A 1.8 kN·m backstop on a cement plant incline was returned as "defective" after 14 months: uneven sprag witness marks, one third of the sprags polished and the rest untouched, and fretting corrosion on the mounting flange face. The bearing itself passed every bench test. The problem was the torque arm — a 180 mm stub welded to the frame at a fixed centre distance, bolted to the backstop's reaction lug with no slot, no spherical washer and no axial clearance.
Backstop bearings rarely fail because the locking elements are too small. They fail because the reaction path was designed as an afterthought. The torque arm is a structural member with its own stiffness, its own thermal behaviour and its own fatigue life, and all three feed back into the sprag contact.
The reaction at the end of the torque arm is F = T / L. Doubling the arm length halves the force on the lug, the bolts and the frame bracket — but it doubles the bending moment the arm must carry at its root and increases tip deflection, which shows up as angular misalignment at the backstop.
A sprag backstop locks only when all sprags contact both races simultaneously. Any angular or radial misalignment of the housing relative to the shaft makes the sprags engage progressively: the first few carry the whole torque, the rest stay idle. Field examinations of failed units typically show 30 to 50 percent of sprags with no contact pattern at all. That single effect can double the local contact stress without changing the calculated torque by a single newton-metre.
If the torque arm is bolted rigidly in all three directions, thermal growth of the shaft is converted into axial thrust on the backstop's internal bearings. For a 300 mm shaft section, ΔL = α·L·ΔT = 11.5×10-6 × 300 × 40 K ≈ 0.14 mm. Restrained, that millimetre-scale freedom becomes kilonewtons of thrust — enough to brinell the internal bearing and, in severe cases, to skew the sprags out of their working angle.
The arm, its bolts and the welded bracket see a fully reversed load every time the backstop locks. A bracket designed for static shear at 2× safety factor can still fail in fatigue at 20,000 lock events if the weld toe detail is poor.
| Parameter | Short arm, L = 180 mm | Long arm, L = 350 mm |
|---|---|---|
| Reaction force F = T/L | 10.0 kN | 5.14 kN |
| Root bending moment M = F·L | 1.80 kN·m | 1.80 kN·m |
| Tube 60×60×5, I = 559,000 mm4 | — | — |
| Bending stress σ = M·c/I | 97 MPa | 97 MPa |
| Tip deflection δ = FL³/(3EI) | 0.33 mm | 2.4 mm |
| Angular deflection at lug | 0.10° | 0.39° |
Note what the table shows: root bending stress is independent of arm length (M = T always), because the force falls exactly as fast as the length rises. What changes is deflection and angular misalignment at the lug — 0.39° on the long arm, which is generally beyond the 0.1–0.2° alignment budget a sprag backstop can absorb without uneven sprag loading. The practical conclusion is not "longer is better" or "shorter is better"; it is that the arm must be stiff enough to keep angular error under about 0.1° at the lug, which usually means a deeper section rather than a longer one.
Add the dynamic case: transient peak of 2.0× raises F to 20 kN and σ to 194 MPa on the short arm, against a typical S355 weld-detail fatigue category of 80 MPa at 2×106 cycles. The bracket, not the bearing, becomes the weak link.
| Series | How it reacts to arm/geometry error | Design guidance |
|---|---|---|
| BS | Standard sprag; sensitive to angular error, integral bearing takes the thrust | Short rigid arm, greased, moderate duty |
| BSEU | Sealed sprag; same alignment sensitivity, better contamination tolerance | Provide axial float to protect seals and internal bearing |
| RSBW | Gearbox-shaft mounted; the reaction lug is a defined single point, alignment governed by the gearbox bore | Use a machined pad face and a fitted bolt; document reaction direction and magnitude |
| FRHD | Large-bore roller holdback; inherently more tolerant of angular error but its reaction forces are large | Size the bracket and anchor bolts for the full reaction with shock, not for nominal torque |
Ask the supplier for: the reaction force magnitude, direction and point of application at rated and peak torque; the maximum permissible angular and radial misalignment at the housing; the axial float requirement; the recommended bolt class and tightening torque; and a drawing showing the reaction lug interface. A backstop datasheet without a reaction-force figure is only half a datasheet.
Q: Is a longer torque arm always better because the force is lower?
A: No. The bending moment at the root stays equal to the torque at any length, and tip deflection grows with L³. What matters is angular deflection at the lug, so stiffness beats length.
Q: Can I weld the torque arm directly to the machine frame?
A: Only if you can guarantee alignment after welding and provide axial relief. Welding after final alignment is one of the most common causes of uneven sprag marking.
Q: How much misalignment is acceptable?
A: Typically 0.1–0.2° angular and 0.05–0.1 mm radial for sprag units, and wider for roller types. Confirm the exact figure with the manufacturer per series.
Q: What is the fastest field check for this problem?
A: Disassemble after the first 500 hours and inspect the sprag contact pattern. If fewer than 80 percent of sprags show even marking, the geometry is wrong regardless of the torque calculation.
