A linear guide rail on a gantry axis kept losing preload after a week. The rail was an HRC25, mounted on an aluminum extrusion beam, bolted with M5 countersunk screws every 300 mm. The customer torqued the screws to 8 Nm with a torque wrench. The rail was straight on installation. After a week, the carriage moved with a notch at the middle. The customer re-torqued the screws. It lasted another week. The rail was moving. The screws were torqued correctly. The beam was flexing.
The beam isn’t a reference surface
A linear rail needs a flat, rigid mounting surface. The aluminum extrusion beam used on this gantry was a 60×60 mm profile, 2000 mm long, simply supported at both ends. Under the rail and carriage load (50 kg moving), the beam deflects. The deflection at center is about 0.15 mm. The rail is bolted to the beam. When the beam flexes, the rail flexes with it. The rail stays bonded to the beam because the screws are tight. But the rail’s straightness follows the beam’s deflection. The carriage runs smoothly at the ends but hits a dip at center. The customer hears a “notch” sound. The preload hasn’t changed. The beam is moving.
The torque wrench approach assumes the mounting surface is rigid. If the surface moves under load, no amount of screw torque will hold the rail straight. The rail follows the surface. Garbage in, garbage out.
The fix
The beam was up-sized from 60×60 mm to 90×90 mm aluminum extrusion. The deflection dropped from 0.15 mm to 0.04 mm. The rail ran straight. The notch was gone. The up-sized beam costs $80 more. The alternative — buying a more expensive linear rail that can tolerate beam deflection — costs $500 and still won’t fix the problem. The beam must be stiffer than the rail’s straightness spec.
Also, the rail mounting was changed. Instead of bolting the rail directly to the extrusion, a steel reinforcing plate (8 mm thick) was sandwiched between the extrusion and the rail. The steel plate is flat and rigid. The rail bolts to the plate. The plate bolts to the extrusion. The plate distributes the clamping force and provides a flat reference. The beam flexes but the rail stays flat because the plate bridges across the flex. This is the standard approach for mounting linear rails on extrusions.
The mounting flatness spec
For an HRC25 rail with C0 preload, the mounting surface flatness must be within 0.02 mm over any 300 mm segment. Over a 2000 mm rail, the total straightness should be within 0.05 mm. An aluminum extrusion as-extruded has a flatness of about 0.1 mm over 1000 mm. That’s too loose. You need either a machined reference surface or a steel reinforcing plate bolted to the extrusion. The extrusion alone isn’t flat enough for a preloaded linear rail.
If you’re building a gantry with extrusions, budget for the machined mounting surface. It costs about $100 to have a miller fly-cut the extrusion top surface flat. It’s worth every cent. The alternative is a rail that drifts out of alignment every week, and the customer will blame the rail manufacturer instead of the beam.
The bolt pattern that works
M5 screws every 300 mm is standard for HRC25 rails. But the screws must be torqued in the right order. Start at the center and work outward. Torque each screw to 5 Nm, then 8 Nm, then re-check the center. If you start at one end and work across, the rail pulls in a curve as the last screws pull it down. The center-out approach minimizes this. It’s a 5-minute procedure but it halves the alignment effort.
The rail follows the mounting surface. If the beam flexes, the rail flexes. Use a stiffer beam or a steel reinforcing plate. Machine the reference surface flat to 0.02 mm per 300 mm. Torque center-out. The “notch in the middle” isn’t a rail defect — it’s the beam deflecting under load.
