Ball screw installation failures rarely come from the screw itself; they come from the bearings, the housing bore concentricity, and the lubrication schedule [S10]. FUYU's 2026 FSK40 module lists position accuracy 0.03 mm, max speed 255 mm/s, and max load 25 kg over a 50–1000 mm stroke — numbers that any installer must verify after mounting, not just trust from the data sheet [S1]. The Steinmeyer 1112 miniature, 3 mm diameter, 0.5 mm pitch, 90 N load rating, ground to accuracy class P1–P5, is the bench-scale reference for what tight tolerance looks like in practice [S2].
This guide walks the four physical installation gates, the lubrication rule, and the two failure patterns that drive warranty returns on ball screw units in industrial modules. It is written for maintenance engineers, retrofit builders, and CNC retrofit techs — not for OEM line builders who already have fixtured cells.
Pre-Mount Inspection: Shaft, Nut, and Bearing Block Inventory
Pre-mount inspection catches 70% of field failures before the screw is lifted into the machine [S10]. Verify the ball screw straightness with V-blocks and a dial indicator at both ends and at mid-span; a 0.05 mm/m TIR is the practical field ceiling for general industrial use, with precision-ground screws like the Steinmeyer 1112 specified tighter [S2][S10]. Confirm nut part number against the screw diameter and pitch — mixing a 16 mm × 10 mm nut onto a 16 mm × 5 mm screw will assemble but preload incorrectly, leading to early brinelling [S1][S10].
Check the angular contact bearings on the support blocks: back-to-back (DB) or face-to-face (DF) pairs must be matched, and the bearing bore must be slip-fit to the journal ground section of the screw, not the rolled thread section [S6][S10]. Confirm the fixed-end and supported-end assignments match the manufacturer drawing; swapping them changes the thermal growth direction and induces end-of-stroke thrust spikes. FUYU's FSL80 (stroke 50–1200 mm, position accuracy 0.02 mm, G1610 ball screw) is a useful reference for which end takes the motor-side fixed bearing [S3].
Bearing Block Mounting and Concentricity Gates
Bearing block concentricity to the linear guide carriage is the single highest-leverage alignment step; a 0.02 mm parallel error across the block-to-rail span will translate to a 0.04–0.06 mm positional error at the nut [S6]. Mount the support blocks on a surface ground to within 0.02 mm/m flatness, and torque the cap bolts in a star pattern to the bearing-housing spec — typically 12–18 N·m for M5–M6 cap screws on a 16 mm shaft, but always the screw-maker's number, not a generic table [S10].
Set the angular-contact preload by the spacer or shim method the manufacturer publishes; do not reuse shims between an old block and a new bearing. The Thomson installation guide is explicit: preload that is too light lets the screw axially walk under load reversal; preload that is too heavy accelerates ball-race spalling and burns the grease [S10]. FUYU's FSK40 max-load figure of 25 kgf sits in the mid-duty range where this preload window is narrowest — light enough that a single oversized shim will push it past the fatigue limit [S1].
Nut-to-Carriage Coupling and Parallelism to the Linear Guide
Nut-to-carriage parallelism to the linear guide rail must be within 0.02 mm over the full nut length, measured with a dial indicator swept along the rail while the carriage is hand-cycled [S6][S10]. Most field complaints about "binding screws" are actually nut-mounting parallelism errors, not screw defects. FUYU's FSL80 is described as having a "fully sealed structure, body width 80 mm, dustproof, oleic acid resistance, high temperature resistance" — those features assume the nut is mounted parallel, or the seal lip will bind on the bellows boot and skew the alignment [S3].
For anti-backlash double-nut assemblies, the pre-tension between the two nuts is set by shimming or a disc spring pack; the Thomson guide warns that over-shimming the spring pack eliminates the backlash-cancellation function and forces one nut to carry full load [S10]. For single-nut ground screws like the Steinmeyer 1112 (P1–P5 accuracy class, available with or without axial play), preload is built in by the nut geometry — do not add external pre-tension, or you will exceed the 90 N dynamic load rating margin [S2].
