I get the same eight questions over and over. Sometimes at 7 a.m. from a plant manager whose machine just went down, sometimes from a procurement lead who has three quotes in front of her and can't tell them apart. So instead of writing another "what is a centrifugal compressor" page, I'm going to answer those eight questions here, in order, with as little fluff as I can manage.

Short version of who's answering: I work in equipment sourcing. I've coordinated 40+ expedited compressor projects over nine years — new builds, rerates, surplus units, rentals used as bridges. I don't build compressors. I find them, vet them, and tell people when what they're asking for isn't physically possible. That last part is most of the job.

1. What does "centrifugal compressor sourcing" actually mean?

It means someone else does the matching work between your process datasheet and the supply that exists in the world right now.

That supply isn't one thing. It's new-build machines from OEMs, surplus units that were ordered and never installed, refurbished API 617 machines with documented rotor history, rerated units (new impellers, same casing), and rental fleets for bridging a gap. A sourcing service doesn't manufacture — it knows which of those five channels actually fits a given pressure, flow, and gas composition, and what each one costs in time.

Put another way: the OEM sells you their machine. Sourcing tries to find the machine your process needs, which is not always the same thing.

2. We need it in three weeks. Is that possible?

For a new-build API 617 machine? No. Not three weeks, not three months. In my experience a new-build centrifugal compressor runs 30 to 60 weeks from purchase order to factory acceptance test, and on large motor-driven trains I've seen longer. Anyone who tells you otherwise either doesn't understand what they're quoting or is quoting something that isn't new.

What is possible in three weeks: a rental unit that matches your flow and pressure closely enough, a surplus or lightly used machine that happens to line up, or — if you're talking plant air rather than process gas — a packaged centrifugal air compressor from stock.

When I first started this work, I assumed the answer to every emergency was "we'll find something." Two years and one very expensive mistake later, I learned that the honest answer is a feasibility call within 24 hours, not a yes. In March 2024 we had a client in the Gulf Coast lose a process compressor 36 hours before a restart. We could not supply a machine. What we could do was confirm within a day that nothing on the market matched their gas composition, which let them pivot to flaring and a temporary reroute instead of waiting on us. That call was worth more than a false yes would have been.

3. Why is the fastest quote often the riskiest one?

Because "in stock, ships in two weeks" usually means "built for somebody else's datasheet."

The nameplate may say API 617 and the head curve may look right, but the details decide whether it drops in or eats six months. Nozzle orientation. Casing split type. Dry gas seal type and supplier. Lube oil console configuration. Driver coupling and baseplate footprint. Materials of construction for wet gas service. I've watched a buyer save eleven weeks on delivery and then spend nineteen weeks and a large pile of money adapting the unit to their piping.

The conventional wisdom is that speed costs money. My experience with surplus centrifugal compressors is that speed costs adaptation money, and the invoice arrives later, which is why people keep falling for it. At least, that's been my experience with machines above 1,000 kW — on small packages the math is different.

4. What actually drives lead time on an industrial centrifugal compressor?

Almost never the compressor assembly itself. The critical path is usually one of these:

  • Impeller forgings and five-axis machining slots
  • Gearbox or casing castings (foundry allocation is where the real queue lives)
  • Dry gas seals and their lead times
  • Large synchronous motors — frequently the true bottleneck on the whole train
  • Test stand availability for mechanical run and performance testing

On testing: performance tests are normally run to ASME PTC 10, and the test slot is a scheduled resource, not an on-demand service. A machine that's mechanically finished can still sit four to six weeks waiting for a stand.

Here's where I have to be honest about my own data. I don't have hard numbers on how much of total industry lead time sits in castings versus motors. What I can say from our project records is that on the expedited jobs we've handled, motor and seal procurement caused more slippage than compressor assembly did. I wish I'd tracked that more rigorously from the start.

5. What changes on a high pressure centrifugal compressor?

Quite a lot, and more than most first-time buyers expect.

Casing design is the big one. Horizontally split casings are easier to service and common at moderate pressures; barrel casings show up as pressures climb, because the bolt-up and sealing arrangement handles the load better. The crossover point isn't a single number — many OEMs move to barrel designs somewhere between 40 and 70 bar depending on gas density, rotor dynamics, and head per stage. I've seen horizontally split machines specified above that range with special designs, so treat any specific threshold as a starting point for a conversation, not a rule.

Then there's everything downstream of the casing: thicker materials, more conservative rotor dynamic analysis, seal and support system complexity, and heavier testing. The governing specification is still API 617, but the chapters that matter shift, and vendors who are competent at 15 bar are not automatically competent at 90 bar.

API 617 (Axial and Centrifugal Compressors and Expander-compressors) remains the specification most refinery, petrochemical, and gas-processing buyers write into the purchase order. If your datasheet doesn't reference it, ask why — and get the answer in writing.

6. Centrifugal air compressor or screw — where's the line?

I'll stay out of the religious war here, because both technologies do their job well and neither is "inferior." The practical line is flow and duty cycle.

Centrifugal air compressors tend to make sense at higher flows and in oil-free applications — food, pharma, electronics — where air quality is specified to ISO 8573-1 classes and where the machine will run at or near design point most of the time. Below a certain flow, a screw package is usually simpler and cheaper to own, and it tolerates part-load far better.

That part-load point matters and gets skipped in a lot of comparisons. Centrifugal machines surge at low flow, so you need inlet guide vanes, blowoff, or a control strategy that keeps them off the surge line. If your plant air demand swings hard between shifts, price the controls and the blowoff losses before you fall in love with the efficiency number at design point.

7. What should I have ready before I call anyone?

More than you think, and I say that as someone who has received "we need a centrifugal compressor" as an entire request.

The minimum useful set: gas composition and molecular weight, k value, flow in actual and normal units, suction pressure and temperature, required discharge pressure, turndown requirement, utilities available (cooling water condition, power supply, steam), area classification (ATEX, IECEx, or Class I Div 2), driver preference, altitude, and ambient design conditions.

Missing any of these doesn't stop the search, but it slows it down and it invites assumptions — and assumptions on gas composition are where expensive surprises come from. Honestly, I'm not sure why the datasheet is usually the last thing to arrive. My best guess is that process engineers get pulled in late on procurement timelines that were set before anyone asked them.

8. What's changed in sourcing since 2020?

Plenty, and some of it is genuinely better.

The surplus and refurbishment market has gotten more transparent — documented rotor histories, digital records of prior service, remote monitoring data on used units. That's made used equipment a defensible option where ten years ago it was a gamble. Dual-sourcing and second-source qualification have moved from a nice-to-have to standard practice on critical trains. And what was best practice in 2019 doesn't all hold up in 2025: assuming a single OEM allocation slot would hold, for instance, is a habit that's burned a lot of people.

What hasn't changed: a correct datasheet, an honest feasibility call, and a documented test protocol still decide whether a project goes well. The fundamentals are the same. The execution around them is not.

Which is really the answer to most of the questions above. There's no trick to centrifugal compressor sourcing. There's just a lot of detail, a handful of honest conversations, and a willingness to say no before someone signs a purchase order.