Loose Bolts on a Reversing Bracket
A 6 kN vertical service load, reversing between zero and full value, is carried 150 mm from the centroid of a bracket fixed to a machine frame by four M12 grade 8.8 bolts on an 80 mm × 80 mm square pattern. The bracket face is painted, and the bolts were tightened with a torque wrench of unknown calibration; the maintenance log records only “tightened firmly”. After three weeks two bolts are found loose, paint is fretting around all four holes, and a 15 mm crack runs from the bracket-to-frame weld toe. The bolts show no elongation or necking. Production wants the machine back in one shift and proposes replacing the M12 bolts with M16 and re-welding the cracked bracket, with no other change. The joint is in single shear, the clamped plates are 10 mm thick, and the faying surfaces are as-painted.
1. Estimate the shear demand per bolt and decide whether the loosening is consistent with joint slip rather than bolt overload.
Read the worked answer
Convert the eccentric load to the bolt-group centroid: M = 6 kN × 0.150 m = 0.90 kN·m. Each bolt is at r = √(40² + 40²) = 56.6 mm, so Σr² = 4(40² + 40²) = 12,800 mm². The moment-induced force per bolt is F_M = Mr/Σr² = (0.90 × 10⁶ N·mm)(56.6 mm)/12,800 mm² ≈ 3.98 kN, perpendicular to r. Direct shear is 6/4 = 1.5 kN. For the worst bolt, combining the components gives approximately √[(1.5 + 2.81)² + 2.81²] ≈ 5.4 kN.
A single M12 8.8 bolt in shear has a capacity of roughly 30 kN or more, so the bolts are not being sheared. Slip resistance, however, is only μ times the clamp force: with about 25 kN preload and painted surfaces (μ ≈ 0.15), that is 3.75 kN per bolt, well below 5.4 kN. The joint therefore slips cyclically, frets the paint, and loses preload - exactly the observed symptoms. No bolt elongation supports slip and loosening rather than overload.
2. Evaluate the M16-plus-reweld proposal and specify a better corrective action, with supporting numbers.
Read the worked answer
Upsizing to M16 does not address the mechanism. If the paint remains and preload is still uncontrolled, slip capacity may barely change, and the larger bolts simply delay the same fretting. Re-welding the toe crack without removing the slip will also re-crack, because the bracket is being flexed every load cycle. The crack is a symptom, not the root cause.
The required clamp force for μ = 0.30 on clean, bare faying surfaces is 5.4/0.30 ≈ 18 kN; applying a factor of 1.3 gives about 23.4 kN, so a 25 kN preload is adequate. M12 8.8 can sustain this (proof load is well above 30 kN). With K ≈ 0.20, tightening torque is T = KFd ≈ 0.20 × 25,000 N × 12 mm ≈ 60 N·m, applied with a calibrated wrench in a cross pattern. Clean paint from the faying surfaces, use hardened washers, and verify preload by torque-turn or ultrasonic measurement on a sample. Only after slip is eliminated should the weld toe be ground, dye-penetrant tested, and repaired with a detail that reduces stress concentration - for example a gusset or a move away from the bracket root.
3. What checks and documentation would you require before signing the machine back into service?
Read the worked answer
I would require evidence, not assurances: a calibration certificate for the torque wrench; a recorded torque value and tightening sequence; confirmation that faying surfaces are bare and dry; and a first-article check of bolt preload on at least one bolt per joint. The weld repair should have a written procedure, dye-penetrant or magnetic-particle acceptance, and a photograph of the finished toe. After one shift of running, re-torque and record any loss, then repeat at 24 hours and one week.
Vibration and relative-slip checks at the bracket should also be logged; a marker line across the joint faces is a cheap slip indicator. The maintenance record should state the root cause (paint and uncontrolled preload causing slip), the action taken, and the revised torque specification, so the next overhaul does not repeat the error. Without these records, a signature would certify work that has not been demonstrated.