The Ball Screw That Lost Its Backlash After a Year
We selected a C7-rolled ball screw for a positioning axis. It had 0.05 mm backlash when new. After a year of 24/7 operation, the backlash grew to 0.15 mm. The axis lost accuracy — the following error doubled. The problem: we didn’t preload the screw. A C7 rolled screw (without preload) has inherent backlash, and as the balls and nut wear, the backlash grows. We should have specified a preloaded nut (double nut or single-nut preload) for a positioning axis. For a conveyor or an axial feed, C7 without preload is fine. For precision positioning, you need preload. The mistake was applying a cost-saving choice (no preload) to a precision axis.
Ball screw preload and backlash selection directly determines the axis’s repeatability. This article covers the preload classes and when each is needed.
Backlash vs Preload
Backlash is the free play between the screw and nut. When you reverse direction, the nut doesn’t move until the backlash is taken up — that’s the lost motion. It directly degrades positioning accuracy.
Preload removes backlash: the nut is loaded against the screw so there’s no free play. The trade-off: preload increases friction and heat (the screw runs warmer). Too much preload and the screw fails early from heat.
Accuracy Grades
- C7 (rolled): Standard, rolled thread. Backlash about 0.05 mm (without preload). Cheap. For general positioning (conveyors, non-critical axes).
- C5 (ground): Ground thread. Backlash about 0.01–0.02 mm. For precision positioning. Usually supplied preloaded.
- C3/C0 (ultra-precision): Ground, lapped. For ultra-precision (micrometer, CNC grinders).
For most automation (assembly, dispensing, pick-and-place), C5 with preload is the sweet spot. C7 without preload is for light, non-critical axes.
Preload Types
Single-Nut Preload (Spring or Oversized Ball)
A single nut with slightly oversized balls (or a spring plunger) that preloads the balls against the race. Simple, compact. But the preload is limited (about 5–10% of dynamic load). Good for light-to-medium preload.
Double-Nut Preload (DNB)
Two nuts connected by a spacer (or a spring). The spacer pushes the nuts apart, preloading the balls. Preload is adjustable (shim the spacer). Standard for precision axes. Preload can be 5–15% of dynamic load.
| Preload Type | Preload Range | Best For |
|---|---|---|
| None (C7) | 0 (0.05 mm backlash) | Conveyors, light feed, non-critical |
| Single-nut (oversized ball) | 5–8% of Ca | Light precision, compact |
| Double-nut (spacer) | 5–15% of Ca | Standard precision positioning |
| Heavy preload (P0/P1) | >15% of Ca | Ultra-precision, high stiffness |
Step 1: When to Preload
Preload is required when:
- The axis reverses direction often (pick-and-place, CNC). Without preload, the lost motion accumulates.
- Positioning accuracy is critical (0.02 mm or better).
- The load reverses (the screw pulls and pushes). Without preload, the nut slaps back and forth.
Preload is NOT needed when:
- The axis moves in one direction only (a feed that pushes and retracts, but doesn’t reverse position).
- The axis carries a constant unidirectional load (the load itself takes up the backlash).
- Accuracy isn’t critical (a conveyor pusher, a clamp).
The preload rule: If the axis reverses direction and accuracy matters, use a preloaded nut (C5 double-nut). The screw that lost backlash after a year was a C7 without preload on a reversing axis. For one-directional feed or a clamped load, C7 without preload saves cost.
Step 2: Preload Amount
The preload (F0) is a fraction of the dynamic load rating (Ca). Typical preloads:
- Light preload (P0, ~5% Ca): General. Low heat, enough stiffness for most automation.
- Medium preload (P1, ~10% Ca): For high rigidity (cutting, grinding). More heat.
- Heavy preload (P2, ~15% Ca): Ultra-precision. Watch the temperature rise.
Too much preload increases the friction torque. The motor must overcome the preload torque:
T_preload ≈ F0 × lead / (2π × η)
For F0 = 500 N, lead = 5 mm, η = 0.9: T = 500 × 0.005 / (2π × 0.9) = 0.44 N·m. Small. But at P2 (15% of Ca = 1500 N), T = 1.3 N·m. The motor must handle this extra torque continuously.
Step 3: Heat and Temperature Rise
Preload creates heat (the balls rub harder). The screw’s temperature rises. For long screws, thermal expansion can cause binding (article 61). Monitor the temperature.
- Light preload: temperature rise about 5–10°C.
- Medium/heavy preload: rise 15–25°C. Need cooling (or a longer expansion joint).
For long screws (over 1.5 m), use a fixed-supported end (article 61) and allow for thermal expansion. Heavy preload on a long screw risks binding.
A Preload Selection Checklist
- Does the axis reverse direction? (If yes, preload.)
- What is the required accuracy? (0.05 mm or 0.01 mm?)
- Is the load unidirectional? (Then preload may not be needed.)
- C7 (rolled, no preload) or C5 (ground, preloaded)?
- Single-nut or double-nut preload?
- Preload level? (P0, P1, P2?)
- Does the motor handle the preload torque?
- Is the temperature rise acceptable? (Long screw?)
- Is the screw fixed-supported? (For thermal expansion?)
- Will backlash be measured at FAT? (As a baseline?)
The Bottom Line
Ball screw preload and backlash choose based on the axis’s reversing duty. The screw that lost accuracy after a year was a C7 without preload on a reversing positioning axis. For reversing precision axes, use C5 with a double-nut preload (P0 or P1). For one-directional feed or unidirectional loads, C7 without preload saves cost. Size the preload torque for the motor, and watch the temperature rise on long screws.
