A lead screw is a threaded rotational-to-translational gear that constrains a rotating screw axis to a translating nut axis through the kinematic relation ωS·L = 2π·vN, where L is the screw lead in mm/turn [S2]. Installation quality is judged against four gates: end-bearing alignment under 0.02 mm/m runout, nut-to-shaft parallelism, lubrication film at the thread helix, and a measured travel-versus-revolution check within ±0.05 mm over 300 mm of stroke.
Spec-first installation separates assemblies that survive 10⁷ cycles at full rated load from those that fail at the nut entrance in the first 10⁴ cycles. The process applies to ACME, trapezoidal, and square-form threads used in CNC axes, press rams, valve actuators, and lead screw driven stages.
Pre-Install Verification: Lead, Hand, and Free-State Straightness
Lead L (mm/turn) is the translational displacement of the nut for one revolution of the screw, and the transmission ratio is RNS = 2π/L, meaning a 5 mm lead converts 1 rad of screw rotation into 5/2π = 0.796 mm of nut travel [S2]. Confirm L from the vendor part drawing against the drivetrain's required mm/rev before mounting; mismatched lead is the most common commissioning fault on retrofits.
Verify thread hand against motor rotation. With a right-hand helix, ωS and vN carry the same sign; with a left-hand helix they carry opposite signs, so swapping hand inverts the commanded direction [S2]. Bench-spin the screw 5 turns by hand and mark the nut start/end positions to read the actual lead with a 0.01 mm-resolution dial indicator before the assembly enters the frame.
Check free-state straightness on V-blocks with a dial indicator run along the OD over the full length. Acceptance for general machine axes is ≤0.05 mm/m; precision stages (≤0.01 mm positioning) require ≤0.01 mm/m. Any reading above the gate disqualifies the screw for that class, no matter how clean the threads look. A useful cross-reference for hand/lead/thread-form selection is the lead screw types and classifications map.
End-Bearing Mounting: Concentricity, Fix-Float, and Thermal Growth
Use one fixed and one floating end bearing. The fixed end (typically an angular-contact pair, back-to-back or face-to-face, preload per the bearing print) takes axial load and resists screw thrust. The floating end (deep-groove ball bearing or needle roller) carries only radial load and absorbs differential thermal expansion between screw and frame. Without the float, a 1 m steel screw grows ≈12 µm per °C and will buckle the end bearings at a 10 °C rise. [S2]
Set bearing housings with shaft-to-housing concentricity ≤0.01 mm TIR for general axes and ≤0.005 mm for high-speed spindles. The screw journals must be clean, dry, and lightly oiled before mounting; any burr on the journal corner will telegraph into nut-runout within one full turn.
Couple the driven end with a flexible coupling rated for ≥1.5× the motor's peak torque, with parallel misalignment capacity ≥0.1 mm and angular misalignment capacity ≥0.5°. Rigid couplings are a frequent root cause of premature thrust-bearing failure; side-loads from misalignment are amplified by the screw's lead and appear as nut-binding long before the bearing itself runs hot.
Nut-to-Shaft Parallelism and Preload Selection
Nut-to-shaft parallelism is the single largest driver of wear life. The mounting surface of the nut housing must be machined flat to ≤0.02 mm across its footprint, and the nut bore must be aligned to the screw axis within ≤0.03 mm over the full travel length. Misalignment past 0.05 mm/m will lift one thread flank into edge contact, raising contact stress by a factor of three and cutting the B-10 life by an order of magnitude. [S2]
Preload options: zero-preload (sloppy, used for hand-driven instrument screws), light preload (≤5% of dynamic load rating Ca, used for positioning stages where stick-slip is unacceptable), and heavy preload (5–10% of Ca, used for rigid CNC axes where stiffness > 50 N/µm is required). The ball screw article documents the same preload bands for recirculating ball nuts; the comparable lead-screw band is 2–4% lower because the sliding contact generates more frictional heat at the same preload.
For self-locking requirements — vertical axes where the load must not back-drive the screw — confirm that the installation operates in the self-locking regime where ηNS < 0, which occurs when cos(α) < k/tan(λ) (acme half-angle α, friction coefficient k, lead angle λ) [S2]. Single-start ACME screws with lead ≤5 mm and steel-on-bronze contact satisfy this gate; a multi-start screw with the same load will back-drive and need a brake.
