I'm the person who gets the call when a spindle lets go at 4 p.m. on a Thursday and the customer needs parts Monday. I coordinate equipment sourcing and emergency machine replacement for a contract machining shop — 60+ machine purchases and rush replacements over 11 years, including three jobs where the replacement lathe had to be cutting chips inside a week.
What follows is the checklist I actually run, in order, when a CNC lathe has to be sourced against a fixed date. Nine steps. You can compress them into a single day if you're desperate. You can't skip them.
Who this is for
Use this if you're sourcing a CNC lathe against a hard date. A machine died. A contract starts in five weeks. A customer moved their schedule up and didn't tell you until the PO arrived. It also applies if what you actually need is CNC lathe machining capacity rather than a machine — steps 1 through 3 are identical, and step 2 will tell you whether you should be buying anything at all.
Skip this if your timeline is measured in months and you're running a normal capital process. This checklist is deliberately blunt. It favors speed and verifiable facts over elegant evaluation matrices.
Phase 1: Before you call a single supplier
Step 1: Freeze the spec. Write it down. Get it signed off.
Almost everyone starts by calling suppliers. That's backwards. If two suppliers quote against two different versions of what you need, you can't compare them — and you'll waste the two days you don't have.
The non-negotiables I write down every single time:
- Max turning diameter and max turning length — the actual numbers you need to cut, not the catalog's "swing over bed" figure
- Bar capacity and spindle bore, if you're bar feeding
- Chuck size and spindle nose (A2-6 and A2-8 are the common ones; spindle nose dimensions are standardized under ISO 702-1)
- Spindle speed range and spindle motor rating
- Turret type — station count, and whether you need live tooling or a Y-axis
- Control brand and model
- Tailstock, steady rest, chip conveyor, bar feeder interface
Get the shop supervisor or the programmer to sign off on that list (not purchasing — purchasing will optimize for price and you will pay for it later in setup time). One page, dated, agreed. This document becomes your filter. Any quote that doesn't meet it goes in the bin without a reply.
Step 2: Decide what you're actually buying — and whether you should be buying at all
Four real options: new, used, remanufactured, or outsource the parts to a job shop. The one people skip is outsourcing. If you need 200 parts in six weeks, buying a lathe is almost certainly wrong — you'll spend four weeks on sourcing and commissioning and six weeks learning the machine. That's a job-shop phone call, not a capital purchase.
On used machines: I've bought them, and under deadline they're often the right answer. A machine somebody else already commissioned and ran for eight years is a known quantity in a way a brand-new build never is. The catch is verification (step 5) — you need spindle hours, a control history, and a test-cut part, not a photo of the paint.
Where I've gotten this wrong: we were looking at a 12-year-old turning center against a new one. The upside was roughly $40,000 in savings and about three weeks of lead time. The risk was a control board failure with a six-week replacement lead on that specific revision. I kept asking myself — is $40,000 worth potentially blowing a contract that bills $220,000 a year? We bought new. I still don't know if that was right, but I know why I decided it.
Phase 2: Quotes, verification, and the things that aren't on the quote
Step 3: Ask for acceptance-test data, not catalog specs
This is the step most buyers skip, and it's the one that separates a machine from a brochure. Every builder prints a positioning accuracy number. Almost none of them tell you how they measured it. You want the actual test protocol and results, on the specific serial number you're buying.
Relevant standards to name explicitly in your RFQ: ISO 230-2 (determination of accuracy and repeatability of positioning of numerically controlled axes), ASME B5.57 (methods for performance evaluation of CNC lathes and turning centers), and for general-purpose parallel lathes, ISO 1708 (acceptance conditions — testing of accuracy). Ask which of these the builder's report follows, and ask for the report itself.
I'm not a metrologist — I can't tell you how to set up a laser interferometer or how to properly apply ISO 230-3 thermal testing. What I can tell you from a sourcing perspective is what to demand and how to read it: you want the test report for your machine, and you want a test-cut part shipped with it. A turned test piece tells you more about real roundness, taper, and surface finish than any spec sheet will.
Step 4: Get the exclusions in writing, not the price
A machine price means nothing without its exclusions. The number I care about is the landed, running, cutting-chips number. Make the supplier list, in writing, what is not included:
- Rigging, crane, and freight — including who unloads at your dock
- Foundation work, grouting, and leveling
- Transformer, disconnect, and the electrical run from your panel
- Air supply and dryer
- Bar feeder, chip conveyor, coolant system, and chip bucket
- Toolholders and the initial tooling package (this one is brutal — see the note below)
- Commissioning, training days, and who pays for travel
- Duties, taxes, and customs brokerage on an import
- Spare parts kit and when the warranty clock actually starts
Two ranges worth budgeting for, from publicly listed prices and quotes I saw in Q1 2026 (verify current rates, they move): a modest but complete tooling package for an 8-inch chuck machine runs roughly $5,000–$25,000, and domestic rigging plus installation plus commissioning typically lands between $3,000 and $12,000 depending on your building and how far the machine has to travel inside it. That's on top of the machine. If your budget didn't include those two lines, your budget is wrong.
