The $2,300 Burnham Callback That Made Me Rewrite Our Boiler-Sizing Rulebook

March 2022: The Callback That Wouldn't Quit

In March 2022, I stood in the basement of a 14-unit apartment building in Albany, New York, listening to a Burnham Series 2 commercial boiler short-cycle for the fifth time that hour. The building manager was behind me with his arms crossed. He'd called us back three times in three weeks. I'd been managing boiler service operations for nearly a decade at that point. I knew this equipment. Or at least, I thought I did.

My name is on a lot of invoices, and so are my mistakes. I've documented 14 of them over 11 years in this role—roughly $18,400 in wasted labor, unnecessary parts, and callbacks combined. This callback was number nine, and it cost about $2,300. That's the one that finally made me rewrite our pre-installation checklist from scratch.

How the Job Started

The building had an old atmospheric natural gas boiler—I won't name the brand, because I want to keep this focused on my mistake, not on the equipment. If I remember correctly, the original unit was rated at 210 MBH input, had been in service since roughly 2004, and was leaking from the block. The property manager got three bids. We won on price and what I thought was a solid proposal.

Here's the thing: I sized the replacement Burnham the way I'd been sizing boilers since I got into this business in 2013. Square footage, rough insulation assumptions, a little buffer. The old unit was 210 MBH, the building was 9,600 square feet, so I spec'd a Burnham Series 2 at 199 MBH. Made sense. Looked responsible. Matched the old one almost exactly.

Install went clean. Four days, no surprises. The crew tested it, it fired, it heated, we got paid. I moved on.

The Callback

Three weeks later, the phone rang. Short-cycling during mild weather—unit would fire for four or five minutes, shut off, then refire a few minutes later. Classic symptom. My tech went out, checked the obvious: flame sensor, thermocouple, high-limit, gas pressure, circulator. All looked fine. He replaced the flame sensor anyway. Two weeks later, same call.

I went out myself the second time. Honestly, I was annoyed. I figured the tenant was messing with the thermostat, or the building manager had the setpoint cranked, or there was air in the loop. I checked all three. None of those things.

Third call, I sat down in the mechanical room for an hour just watching it run. That's when I noticed the return water temperature was climbing much faster than I expected. The boiler was reaching its high limit, shutting down, then waiting for the return water to cool before it fired again. It wasn't a flame problem. It wasn't a control problem. It was a load problem—or rather, a lack of one.

The Surprise

Never expected the building to be that tight. Turns out the property owner had done a full weatherization retrofit in 2019—new windows, air sealing, added attic insulation, the whole thing. The actual heat load had dropped to somewhere around 110–120 MBH. We'd installed a 199 MBH boiler into a system that needed less than two-thirds of that.

The surprise wasn't that the boiler was oversized. It was that nobody calculated the actual load. Me included. I'd relied on the old nameplate and a square-footage rule of thumb, exactly the way I was trained.

I called Burnham technical support at that point, mostly to confirm I wasn't missing something obvious in the controls. They were helpful—and patient—but the conversation kept circling back to one question: "What's your actual load?" Which was fair. I didn't have a good answer.

"What was best practice in 2015 may not be accurate in 2025."

I should add that this building was a better candidate for a modulating condensing boiler—something with a turndown ratio that could match the real load. Instead, we'd installed a two-stage cast iron unit because that's what the old one looked like. The Burnham Series 2 is a solid boiler. Cast iron sections, well-tested controls, and when it's sized right, it'll run 20–25 years. But we didn't give it a fair shot. We gave it a load it couldn't modulate down to.

What It Cost

Here's the breakdown, roughly, if I remember the invoice correctly:

  • Three service calls: about $720 in truck roll and diagnostic time
  • Unnecessary flame sensor and thermocouple replacement: $280
  • Return trip to install a buffer tank and re-pipe for primary/secondary: $1,100
  • Credited labor for the callback visits: $200

Total: roughly $2,300. Not catastrophic. But the credibility hit was worse. A property manager who gave us three buildings is now down to one. He told me later—direct quote—"I don't need a boiler that's bigger than my whole heating load." Ouch. Fair.

When a customer types "burnham boiler service near me" three weeks after an install, that's not a compliment. That's a person who's starting to wonder if they picked the wrong contractor.

Part of Me Wishes We'd Just Done the Math

I have mixed feelings about how this played out. On one hand, the mistake forced us to update our process, and that process has caught real issues since. On the other hand, I basically wasted $2,300 and a client relationship learning something that a $200 load-calculation subscription and the Burnham installation manual could have told me upfront.

Part of me is grateful this happened on a 14-unit building and not a 140-unit commercial project. Another part knows that "grateful for the small version of the mistake" isn't a real strategy.

The Checklist That Came Out of It

After the third callback, I sat down and wrote out our current pre-installation checklist. It's five items. It's boring. It works.

  1. Actual heat-load calculation on every replacement. Not square footage. Not "match the old nameplate." A real calculation—Manual J or equivalent—regardless of building size.
  2. Document any envelope changes. Windows replaced, insulation added, air sealing—ask the property manager directly and write it down. If they don't know, pull the last three years of utility bills and look at the trend.
  3. Verify distribution-system capacity. Old piping and old terminal units don't always work well at the lower return temperatures modern boilers prefer.
  4. Confirm the venting path. Existing chimneys don't always handle new boiler venting requirements, especially if the replacement has different flue characteristics than the unit it's replacing.
  5. Match turndown to the load range. If the load is going to swing (shoulder seasons, partial occupancy, that sort of thing), document the range and pick equipment that can modulate down to it.

We've caught 47 potential issues with this checklist in the past 18 months. Some were minor. A few were the same class of mistake I made in Albany—wrong-sized equipment, wrong assumptions about the building envelope, missing information about prior retrofits.

The industry's moved on from rule-of-thumb sizing. AHRI and ASHRAE standards have been pushing for actual load-based selection for years, and DOE efficiency benchmarks are steadily making oversized equipment both harder to justify and more expensive to operate. What worked in 2015 is basically a liability in 2025. The fundamentals of boiler service haven't changed—water, heat, venting, controls. But the execution has. I just needed a $2,300 callback to catch up with the execution part.

If you're running boiler installs and you're still sizing by nameplate, do yourself a favor. Do the load calc first. Trust me on this one—take it from someone who didn't.

author-avatar
Elisa Nordberg

Elisa Nordberg writes about air-cooled and water-cooled industrial chillers, modular glycol systems, and screw, scroll, and centrifugal configurations for process and comfort cooling. Her evaluations reference ISO 5149 and AHRI 550/590 practices while comparing cooling capacity, COP, IPLV, compressor lift, fluid flow, and evaporator approach temperature. She helps plant engineers and sourcing teams size dependable chiller packages, interpret part-load performance, and balance energy use, redundancy, maintenance access, and lifecycle cost.

Leave a Reply

Your email address will not be published. Required fields are marked