I manage a $2.3 million annual equipment and maintenance budget for a 140-person site services company. Over the past 11 years, I've approved 37 compressor purchases—new, used, leased, and a handful I'd rather forget. Every single one came down to two decisions: buy new or buy used, and spec heat recovery or just a cooling fan.
Those are the two comparisons I'll run here, using the same TCO spreadsheet I actually use at work. Same categories, same blind spots, no vendor talking points.
Here's the framework up front so you can see where this is going:
If you've ever signed off on a "great deal" that turned into a maintenance headache, you already know why the first number on the quote is the least interesting one.
A new Atlas Copco XAS 185, depending on configuration and dealer, ran somewhere in the $18,000–$24,000 range in the Q4 2024 quotes I collected. A used unit with 3,000–5,000 hours typically listed for $9,000–$14,000 on the secondary market.
On sticker alone, the used unit wins by $8,000–$10,000. That's not the number I care about.
I assumed "4,000 hours" meant "half the service life left." Didn't verify the maintenance log. Turned out the unit had skipped two oil changes and the airend was already showing wear. Learned never to assume hours equal condition—the log tells you more than the hour meter does.
Five-year TCO for the used route, in my actual spreadsheet: purchase price, plus a realistic overhaul allowance for the airend, plus the higher consumable burn rate a worn machine runs, plus one extra week of rental coverage per year while it's in the shop. That usually lands 15–25% above the purchase price. The new unit's TCO lands roughly 8–12% above its purchase price, dominated by scheduled service.
Bottom line: the used unit still usually wins on five-year cost—but only by 5–8%, not by 40% like the sticker suggests. If you're buying used to save 40%, you're probably going to be disappointed.
An OEM air filter for the XAS 185 lists around $85–$110 per unit, based on my last dealer quote in Q4 2024. Aftermarket equivalents run $25–$45.
On a per-change basis, aftermarket looks like a no-brainer. We change filters roughly every 500 hours—call it four times a year on a machine running two shifts. That's $340–$440 a year in OEM filters versus $100–$180 aftermarket. A $200-ish annual spread against a $20,000 machine.
Here's where the assumption bites you. On a used machine, I assumed the filter was the filter—same micron rating, same media. Didn't check the spec sheet. Turned out the aftermarket unit had a coarser pre-filter layer and the inner element was loading up twice as fast. We saved $200 on filters and burned $600 in reduced change intervals plus a fouled sensor that cost me a half-day of diagnostics.
To be fair, not all aftermarket filters are bad. Some meet the same ISO 8573-1 filtration class (the compressed-air purity standard) and work fine. But "meets the class" and "meets the OEM media spec" are two different claims, and the second one is the one your warranty cares about. On a machine still under Atlas Copco's service plan, pay for the OEM filter. On a 6,000-hour unit you're already planning to trade, aftermarket is fair game.
This is the comparison most people skip, and it has the widest spread on the whole spreadsheet.
A standard air-cooled compressor dumps essentially 100% of its electrical input into the room as heat, and the cooling fan's only job is to move that heat outside. Think of it as an air conditioner running in reverse—you're paying good money to remove heat you already paid good money to create.
A heat recovery module is the heat-pump side of the analogy: instead of dumping the waste heat, it captures 50–90% of it and puts it to work preheating process water or space heating. According to the U.S. Department of Energy's Compressed Air Challenge materials (energy.gov, accessed December 2024), recoverable heat from oil-injected rotary screw compressors can reach up to 90% of input energy, depending on operating temperature and demand profile.
Run the ballpark on a 75 kW fixed-speed unit at 4,000 hours a year. That's roughly 300,000 kWh of input. Recover even 60% as hot water and you're displacing a meaningful chunk of natural gas or electric water heating—on our site, that worked out to about $6,800 a year in avoided fuel cost. The module runs $9,000–$14,000 installed. Payback: somewhere between 18 and 30 months, depending on how much hot water you actually use.
The cooling fan, by contrast, costs $0 to add and saves you nothing. It just moves heat out the wall.
That said—heat recovery only makes sense when there's a year-round hot water demand within pipe distance. If your compressor runs 1,200 hours a year on a site that doesn't need process heat, the module never pays back. I've also seen installations where the piping run was so long the pumping cost ate the savings. It's situational, not universal.
This is the line item that never shows up on a quote, and it's the one that actually changes decisions.
We didn't have a formal spare-parts buffer for mobile units. Cost us when a used XAS 185 threw a fan belt on a Friday afternoon and the nearest replacement was a Monday-morning delivery. The rental coverage for three days plus the crew standby came to $4,100—more than the belt, the filter, and the labor combined.
New units under a service plan don't eliminate downtime, but they compress it: dealer stock, next-day parts, and a warranty that covers the component if it fails early. Used units, especially ones with patchy logs, carry a wider variance—sometimes they run fine for years, sometimes they strand you on a job site. My spreadsheet now carries a flat $3,500-per-year downtime allowance on any used mobile compressor. That number is not scientific; it's what our last four years of incident reports averaged out to.
The number nobody puts in the TCO model, and everyone should.
Atlas Copco compressors hold resale value fairly well—as of late 2024, I was seeing 6,000-hour XAS 185 units listing at 40–50% of new price on the secondary market, provided the service records are clean. A new unit bought today is worth somewhere in the $8,000–$11,000 range in five years. The used unit you buy for $11,000 today is worth maybe $5,000–$6,500 in five years, and only if you've kept the log current.
If you strip the resale value out of both routes, the five-year net cost gap between new and used narrows to roughly 3–5%. The used route still wins on cash out the door—but it's a lot closer than the sticker suggests.
Here's how I actually decide, after 37 purchases and a few mistakes I'd rather not repeat:
Buy new when: the machine is a production asset, the job has penalty clauses, or the unit will run more than 2,000 hours a year. The service plan and parts availability are worth the premium, and the resale value recovers most of it.
Buy used when: you have a competent in-house mechanic, you can inspect the maintenance log before signing—not after—and the unit will run under 1,500 hours a year on non-critical work. Verify the log, not just the hour meter. If I remember correctly, our best used purchase was a 4,200-hour unit with a fully documented service history; our worst was a 3,800-hour unit with a clean-looking hour meter and a log that stopped two years before we bought it.
Spec heat recovery when: the compressor is stationary, runs year-round, and there's a hot water demand within 30 meters of the discharge. Otherwise, the cooling fan is fine—it's not wasteful, it's just not recovering anything.
Skip heat recovery when: the unit runs seasonally, the site has no process heat demand, or the piping run would exceed the savings. A $12,000 module that never pays back is worse than a $400 fan that does its job.
The one thing I'd tell anyone shopping Atlas Copco compressors for sale—new or used—is this: build the TCO spreadsheet before you fall in love with a quote. The cheapest sticker is almost never the cheapest machine, and the most expensive one isn't automatically the safest bet either. Do the math, then decide.