A 1.8 kN·m backstop on a cement plant incline was returned as "defective" after 14 months: uneven sprag witness marks, one third of the sprags polished and the rest untouched, and fretting corrosion on the mounting flange face. The bearing itself passed every bench test. The problem was the torque arm — a 180 mm stub welded to the frame at a fixed centre distance, bolted to the backstop's reaction lug with no slot, no spherical washer and no axial clearance.
Backstop bearings rarely fail because the locking elements are too small. They fail because the reaction path was designed as an afterthought. The torque arm is a structural member with its own stiffness, its own thermal behaviour and its own fatigue life, and all three feed back into the sprag contact.
The reaction at the end of the torque arm is F = T / L. Doubling the arm length halves the force on the lug, the bolts and the frame bracket — but it doubles the bending moment the arm must carry at its root and increases tip deflection, which shows up as angular misalignment at the backstop.
A sprag backstop locks only when all sprags contact both races simultaneously. Any angular or radial misalignment of the housing relative to the shaft makes the sprags engage progressively: the first few carry the whole torque, the rest stay idle. Field examinations of failed units typically show 30 to 50 percent of sprags with no contact pattern at all. That single effect can double the local contact stress without changing the calculated torque by a single newton-metre.
If the torque arm is bolted rigidly in all three directions, thermal growth of the shaft is converted into axial thrust on the backstop's internal bearings. For a 300 mm shaft section, ΔL = α·L·ΔT = 11.5×10-6 × 300 × 40 K ≈ 0.14 mm. Restrained, that millimetre-scale freedom becomes kilonewtons of thrust — enough to brinell the internal bearing and, in severe cases, to skew the sprags out of their working angle.
The arm, its bolts and the welded bracket see a fully reversed load every time the backstop locks. A bracket designed for static shear at 2× safety factor can still fail in fatigue at 20,000 lock events if the weld toe detail is poor.
| Parameter | Short arm, L = 180 mm | Long arm, L = 350 mm |
|---|---|---|
| Reaction force F = T/L | 10.0 kN | 5.14 kN |
| Root bending moment M = F·L | 1.80 kN·m | 1.80 kN·m |
| Tube 60×60×5, I = 559,000 mm4 | — | — |
| Bending stress σ = M·c/I | 97 MPa | 97 MPa |
| Tip deflection δ = FL³/(3EI) | 0.33 mm | 2.4 mm |
| Angular deflection at lug | 0.10° | 0.39° |
Note what the table shows: root bending stress is independent of arm length (M = T always), because the force falls exactly as fast as the length rises. What changes is deflection and angular misalignment at the lug — 0.39° on the long arm, which is generally beyond the 0.1–0.2° alignment budget a sprag backstop can absorb without uneven sprag loading. The practical conclusion is not "longer is better" or "shorter is better"; it is that the arm must be stiff enough to keep angular error under about 0.1° at the lug, which usually means a deeper section rather than a longer one.
Add the dynamic case: transient peak of 2.0× raises F to 20 kN and σ to 194 MPa on the short arm, against a typical S355 weld-detail fatigue category of 80 MPa at 2×106 cycles. The bracket, not the bearing, becomes the weak link.
| Series | How it reacts to arm/geometry error | Design guidance |
|---|---|---|
| BS | Standard sprag; sensitive to angular error, integral bearing takes the thrust | Short rigid arm, greased, moderate duty |
| BSEU | Sealed sprag; same alignment sensitivity, better contamination tolerance | Provide axial float to protect seals and internal bearing |
| RSBW | Gearbox-shaft mounted; the reaction lug is a defined single point, alignment governed by the gearbox bore | Use a machined pad face and a fitted bolt; document reaction direction and magnitude |
| FRHD | Large-bore roller holdback; inherently more tolerant of angular error but its reaction forces are large | Size the bracket and anchor bolts for the full reaction with shock, not for nominal torque |
Ask the supplier for: the reaction force magnitude, direction and point of application at rated and peak torque; the maximum permissible angular and radial misalignment at the housing; the axial float requirement; the recommended bolt class and tightening torque; and a drawing showing the reaction lug interface. A backstop datasheet without a reaction-force figure is only half a datasheet.
Q: Is a longer torque arm always better because the force is lower?
A: No. The bending moment at the root stays equal to the torque at any length, and tip deflection grows with L³. What matters is angular deflection at the lug, so stiffness beats length.
Q: Can I weld the torque arm directly to the machine frame?
A: Only if you can guarantee alignment after welding and provide axial relief. Welding after final alignment is one of the most common causes of uneven sprag marking.
Q: How much misalignment is acceptable?
A: Typically 0.1–0.2° angular and 0.05–0.1 mm radial for sprag units, and wider for roller types. Confirm the exact figure with the manufacturer per series.
Q: What is the fastest field check for this problem?
A: Disassemble after the first 500 hours and inspect the sprag contact pattern. If fewer than 80 percent of sprags show even marking, the geometry is wrong regardless of the torque calculation.
english
français
Deutsch
Italiano
Русский
Español
português
Nederlandse
ελληνικά
日本語
한국
العربية
हिन्दी
Türkçe
indonesia
tiếng Việt
ไทย
বাংলা
فارسی
polski