Linear Guide

What is a Linear Guide

A linear guide is a rolling-element linear bearing built around a profiled rail. The rail is a hardened steel bar with two or four precision-ground raceway grooves running its full length. A carriage, or block, wraps around the rail and contains return channels in which balls or cylindrical rollers continuously recirculate: as the carriage moves, rolling elements enter the loaded zone between block and rail, travel the length of contact, then divert into a return passage and loop back to the start. Because the elements never leave the carriage, stroke length is limited only by rail length, not by the bearing itself.

The defining advantage over a plain sliding way is the contact mechanism. A dovetail or box way slides metal on metal with a friction coefficient near 0.1 and significant stick-slip at low speed. A profiled rail rolls, with a friction coefficient on the order of 0.002 to 0.004 excluding seal drag, which is approximately one fiftieth of sliding contact. Lower friction means less drive force, less heat, smoother creep at micrometer speeds, and far better positioning repeatability, which is why profiled rails displaced sliding ways in most precision machinery from the 1980s onward.

Structurally a linear guide has three functional groups. First, the rail: through-hardened or induction-hardened bearing steel, typically 58 to 62 HRC on the raceways, ground straight and parallel, and counterbored for socket-head mounting bolts (top mounting) or drilled from below (bottom mounting). Second, the carriage: a steel block housing the loaded raceways, the recirculation end caps, the grease reservoir, and the seals. Third, the rolling-element circuit: balls in point contact or rollers in line contact, often held in a polymer cage or chain that spaces them apart to eliminate element-to-element rubbing and reduce noise, the feature THK markets as Caged Ball and HIWIN as SynchMotion.

A single carriage on a single rail already resists load in four directions, downward, upward (pull-off), and both lateral directions, and resists pitch and yaw moments through the spacing of its loaded balls. This five-degree-of-freedom constraint is what separates a profiled rail from a round-shaft linear bushing, which carries no moment on its own. A typical machine axis uses two parallel rails with two carriages each, four blocks total, which adds roll constraint and distributes load, but many compact stages run a single rail with a long carriage where space or cost rules out a second rail.

The category spans an enormous size and load range. Miniature stainless rails as narrow as 1 to 7 mm in width serve medical devices, optics, and 3D printers. Mainstream industrial sizes, designated by rail width in millimeters, run 15, 20, 25, 30, 35, 45, 55, and 65, with basic dynamic load ratings from roughly 14 kN at the small end to over 200 kN for the largest ball blocks, and higher still for roller rails. The number that procurement engineers must internalize is that a single nominal size hides a matrix of carriage lengths, flange shapes, accuracy grades, preload classes, and seal packages, and that matrix is where selection lives.

Types and Classification

Linear guides are classified along several independent axes: rolling element (ball or roller), carriage geometry (flanged or compact, standard or low profile), raceway contact angle, and rail mounting method. Choosing the wrong carriage geometry rarely causes an outright failure the way an incompatible material does, but it does drive cost, footprint, and rigidity, so it is the first fork in the decision tree. The table below compares the common carriage and rail families found across the major catalogs.

Contact geometry is the most important hidden variable. Most modern ball guides use a Gothic-arch (two-point per groove) raceway with a nominal 45-degree contact angle on four rows, which gives equal load capacity in all four radial directions and good moment capacity. This four-way equal-load design, established by the THK HSR and SHS families, became the de facto global footprint, which is why sizes 15 to 65 are dimensionally interchangeable across most brands. A circular-arc single-point groove, by contrast, gives higher capacity in two directions but unequal performance, and is now mostly historical.

Carriage geometry divides into flanged blocks, which bolt down through a wide flange from the top and are the most common, and compact or square blocks, which bolt down through the body and suit narrow envelopes or carriages mounted on their side. Low-profile variants trade a little moment capacity for reduced overall height, useful in flat XY tables. Long and extra-long carriages increase the loaded ball count, raising both load rating and moment capacity at the cost of length.

Rail mounting is top mounting, where counterbored holes accept socket-head cap screws driven from above, or bottom mounting, where the rail is tapped and bolted from beneath, used when the top face must be clear or when the rail forms a structural beam. Rails ship in standard lengths up to roughly 3,000 to 4,000 mm depending on size and can be butt-jointed with matched ends for longer travels, with the joint ground and the hole pattern continued so a carriage crosses it without a bump.

Material grade branches into standard bearing steel for dry, indoor service and corrosion-resistant variants, either martensitic stainless rail and carriage or a surface treatment such as black chrome, electroless nickel, or thin dense chrome, for washdown, coolant, food, and medical duty. Stainless and coated rails carry a small load-rating penalty but are the only durable answer to humidity, mild chemicals, and cleaning chemistry.

Ball versus Roller Technology

The single largest performance and cost fork is whether the rolling elements are balls or cylindrical rollers. Balls contact the raceway at a point; rollers contact along a line. That geometric difference cascades into rigidity, load capacity, life exponent, speed, friction, and price. There is no universally better choice; the table below compares the engineering metrics that drive the decision.

