In linear guide selection, two performance targets appear to be diametrically opposed: rigidity — the guide system’s resistance to deflection under load — and low friction — the ease with which the carriage moves under drive force.
Increase preload to gain stiffness, and you pay for it in friction, heat, drive motor sizing, and bearing life. Reduce preload for smooth, efficient motion, and you sacrifice the positional accuracy and vibration damping your application demands.
This tension is not just theoretical. It determines whether a semiconductor wire-bonding machine holds its 1 µm placement tolerance under vibration, or whether a CNC machining center develops thermal drift after two hours of operation.
Getting the preload specification wrong is one of the most common — and most expensive — errors in linear motion system design.
This guide builds the complete framework: the mechanics of preload, how precision grades interact with preload selection, how to calculate required stiffness, and the application matrices that eliminate guesswork from the selection process.
What Is Preload and Why It Matters
Preload in a linear guide is a deliberately applied internal force between the rolling elements (balls or rollers) and the raceways — created by using oversized rolling elements, by offsetting opposing carriages, or by grinding the raceway geometry to create an interference fit.
This pre-compression eliminates internal clearance, and that elimination is the source of both the benefits and the penalties of preload.
Without preload, any external load first eliminates the internal clearance before the rolling elements can actually transmit force — creating a “dead band” of zero stiffness that manifests as positional error, vibration susceptibility, and chatter in machining applications.
With preload, the rolling contact is always under compression, and the system responds immediately and elastically to external loads.
Core Definition Preload force is typically expressed as a percentage of the guide’s basic dynamic load rating (C): Light preload = 2–5% of C. Medium preload = 5–8% of C. Heavy preload = 8–12% of C. Ultra-heavy preload = 12–15% of C. This standardization allows direct comparison across manufacturers and size classes. |
The Mechanics: Stiffness vs. Friction Tradeoff
The relationship between preload and system stiffness is governed by Hertz contact theory.
As preload increases, the contact ellipse between ball and raceway grows, and the local elastic deformation decreases relative to the applied load — meaning the system "feels stiffer."
However, this expanded contact area also increases rolling resistance through a mechanism called elastic hysteresis.
The stiffness curve is concave — the first increment of preload produces the largest relative stiffness gain. Moving from zero clearance to light preload might double radial stiffness. Moving from light to heavy preload might add only 40% more.
Meanwhile, friction increases roughly linearly with preload force, and the heat generated (which scales with friction × velocity) can cause thermal expansion of the guide rail — ironically reducing the positional accuracy you added preload to achieve.
⚠ The Thermal Paradox Heavy preload in a high-speed application creates a self-defeating loop: excess friction → heat generation → thermal expansion of rail → positional drift. In a 1-meter ballscrew-driven axis running at 1 m/s with C2-grade preload, thermal expansion can account for 5–15 μm of positional error per hour — often exceeding the guide’s inherent accuracy grade. |
This is why the preload specification cannot be made in isolation from speed. The preload-speed product (analogous to the DN value in rotary bearings) is the practical limit: high preload + high speed = unacceptable thermal load.
The correct engineering process determines required stiffness first, then selects the minimum preload grade that meets that stiffness requirement.
“The perfect preload is not the highest preload your application can tolerate — it is the lowest preload that meets your stiffness requirement, at your operating speed, in your thermal environment.”
Preload Grades Explained
Most major manufacturers (THK, NSK, IKO, Hiwin, Bosch Rexroth) use a consistent 5-grade preload classification system, though designation letters vary by brand.
The table below uses the THK/NSK convention as the industry reference.
Grade | Force (% C) | Stiffness Gain | Friction Increase | Max Speed | Thermal Risk |
Z0 Light | ~2% | Baseline | Baseline | No limit | Minimal |
C0 Light-Med | ~5% | +35–50% | +25–40% | Full rated | Low |
C1 Medium | ~8% | +60–80% | +50–70% | ≤80% rated | Moderate |
C2 Heavy | ~12% | +90–120% | +80–110% | ≤50% rated | High |
C3 Ultra-Heavy | ~15% | +120–150% | +120–150% | ≤25% rated | Very High |
Precision Classes and Their Interaction with Preload
Precision grade and preload grade are independent specifications — but they interact critically. A high-precision, lightly-preloaded guide is not the same as a standard-precision, heavily-preloaded guide, even if their stiffness numbers appear similar on paper.