Lubrication: Grease Grade, Interval, and Re-Grease Procedure
Grease selection and interval is the most under-documented gate in the field. The Thomson guide specifies re-lubrication based on duty cycle, not calendar time — a screw running 2 m/s continuous needs more frequent re-grease than one running 0.1 m/s intermittent, even if the latter sits in a dirtier shop [S10]. Lithium-complex or polyurea greases with ISO VG 100–150 base oil are the common industrial picks; for cleanroom or medical applications like the FSL80 use case, low-particle PTFE-thickened greases are standard [S3][S10].
Re-grease procedure: purge old grease through the wipers by pumping fresh grease until clean grease exits both wipers, then wipe the excess. Do not "top off" — layered greases of incompatible thickeners (lithium over polyurea, or vice versa) will separate and starve the ball track [S10]. For vertical-axis screws (FSK40 lists "vertical" as a configurable orientation), add a nut-flood cycle every 500 hours or every 6 months, whichever comes first, to ensure the lower return circuit stays wet [S1][S10].
First-Run Acceptance Test: Backlash, Runout, and Temperature
First-run acceptance is a measured gate, not a visual one. Lock the screw from rotating, apply 50 N axial load in both directions with a force gauge, and measure the lash at the nut — the figure should match the manufacturer spec (P1 grade ground screws: 0–0.01 mm; rolled screws: 0.05–0.10 mm typical) [S2][S10]. Run a full-stroke pass at the spec velocity (FSK40: 55–250 mm/s; FSL80: 60–220 mm/s) and record the motor current trace — any step or bump indicates a misalignment, a damaged ball track, or a swarf-loaded return tube [S1][S3][S10].
After 2 hours of duty cycle, measure the housing surface temperature with an IR thermometer or thermocouple; a stabilized 10–15 °C rise above ambient is normal at rated load, more than 20 °C rise flags lubrication starvation or preload overshoot [S10]. For high-precision ground screws like the Steinmeyer 1112 series, this gate is non-negotiable; the accuracy class is paid for, but only if thermal stability is verified [S2].
Common Failure Modes and When to Replace, Not Repair
Three failure modes dominate field returns. (1) Brinelling at the ball track from shock load or overload — visible as pitch-line craters, requires screw replacement, not retumbling [S10]. (2) Return-tube contamination, where swarf or hardened grease blocks the ball circuit and forces the screw to lock mid-stroke — repairable by flushing with solvent and re-greasing, but only if caught before the ball track is scored [S10]. (3) Nut-to-screw axial play creep, where the nut has worn past the preload window — a double-nut anti-backlash assembly can be re-shimmed; a single-nut ground screw like the Steinmeyer 1112 cannot, and the whole screw must be replaced [S2][S10].
Replacement triggers that are not repair candidates: any ball-track spalling visible under 10× magnification, any axial runout at the journal exceeding the manufacturer TIR by more than 2×, and any preload drift exceeding the original spec on a re-test. The CNC retrofit community treats these as hard cutoffs, since a screw that fails mid-cut on a mill can take the spindle with it [S6]. For sourcing replacement screws in 2026, the Ball Screw Supply Chain 2026 report maps current lead times and the steel-alloy levers that drive cost. When staging a retrofit alongside other mechanical hardware, an installer-level overview of slewing-bearing bolting is given in the Slewing Drive Installation field guide, which shares the same concentricity-and-preload logic.
Track the 2026-07-20 Nanjing Shuntai catalog update for SFU/SFJ 2000 mm stainless nut variants [S9], the Kuroda Precision selection software release (2026-03-10) for duty-cycle and life calculation [S8], and any revision of the Thomson installation/maintenance PDF [S10] — these three nodes are the most reliable indicators of when the installation rules shift versus when only the marketing changes.