Ball guides dominate by volume. Point contact deforms easily under preload, so friction stays low and uniform, the carriage runs smoothly down to micrometer creep speeds, and the system tolerates a degree of mounting surface error because the contact ellipse can shift to accommodate misalignment. Ball guides reach high speeds and accelerations, suit servo-driven automation, and cost meaningfully less than rollers. The recirculating ball linear guide is the right answer for the large majority of automation, semiconductor, packaging, robotics, and light-to-medium machine-tool axes. Caged-ball designs add a polymer chain between balls that eliminates ball-to-ball collision, cutting noise and extending grease life.

Roller guides exist for rigidity and load. Line contact spreads force over a much larger area, so a roller block of a given size carries substantially more dynamic load and deflects far less under the same force, typically two to three times stiffer than the equivalent ball block. The fatigue model also rewards rollers: the life-distance exponent rises from 3 for balls to 10/3 for rollers, and the rated-life reference travel is conventionally 100 km rather than 50 km, both of which lengthen calculated life for a given load. The trade is cost, slightly higher friction is offset by lower deflection, and a markedly tighter sensitivity to mounting flatness, because a stiff line contact cannot redistribute load around a high spot the way a compliant ball can. Roller rails are therefore reserved for heavy-cutting machining centers, grinding machines, large boring mills, presses, and any stage where deflection under cutting force or self-weight must be minimized.

A practical heuristic: start with a ball guide and size it. If the calculated rigidity is insufficient, the deflection budget is blown, or the load is dominated by cutting or impact forces, step up to a roller guide of the same or one smaller nominal size, because the roller block recovers the rigidity and capacity. If speed, smoothness, cost, or tolerance to an imperfect base surface dominate, stay with the ball guide. Roller rails also demand a more carefully scraped and aligned mounting surface, which adds installation labor that should be costed into the comparison.

A third element type, the crossed-roller way (not a recirculating profiled rail but a finite-stroke V-groove system), serves ultra-high-rigidity, short-travel precision stages where zero backlash and the smoothest possible motion outweigh the limited stroke. It sits outside the recirculating profiled-rail family covered here but is worth noting when travel is short and precision demands are extreme.

Load Ratings, Life, and Standards

Linear guides are rolling bearings, so they are rated and their life is predicted with the same fatigue framework as ball and roller bearings, codified in ISO 14728 for linear motion rolling bearings. Three numbers anchor every datasheet: the basic dynamic load rating C, the basic static load rating C0, and the static moment ratings about the three axes. Understanding what each one bounds is the difference between a sound selection and a premature failure.

Basic dynamic load rating C is the constant load under which a population of carriages reaches a nominal rated travel before fatigue, conventionally 50 km for ball guides and 100 km for roller guides. It is the reference for life calculation, not a load the guide should actually carry continuously. Across mainstream ball families C ranges from roughly 14 kN for a size 15 block to over 200 kN for the largest size 65 blocks; roller blocks of equal size carry more. Basic static load rating C0 is the load that produces a defined permanent deformation (about 0.0001 of the rolling-element diameter) at the most heavily loaded contact, and it bounds short-duration peaks, clamping, and stationary impact rather than running load.

Service life is computed, not guessed. The nominal L10 life, the distance 90 percent of a population survives, is:

For ball guides L10 equals (f_H f_T f_C C divided by f_W P) cubed, multiplied by the 50 km reference travel; for roller guides the exponent is 10/3 and the reference is 100 km. Here P is the equivalent dynamic load computed over the real motion profile, including the moment contribution converted to an equivalent radial load. The load factor f_W is the one engineers most often get wrong: a smooth, slow conveyor sits near 1.0, but a machine with frequent reversals, impact, or vibration sits near 2.0, and because the term is cubed, doubling f_W cuts calculated life by a factor of eight. The lesson is to size against the equivalent load over the duty cycle and an honest shock factor, never against a single static peak.

The relevant standards form a small, stable set. ISO 12090-1 and ISO 12090-2 fix boundary dimensions and tolerances for profiled rail guides (series 1 to 5; a draft extends the series), which is why carriages are cross-brand interchangeable in their mounting envelope. DIN 645 covers the same dimensional ground in the German system. ISO 14728-1 and -2 define dynamic and static load ratings and life calculation for linear motion rolling bearings. ASME B5.64 specifies methods for the performance evaluation of single-axis linear positioning systems, the framework for stating repeatability and accuracy of a complete axis. JIS B 1192 in Japan parallels the ISO bearing-life work. A datasheet that cites these by number, rather than vague marketing accuracy claims, is the first sign of a vendor worth shortlisting.

Accuracy and Preload Parameters

Two class systems, accuracy grade and preload class, define how a linear guide will perform on the machine, and they are shared in structure (if not always in symbol) across THK, HIWIN, Bosch Rexroth, IKO, and NSK. A purchasing engineer who can read these two systems can compare any two catalogs. Everything else on the datasheet, sizes, load ratings, and seals, is secondary to getting these two right.

