is cheap. The cause is where the real invoice was all along.