Precision Grade | Running Parallelism | Height Var. | Width Var. | Typical Application |
Normal (N) | ≤ 50 μm/300 mm | ≤ 40 μm | ≤ 40 μm | General automation, logistics, packaging |
High (H) | ≤ 30 μm/300 mm | ≤ 20 μm | ≤ 20 μm | CNC routers, laser systems, light machinery |
Precision (P) | ≤ 15 μm/300 mm | ≤ 10 μm | ≤ 10 μm | Machining centers, CMMs, grinding machines |
Super Precision (SP) | ≤ 7 μm/300 mm | ≤ 5 μm | ≤ 5 μm | Grinding machines, EDM, precision assembly |
Ultra Precision (UP) | ≤ 3.5 μm/300 mm | ≤ 3 μm | ≤ 3 μm | Semiconductor fab, optical equipment, wafer inspection |
The critical interaction: higher precision grades require higher preload to realize their accuracy potential.
A UP-grade guide with Z0 (light) preload will exhibit micro-vibration and sub-µm stochastic positioning errors because the minimal preload doesn't suppress the rolling element dynamics that create noise in the motion profile.
Conversely, applying C2 (heavy) preload to a Normal-grade guide is wasteful — the guide's own geometric tolerances are larger than the positional improvement the heavy preload provides.
Design Rule Match preload grade level to precision grade level as a starting point: Normal precision → Z0–C0 preload. High precision → C0–C1. Precision grade → C1–C2. SP/UP grade → C1–C2 with enhanced lubrication and thermal management. Never specify C3 preload for a UP-grade guide running at speed — the thermal effects will degrade the very accuracy you paid the premium precision grade to achieve. |
Stiffness Calculation Framework
The correct selection sequence starts with stiffness requirements — not with a preload grade shortcut. Here is the four-step engineering process:
01 | Define Allowable Deflection Maximum acceptable position error under worst-case load. For a milling spindle: typically 1–3 μm under full cutting force. For a semiconductor wire bonder: 0.1–0.5 μm. This number drives everything downstream. |
02 | Calculate Required Stiffness K_required = F_max / δ_allowable. Units: N/μm. A 5,000 N cutting force with 2 μm allowable deflection = 2,500 N/μm required guide system stiffness. |
03 | Account for System Compliance Guide stiffness is in series with mount, spindle, and workpiece stiffness. The system stiffness formula is: 1/K_system = 1/K_guide + 1/K_mount + 1/K_other. System stiffness is ALWAYS less than any individual element. |
04 | Select Minimum Preload Grade Choose the lowest preload grade that achieves K_required, considering speed constraints and thermal environment. Often, adding carriages is more efficient than upgrading preload grade. |
Published stiffness values from manufacturer datasheets are typically given for the standard preload grade (C0). The multipliers for other grades are approximately:
Direction | Z0 Multiplier | C0 (Base) | C1 Multiplier | C2 Multiplier | Notes |
Radial (vertical load) | 0.70× | 1.00× | 1.45× | 1.90× | Most critical load direction |
Reverse radial (upward) | 0.65× | 1.00× | 1.50× | 2.00× | Critical for inverted mounting |
Lateral (side load) | 0.70× | 1.00× | 1.40× | 1.85× | Important for lateral cutting forces |
Pitching moment | 0.60× | 1.00× | 1.55× | 2.10× | Sensitive to carriage span |
✓ Worked Example Required radial stiffness: 1,800 N/μm. Catalog shows 25-size guide at C0 = 1,200 N/μm. C1 multiplier = 1.45×, giving 1,740 N/μm — marginal. C2 gives 2,280 N/μm. However, the axis runs at 2 m/s continuously. C2 at this speed creates unacceptable thermal load. Solution: use two carriages per rail, which approximately doubles system stiffness at C0 preload — achieving 2,400 N/μm without speed penalty. |
Lubrication’s Critical Role in the Equation
Lubrication is the hidden variable that makes or breaks the preload decision. Proper lubrication reduces the friction penalty of high preload by up to 40%, extends guide life by 3–5×, and — critically — changes the stiffness behavior at ultra-light preload levels.
