An owner calls: their 7.5-ton rooftop unit is down again, the repair quote is $3,800, and they want to know if they should fix it or replace it. The lazy answer is a rule — "$5,000 rule," "half the replacement cost," pick your shortcut. The honest answer is five questions. Here's how I actually work it.
1. How old is it, really?
Packaged rooftop units run 15 to 20 years. The back half of that life is where reliability falls off and parts get scarce. My rough framing:
- Under 10 years: Repair, almost always. It's got life left and the major components are still supported.
- 10–15 years: The gray zone. Now the size of the repair and the energy numbers decide it.
- Over 15–18 years: Lean hard toward replacement. You're putting money into a unit that's living on borrowed time, and the next failure is around the corner.
2. What's the refrigerant?
This one ends a lot of conversations by itself. If the unit runs R-22, it's old enough that R-22 has been phased out of production. The refrigerant is expensive and getting more so, and a unit with a refrigerant leak in a sealed system you can't affordably recharge is effectively done. A compressor or coil failure on an R-22 unit almost always tips to replacement — not because the repair is impossible, but because you'd be sinking real money into a dead-end platform.
"It's an R-22 unit with a refrigerant leak" is, nine times out of ten, the whole answer. Replace it.
3. What does the repair buy you?
Not the dollar amount in isolation — the dollar amount relative to a new unit and relative to how much runway it buys. A $3,800 repair on a unit that's good for another eight years is money well spent. The same $3,800 on a 17-year-old R-22 unit buys you maybe a season before the next failure. Same number, completely different decision.
I look at the repair as a percentage of installed replacement cost. Installed, a packaged RTU swap runs roughly $1,800–$2,600 per ton depending on the market, the unit, and the access. If the repair is climbing past 40–45% of that replacement number on an aging unit, repair stops making sense.
4. What's it costing to run?
An old unit isn't just unreliable — it's inefficient. Compressors and motors degrade, coils foul, and the efficiency the unit shipped with 15 years ago was lower than today's minimums to begin with. A new high-efficiency unit can cut the energy attributable to that equipment meaningfully. When I run the numbers I'll assume something like an 18% energy savings as a planning figure, more on a really old unit.
That savings turns into a simple payback: replacement cost divided by annual savings. If the new unit pays for a chunk of itself in energy over its life, replacement gets a lot easier to justify to ownership — it's not just avoiding repairs, it's reducing the operating cost every month. The replace-vs-repair calculator runs all of this: replacement cost, energy savings, and payback, with a recommendation.
5. What's the cost of it being down?
This is the one that never shows up in a formula but matters enormously to the owner. A rooftop unit serving a restaurant dining room in July, a server room, or a retail floor on a Saturday isn't just an equipment cost — it's lost business and angry customers every hour it's down. For that owner, the reliability of a new unit is worth a premium over nursing an old one through another failure. For a unit serving a warehouse that's comfortable enough either way, downtime costs less and you can afford to repair.
Age and refrigerant decide most of these before you ever open a spreadsheet.
Putting it together
Take that opening example — 7.5-ton unit, 16 years old, $3,800 repair, $4,200/year to run. Installed replacement is around $16,500. The repair is roughly 23% of replacement, which alone says "repair." But it's 16 years old, the energy savings of a new unit are real, and if it's R-22 with a sealed-system problem, that's the ballgame. That's why it lands in "lean replace" — the age and the platform outweigh a repair that looks affordable on paper. Run your own numbers here.
How I present it to the owner
I don't hand them a verdict. I hand them the five answers: here's the age, here's the refrigerant situation, here's the repair as a percent of a new unit, here's the energy payback, and here's what downtime costs you. Then the decision is theirs and they understand it. An owner who understands why replacement makes sense buys a unit. An owner who feels sold a unit fights you on every line. The honest math sells itself.
About the author: Gregory Frazier is a heavy commercial HVAC estimator working Maryland, Pennsylvania, Delaware, Virginia and Florida for 25+ years. Read his story →