Accuracy grade bounds five things: the tolerance on carriage height H, the tolerance on carriage width, the height difference between paired carriages on the same plane, the width difference, and, most importantly, the running parallelism of the carriage reference faces relative to the rail mounting faces over the full stroke. The grades, from loosest to tightest, are C (normal), H (high), P (precision), SP (super precision), and UP (ultra precision). The table maps each grade to typical application classes; exact micrometer values vary with rail length and must be read from the specific catalog.

The critical caveat: the published accuracy grade is achievable only if the mounting surface is flat and the rail is aligned to that flatness. A UP-grade rail bolted to a wavy plate runs no better than the plate. Tighter grades therefore carry a hidden cost in base machining and installation labor, and over-specifying accuracy wastes money without improving the machine. Match the grade to what the application actually needs, then verify the base surface can support it.

Preload class sets the internal negative clearance, the interference fit between rolling elements and raceways created by installing slightly oversized balls or rollers. Preload removes backlash, raises rigidity, and improves running accuracy under load, but it also raises friction, drive force, and heat generation. It is expressed as a fraction of the basic dynamic load rating C, and three classes are standard:

THK expresses the same idea with radial-clearance symbols: Normal clearance for fixed-direction, low-vibration, twin-rail installs where low resistance matters; C1 (light preload) for overhang or single-rail moment loads needing accuracy; and C0 (medium preload) where high rigidity is required under vibration and impact, as in heavy-cutting machine tools. The numeric Z0/ZA/ZB scheme and the THK clearance scheme describe the same physical continuum from clearance to heavy interference. Two further cautions apply. First, preload directly raises heat: a battery-stacking machine that was mistakenly specified with heavy preload suffered rail deformation from thermal expansion, a real and avoidable field failure. Second, preload relaxes over the first running-in period and across life, so a guide that needs sustained zero backlash should be sized one class up from the minimum that meets initial rigidity, and rechecked after run-in.

Beyond grade and preload, a complete spec line also names rail size, carriage style and length, number of carriages per rail, accuracy grade, preload class, seal package, lubrication type and port, material or coating, and the rail length with end-machining and joint detail. Reading two competing datasheets means lining up these fields one for one; any blank field is a question for the vendor, not an assumption to make.

Selection Decision Factors

To turn the preceding five chapters into a specific part number, work the decision sequence below in order. Most selection mistakes come not from a single wrong number but from deciding a downstream parameter before an upstream one is fixed. These nine steps double as an RFQ template.

1. Define the load and motion profile: List the magnitude and direction of forces and moments on the carriage in each phase of the cycle, the stroke, speed, acceleration, orientation (horizontal, vertical, wall-mounted), and the duty cycle. This is the input to every later step; skipping it makes load rating meaningless.
2. Choose element type: Ball for speed, smoothness, cost, and tolerance to base error; roller for maximum rigidity and load, heavy cutting, or a strict deflection budget. Start with ball and escalate only if rigidity or capacity fails.
3. Size against life, not peak: Compute the equivalent dynamic load P over the duty cycle, including the moment-to-radial conversion, then solve the L10 life equation for the required basic dynamic load rating C using an honest shock factor f_W. Pick the smallest size that meets the target life with margin.
4. Verify static safety and moments: Check the static safety factor (C0 divided by the worst stationary or impact load, commonly 2 to 4 minimum) and confirm each static moment rating against the applied pitch, yaw, and roll moments, especially for single-rail or cantilever layouts.
5. Set accuracy grade: Match C, H, P, SP, or UP to the real positioning requirement, not the most impressive number. Confirm the mounting surface flatness can actually realize the chosen grade.
6. Set preload class: Light for high speed and twin rails, medium as the general default, heavy only for high-rigidity, single-rail, or high-vibration duty, accepting the added friction and heat.
7. Specify carriage style, mounting, and rail length: Flanged or compact, standard or long carriage, top or bottom mounting, number of carriages per rail, total rail length, and end-machining or butt-joint detail for travels beyond standard rail length.
8. Select sealing, lubrication, and material: Seal package to suit the contamination (standard end seals, double seals, scrapers, bellows, or cover strip), grease type and relubrication interval and port, and standard, stainless, or coated material for the environment.
9. Total cost of ownership: Purchase price plus base-surface machining, installation and alignment labor, lubrication and relubrication over life, and the cost of downtime if the axis is critical. A cheaper rail that needs a flatter, more expensive base, or that wears out and loses accuracy early, is rarely the lowest lifetime cost.

One dimension that is easy to overlook at the purchasing stage but decisive over a ten-year service life is serviceability and supply: local stock of carriages and rails, availability of matched-pair and interchangeable carriages so a worn block can be replaced without rebuying the rail, relubrication accessibility, and a vendor that publishes life-calculation tools and honors the ISO 12090 footprint so a future replacement is not single-sourced. THK, HIWIN, Bosch Rexroth, IKO, NSK, and Schaeffler INA all maintain regional stock and engineering support, which is why they are safe defaults for production machinery, while dimensionally compatible value brands such as STAF, ABBA, and CSK can serve non-critical axes at lower cost. Because sizes 15 to 65 follow a common boundary footprint, a carriage from one brand will usually fit another brand's rail in form and dimension, but preload, accuracy, and seal options never transfer, so always confirm the full spec line rather than the size alone.