Purchase used CNC lathes: how the iron actually degrades

• 1
  What purchase used CNC lathes really buys you
• 2
  Where a used CNC lathe loses accuracy first
• 3
  Control support decides the machine's real life
• 4
  Geometry, turret tooling, and bar capacity
• 5
  When a rebuild beats buying new iron
• 6
  Used lathe condition: what the number means

A lathe is a frame, a spindle, a turret, and a control. The frame barely changes over 20 years. The spindle and turret wear in measurable ways, and the control ages in software, not in metal. So the useful question is not how old the machine is. It is which of those four systems still holds tolerance.

Most turning centers are built around a cast iron or welded steel bed with hardened box ways or linear guides. Box ways wear slowly and can be scraped back into geometry. Linear guides wear as a set and are replaced, not adjusted. Knowing which guide type sits under the turret tells you most of what a rebuild will cost later.

Spindle bearings dominate accuracy on any lathe. A machine that cut aluminium for a decade may show 2–3 μm of radial runout, while a machine that hogged stainless for three years can show five times that. Hours in cut matter more than calendar years. Ask for spindle load history if the control stores it.

The payoff is real: a well-kept turning center costs a fraction of a new one and can hold the same ±0.005 mm on standard work. The catch is that you are buying the maintenance record, not the badge on the door.

• 1
  FrameRarely the limiting factor; check for cracks and leveling history
• 2
  SpindleBearing preload and runout decide your achievable roundness
• 3
  TurretRepeatability drifts with index wear and coupling wear
• 4
  ControlSoftware and spare-part support define the machine's remaining life

Radial and axial runout at the spindle nose is the first number to take. Dial the taper and the face separately. Under 5 μm on both is healthy for a mid-size turning center. Between 5 and 15 μm means new bearings soon. Above 15 μm, price the job as a spindle rebuild before you negotiate.

Turret repeatability comes next. Index the same station ten times and indicate a test bar. A good turret repeats within 5 μm. The wear usually sits in the curvic coupling or the index motor brake, and both are serviceable. If only one station drifts, the coupling has a worn zone; if all stations drift, the clamp is loose.

Guide wear shows up as taper over a long cut. Turn a 300 mm bar and measure both ends. More than 0.02 mm of taper across that length points to way wear or a twisted bed. Check the leveling pads and the floor before you blame the machine. A lathe that was never re-leveled after a move can cut tapered and still be mechanically sound.

Thermal behavior is the wear you cannot see. Run the spindle at 4,000 rpm for 30 minutes and re-check runout. Growth of 10–15 μm is normal. Growth beyond that means preload is set wrong or lubrication is failing, and no amount of cleaning will fix it.

• 1
  Spindle runoutTake radial and axial readings at the taper
• 2
  Turret repeatTen indexes, one test bar, one indicator
• 3
  Taper test300 mm bar, measure both ends
• 4
  Thermal growthRe-measure after 30 minutes at speed

A mechanically sound lathe with a dead control is a parts donor. Before you buy, confirm that the control vendor still sells boards, drives, and parameter backups for that model. FANUC and Siemens hold support longest, but even those families have cut-off generations.

Ask what happens when a drive fails. If the answer is a third-party repair shop with a two-week turnaround, factor that into uptime. If the answer is a supported replacement within days, the machine can carry production work. The control generation matters more than the hour meter on the spindle.

Battery-backed parameter memory is a quiet killer. Controls that sat unpowered for months can lose parameters, and reloading them requires a backup that the previous owner may never have made. Power the machine up in front of you and check for alarms before any money moves.

Servo and spindle drives are the expensive consumables. Pull the alarm history if the control keeps one. Repeated overcurrent or overheat faults on one axis suggest a failing drive or a mechanical bind that will return.

• 1
  Supported generationConfirm board and drive availability from the OEM
• 2
  Parameter backupAsk for a copy; test that the machine boots clean
• 3
  Alarm historyRepeat faults on one axis are a red flag
• 4
  Spares pathKnow the lead time before you commit

Bed geometry is checked with a straightedge and a level, not with a test cut. Twist of 0.02 mm per meter is the practical limit for a lathe that will cut long shafts. Beyond that, you are chasing taper with offsets and the operator pays for it every shift.

Turret tooling follows the machine. A used lathe with 12 stations and a full set of block holders is worth more than the same machine bare. Odd tooling formats, VDI versus BMT, decide which holders you can still buy and at what price. Check that holders seat without shimming.

Bar capacity and spindle bore set the upper limit on what the machine can feed. A 65 mm bore will not run 70 mm bar, no matter what the brochure said. Measure the bore if the spec sheet is missing. The same goes for chuck size and the maximum swing over the cross slide.

Chip management is unglamorous and it decides whether the machine runs unattended. A conveyor that jams or a coolant system that has sat dry for a year costs more to fix than the machine saves. Look inside the coolant tank. Sludge tells you how the machine was treated.

• 1
  Bed twistLevel and straightedge; 0.02 mm per meter is the limit
• 2
  Tooling formatVDI or BMT; confirm holders are available
• 3
  Spindle boreMeasure it; do not trust the spec sheet
• 4
  Coolant and conveyorCheck for sludge and seized parts

A used lathe plus a spindle rebuild plus new guide rails can still land well under the price of a comparable new machine. That math works when the frame is straight and the control is supported. It fails when two of the four systems need work at the same time.

The break-even is usually about geometry. If the bed is straight and the spindle can be re-ground or re-bearinged, a rebuild makes sense for shops cutting standard tolerances. If the bed is twisted and the control is obsolete, you are buying a project, not a machine.

Rebuild time matters as much as rebuild cost. A spindle rebuild can take weeks, and those weeks are lost capacity. If you need the machine cutting this month, buy one that is already running and inspect it under power. If you have slack in the schedule, a rebuild can be the cheaper route.

For parts that need ±0.005 mm and Ra 0.8–1.6 μm, a fresh rebuild on a straight frame will hold those numbers. For tight bores or hard materials, the rebuilt machine may still need a finishing pass on a newer spindle. Plan the process around the machine's real capability, not the nameplate.

• 1
  Rebuild winsStraight frame, supported control, one worn system
• 2
  Replace winsTwisted bed, obsolete control, multiple failures
• 3
  Time costWeeks of rebuild are weeks of lost capacity
• 4
  Process fitMatch the machine to the tolerance you actually need