Lubrication Film, Anti-Backlash, and Surface Treatment Gates
Lubricant selection depends on load and speed. For PV ≤ 0.5 MPa·m/s, a NLGI #2 lithium EP grease with MoS₂ is typical. For PV between 0.5 and 1.5 MPa·m/s, switch to a synthetic PAO oil with ISO VG 68–100 viscosity, applied through an automatic lubricator at 0.5–1.0 mL per 100 mm of travel per hour. Above 1.5 MPa·m/s, an externally pressurized oil-mist or oil-air system with VG 220 is the conservative path.
Thread surface treatments extend life: PTFE coatings lower the breakaway torque by 30–60% and are common on leadscrews in cleanroom and food-grade equipment; hard chrome (20–30 µm) lifts surface hardness to 60+ HRC for abrasive environments; black-oxide plus oil is the budget default for indoor machine-tool axes. Untreated steel-on-bronze is acceptable for low-cycle applications but should never be specified for >10⁵ cycles at rated load.
For anti-backlash on a single nut, install a split nut with a spring or shim stack that biases the two halves apart by 0.05–0.10 mm; verify the bias by measuring the axial play with a dial indicator under 5 N of nut-side load. Play under 0.02 mm is over-tight and will burn the nut; play over 0.15 mm will show as lost motion at reversal. The option that lives between single-nut split and a twin preloaded nut is the ball screw — review its spec gate before committing to a heavy-preload lead-screw retrofit.
Commissioning Tests: Backlash, Drag Torque, and Travel Verification
Backlash test: lock the screw, apply a 10 N axial load to the nut in the driving direction, and read position. Reverse the load to the opposite direction and re-read. The difference is backlash; the gate is ≤0.05 mm for general positioning and ≤0.01 mm for precision stages. Values above the gate indicate either misaligned end bearings or a worn nut, not just loose mounting. [S2]
Drag-torque test: with the screw unpowered, rotate it by a torque wrench through one full turn at the drive end. Compare the measured drag against the vendor's published breakaway and running torque (typical breakaway for a 16 mm ACME lead screw with bronze nut: 0.2–0.5 N·m; running after one full revolution: 0.10–0.25 N·m). Spikes above 2× the running value point at lubrication starvation, thread damage, or end-bearing preload that is too high.
Travel verification: command 300 mm of travel from the controller, measure the actual nut displacement with a laser interferometer or a calibrated linear scale, and compare against commanded. Acceptance: ±0.05 mm over 300 mm for general axes, ±0.01 mm for precision. This step is the field-level equivalent of the kinematic constraint ωS·L = 2π·vN and is the single best check that lead, hand, and reduction ratio are wired correctly [S2].
Common Failure Modes and When to Replace, Not Repair
Symptom → root cause → action map: (1) Nut binds at one end of travel → misaligned end bearing or twisted screw → re-shim the fixed bearing, replace the screw if straightness reading has shifted by more than 0.02 mm/m since commissioning; (2) Audible clicking each revolution → chipped thread tooth from overload or contamination → replace nut, inspect screw flanks with 10× magnifier, replace screw if gouges exceed 0.2 mm depth; (3) Rising drag torque after the first 10⁴ cycles → lubrication starvation or wrong viscosity → re-lubricate and re-test; if torque does not recover within 1×10³ cycles, replace the nut; (4) Backlash doubles within 10⁵ cycles → wear past the preload budget → either re-shim the split nut (if the design allows) or replace the nut assembly. [S2]
Do not repair a screw with a bent journal, a thread profile measured more than 0.1 mm off-print, or pitting corrosion covering more than 5% of the active thread length — replace. Field-repair of a lead screw is limited to cleaning, re-lubrication, and re-shimming of the split nut; thread re-cutting, plating, or straightening should be sent back to a specialty shop and is rarely economic for standard ACME parts below 25 mm diameter.
Track two signals post-installation: drag-torque trend at 1×10³ cycle intervals for the first 10⁴ cycles, and backlash reading every 5×10⁴ cycles thereafter. Cross-check the alignment of the surrounding axis with a linear guide straightness read, because a worn guide rail will load the nut asymmetrically and mask the real failure source. Further context on the mechanical envelope of an installed axis — including load, span, and base-prep gates analogous to a foundation pour — is covered in the expansion anchor installation field walk-through.