The viscosity of the lubricant also matters at high preload. A grease with too high a base oil viscosity at operating temperature becomes a viscous drag that adds to the preload friction — the guide effectively "feels" like it has more preload than specified.
For high-speed, high-preload combinations, low-viscosity base oil greases (ISO VG 32–46) or oil-air lubrication are essential to avoid exceeding the thermal design limit.
Method | Best Preload | Max Speed | Friction Reduction | Maintenance | Environment |
Oil-air mist | C0–C3 | 100% rated | −40% | Continuous (system) | All environments |
Grease (periodic) | Z0–C2 | 80% rated | −20% | 500–2,000 km | General industrial |
Long-term packed grease | Z0–C1 | 60% rated | −15% | 1,000–5,000 km | Clean/sealed/medical |
PTFE dry film | Z0 only | 100% rated | −30% | Reapply periodically | Vacuum, clean room |
Master Selection Table
Cross-reference your application against this matrix to establish the baseline specification:
Application Type | Precision Grade | Preload Grade | Lubrication | Speed | Primary Constraint |
Semiconductor wire bonding | UP | C1 | Oil-air | High | Sub-μm positioning, vibration damping |
Wafer inspection / AOI | UP/SP | C0–C1 | Oil-air | Very High | Speed + repeatability at nanometer level |
5-axis machining center | P | C1–C2 | Periodic | Medium | Rigidity under cutting forces |
CNC turning center | P/H | C1 | Periodic | Med-High | Balance of stiffness and thermal |
Grinding machine | SP | C1–C2 | Oil-air | Low-Med | Vibration, roundness error |
CMM / metrology | UP/SP | Z0–C0 | Long-term | Very Low | Friction-induced positioning error |
Medical imaging gantry | P | C0 | Long-term | Low | Smooth motion, low noise, cleanability |
Laser cutting / engraving | H/P | C0 | Periodic | High | Speed, lightweight carriage |
Packaging / pick-and-place | H/N | Z0–C0 | Periodic | Very High | Speed, low inertia, cost |
Heavy planer / boring mill | H | C2–C3 | Oil-air | Very Low | Maximum rigidity, heavy cutting |
3D printer / additive mfg | H | Z0 | Long-term | Medium | Low cost, low friction, easy assembly |
Optical alignment stage | UP | Z0 | PTFE dry film | Micro-step | Zero stick-slip, cleanroom compatible |
Application-Specific Deep Dives
Semiconductor & Precision Optics: The Ultra-Low Friction Imperative
In wire bonding, die attach, and wafer inspection equipment, the conflict between preload and friction manifests as stick-slip — the micro-jumping phenomenon where static friction momentarily locks the carriage, then releases suddenly, causing a position overshoot. At nanometer positioning scales, stick-slip events of even 50 nm are catastrophic.
The correct specification is C1 preload with oil-air lubrication. The continuous oil delivery eliminates the transition between static and dynamic friction coefficients, suppressing stick-slip. The C1 preload provides sufficient damping to reject floor vibration. The oil-air system’s cooling keeps thermal expansion below 1 μm over an 8-hour shift.
Machining Centers: The Rigidity-Thermal Balance
In a horizontal machining center, the X-Y table guides face cutting forces up to 15,000 N and require radial stiffness of 3,000–5,000 N/μm to maintain tolerances under interrupted cutting.
This demands C1–C2 preload on 35–45 size guides with 4 carriages per axis. However, table traverse speeds of 30–60 m/min at C2 preload would generate enough heat to defeat the machine’s thermal compensation system.
The typical engineered solution: C1 preload for standard operation, with C2 available via a second carriage set for heavy roughing. Some manufacturers now offer variable-preload guides with hydraulically adjustable preload that can shift between C0 (rapid traverse) and C2 (cutting) — though these add significant cost and system complexity.
CMMs & Metrology Equipment: Minimum Friction Above All
Coordinate measuring machines present the inverse problem. Cutting forces are absent; the “load” is the probe contact force of 0.01–1 N. What destroys metrology accuracy is friction-induced drag that varies with probe direction and carriage position — creating a systematic position error that appears as a “friction map” in the measurement space.