Step 5: Verify the supplier and the specific machine — separately
These are two different questions and people conflate them. A good supplier with a bad machine is still a bad purchase.
For the machine: live video walkthrough (not a pre-recorded clip), photo of the data plate with the serial number, and on a used machine, the control's operating hours — spindle hours and servo hours, which most modern controls will show you. If the machine is "low hours" but the spindle counter says 31,000 hours, you've learned something.
For the supplier: three references with machines of similar size and configuration, and actually call them. Ask what broke and how fast it got fixed. And here's the thing that matters more than people admit — the vendor who says "this isn't our strength, here's who does it better" earns my trust for everything else. I'd rather work with a specialist who knows their limits than a generalist promising one-stop everything. "One-stop" usually means "nothing best-in-class."
I once approved a used lathe off a video walkthrough alone, because the seller was four states away and we were out of time. The machine was fine. The two weeks between wire transfer and first chip were miserable — every morning I was second-guessing whether the turret index I'd seen on video was a real index or a nudge. We got lucky. I wouldn't do it again without a third-party inspection, even at $800 for the trip.
Step 6: Check your own building before you commit to anything
This is the failure mode that has nothing to do with the machine. Confirm, in writing, with your facilities people:
- Available voltage, phase, and the ampacity of the circuit you intend to use
- Door width, aisle width, ceiling height, and floor loading on the route in
- Whether a crane or forklift can reach the final position, and whether the machine needs to go in before something else does
- Floor slab thickness and whether the machine needs a poured pad
- Compressed air volume and pressure at the machine location
People buy the machine first and discover the door is 2 inches too narrow second. That's a week, minimum, and a wall.
Phase 3: Installation, spares, and ramp-up
Step 7: Book the commissioning window and name the people
Riggers, electrician, millwrights, the programmer, the operator, and the supplier's commissioning engineer all have to be in the same place on the same day. Name every one of them by name in an email. "Bob's doing the electrical" is how you lose a week. "Bob, Tuesday the 14th, 7 a.m., panel B" is how you don't.
Step 8: Lock the warranty and the spares kit in writing
Which parts are covered. Who pays travel and labor. Whether the warranty starts at delivery or at commissioning (get it at commissioning — this is worth arguing about). Note that the control manufacturer's warranty is usually separate from the machine builder's, with a different clock and a different phone number.
Then order the spares kit on day one, not after the first breakdown: a set of fuses, the control battery, the coolant pump, and one of whatever the most likely failure part is for that model. A $600 kit has saved me three days before.
Step 9: Ramp up in stages
Don't go from installation to three shifts. First article inspection and sign-off, then tooling trials with production toolholders (not the ones the engineer brought), then a slow shift, then full rate. Your setup people need to prove the program out on real workpieces before the schedule depends on it.
Common mistakes I keep seeing
- Comparing machine prices instead of landed costs. The cheaper quote is often the more expensive purchase once rigging, tooling, and power are added.
- Accepting a spec sheet as evidence. Ask for the report and a test-cut part. If the supplier can't produce either, that's your answer.
- Buying the machine the salesperson has in stock rather than the machine your spec calls for. Re-read step 1. It exists for this reason.
- Forgetting that operator familiarity is a real cost. A control your team has never run adds weeks of learning you didn't budget for.
- Skipping the third-party inspection on a distant used machine. It costs a few hundred dollars. Not doing it costs a lot more.
Where my experience stops
My experience is based on roughly 60 machines, almost all 2-axis turning centers with 6- to 12-inch chucks and beds under about 12 feet. If you're sourcing a Swiss-type, a mill-turn, a large-bore oil country lathe, or anything with a sub-spindle, parts of this checklist will not transfer cleanly — the verification questions change, and the commissioning timeline is longer than anything I've described here. I've only bought domestically, too, so I can't speak to the customs and brokerage side beyond what I wrote in step 4.
Bottom line: freeze the spec, demand the test data, price the exclusions, verify the supplier and the machine separately, and check your own building before you sign. Nine steps, in that order. Under deadline, the order is the whole point.