CMMs specify Z0 or C0 preload with PTFE-coated rolling elements in some cases, and rely on air bearings (zero contact, zero friction, zero preload) for the highest accuracy classes. Where recirculating ball guides must be used, the focus shifts to minimizing preload variation across the travel range — achieved through ultra-precision rail grinding and strict temperature conditioning of the measurement environment.
Common Selection Mistakes
# | Mistake | Consequence | Correct Approach |
1 | Defaulting to C1 “for safety” | Excess friction, drive motor oversizing, thermal drift in speed applications | Calculate required stiffness; select minimum preload that meets it |
2 | Specifying UP precision with Z0 preload | Stick-slip at nanometer scale; sub-μm stochastic errors degrade accuracy | Match preload grade to precision grade level (Section 4) |
3 | Ignoring thermal design | Heavy preload + high speed → thermal expansion exceeds positional accuracy | Calculate preload-speed thermal load; verify against thermal compensation capability |
4 | Using series stiffness incorrectly | Specifying guide stiffness equal to required system stiffness | System stiffness < guide stiffness: 1/K_sys = 1/K_1 + 1/K_2 + ... |
5 | Over-lubricating under high preload | Excess grease creates viscous churning heat on top of preload friction | Follow manufacturer’s grease quantity chart; use low-viscosity base oil at C2+ |
6 | Mixing precision grades on same axis | Lower-accuracy carriage limits system accuracy to its grade | Specify identical precision and preload grade for all carriages on a shared rail |
7 | Ignoring mounting surface flatness | Non-flat rail mount pre-loads guide non-uniformly, varying stiffness across travel | Surface grind or scrape mounting surface to ≤ half the rail precision grade tolerance |
Frequently Asked Questions
Q: Can I increase preload after installation to gain more stiffness?
Generally no. Preload in recirculating ball linear guides is set by the physical size of the rolling elements and the raceway geometry — it cannot be adjusted post-installation. Some manufacturers offer double-nut or adjustable preload carriages for heavy-duty applications, but for standard carriages, preload is fixed at the time of manufacture. The only field adjustment is adding additional carriages per rail to increase effective system stiffness.
Q: What is the effect of preload on bearing life?
Preload reduces bearing life because it adds a permanent internal load to the contact stress. The ISO life calculation requires adding the preload force (Fp) to the applied equivalent load (P) before computing L10 life: P_effective = P + Fp. At C2 preload (~12% of C), this can reduce calculated L10 life by 30–40% compared to the same guide at Z0 preload under identical external loading. Heavy preload should therefore only be specified when the stiffness requirement genuinely demands it.
Q: How does roller-type linear guide compare to ball-type for the preload/stiffness tradeoff?
Roller-type guides (using cylindrical or crossed-roller elements) have fundamentally better stiffness-to-preload ratios than ball guides, because roller contact is line contact rather than point contact. At the same preload force, a roller guide achieves 2–3× higher stiffness than a ball guide of the same size class. The penalty is higher friction than ball guides at equivalent preload, and greater sensitivity to mounting alignment errors.
Q: Does temperature affect preload?
Yes — significantly. Since preload is created by interference between oversized balls and the raceways, any differential thermal expansion between the carriage and rail material changes the effective preload. For standard steel guides, a 10°C increase in carriage temperature relative to the rail will increase preload by approximately 5–8% above the specified value. This is one reason high-preload, high-speed applications require active cooling or oil-air lubrication.
Q: How many carriages per rail should I specify?
The minimum is two carriages per rail for any application with moment loads or overhanging masses. For applications requiring maximum stiffness without increasing preload grade, adding carriages is often more effective than upgrading preload grade — four carriages at C0 can match the stiffness of two carriages at C2, without the thermal penalty. The carriage span should typically be 1.5–3× the carriage length for optimal moment resistance.
Conclusion
The preload-precision balance in linear guide selection is not a compromise — it is an optimization problem with a correct answer for each application. The engineering sequence is clear: define allowable deflection → calculate required stiffness → determine minimum preload that meets stiffness at your operating speed → match precision grade to preload grade → select lubrication appropriate to the preload-speed combination.
Engineers who start with a preload grade rather than a stiffness requirement will consistently either over-specify — paying thermal and life penalties they don’t need — or under-specify, discovering positional errors and vibration problems during commissioning. The framework in this guide eliminates both failure modes.
About the Author
