A conveyor gearmotor failed at the housing flange after two years. The unit was a 5.5 kW helical inline reducer with a hollow output shaft, mounted on the conveyor shaft with a torque arm bolted to the frame. The crack ran from the torque arm lug across the housing. The customer blamed the gearbox. The manufacturer pointed at the torque arm. Both were partly right — the gearbox was fine, and the torque arm was installed the way everyone installs them, which was the problem.
Why a torque arm is a load path, not a mounting bracket
A hollow-shaft gearmotor transmits torque to the driven shaft through the shrink disc or keyed bore. The reaction torque — equal and opposite to the output torque — has to go somewhere. If the gearmotor is bolted directly to a machine frame, the reaction goes through the housing feet. If it’s a hollow shaft on the driven shaft, the housing wants to rotate around the driven shaft, and a torque arm stops it. The torque arm carries the full reaction torque. The force on the arm is the output torque divided by the arm length. It’s not a small force.
For the failed conveyor: output torque 1200 Nm, torque arm length 200 mm from shaft center. The force on the arm is 6000 N — half a ton of force, pulling on the housing lug every time the conveyor runs. The lug was bolted to the frame through a rigid bracket with a solid bolt. No compliance. The entire 6000 N went into the housing as a bending load at the lug, on top of the torque it was already carrying.
The mistake: rigid mounting with no compliance
The torque arm needs to absorb the mounting tolerances and the small housing movements — the gearmotor housing rocks a fraction of a millimeter as the drive train loads and unloads, and it breathes thermally. A rigid bolt-through bracket transmits every one of those movements into the housing as a bending stress. The stress cycles at the lug every cycle of the conveyor. Two years of cyclic bending cracked the housing.
The fix that the manufacturer recommends and most installs ignore: mount the torque arm with rubber bushings or a flexible clevis at both ends. The bushings allow the ±1 mm of housing movement without loading the housing. The arm still carries the 6000 N reaction — the bushing is stiff in the load direction — but the housing no longer sees bending from mounting error. The replacement gearmotor with bushed torque arm has run five years.
Getting the torque arm length and alignment right
The torque arm force is output torque divided by arm length. A longer arm means less force. A 200 mm arm on 1200 Nm gives 6000 N; a 400 mm arm gives 3000 N. But a longer arm also has more leverage to twist the housing if the arm isn’t perpendicular to the reaction direction. The arm should be perpendicular to the line from the shaft center to the anchor point, and the anchor point should be chosen so the arm runs roughly at 90 degrees to the gravity vector when loaded. The manufacturer’s torque arm kits come with the right geometry. The improvised angle-iron brackets that plants build usually get it wrong — either too short, misaligned, or rigid.
The misalignment that kills: the arm anchor point is on a different plane than the output shaft center, so the arm pulls at an angle. The pull has a component that bends the housing axially, not just in the reaction plane. The housing crack on the failed unit ran exactly along that path — the arm was pulling slightly forward as well as sideways.
The check on an installed unit
Look at the torque arm on any hollow-shaft gearmotor in the plant. Three things. Is there a rubber bushing or flexible joint at either end? If the arm is a solid bolt through a solid bracket, that’s the risk. Is the arm roughly perpendicular to the reaction line, and is the anchor on the same plane as the shaft? Can the housing move a millimeter without the arm resisting it? If the answer to the first is no, add bushings — it’s a $30 part and an hour of work. The alternative is a cracked housing and a full gearmotor replacement at $1500 plus downtime.
When direct mounting is better than a torque arm
For gearmotors with a solid output shaft (not hollow), direct mounting on a rigid frame through the housing feet is standard and fine — the feet are designed for the reaction. The problem is specific to hollow-shaft units, where the housing is meant to float on the driven shaft and the torque arm is part of the design. If the application allows it, a flange-mounted gearmotor (the housing bolts to a machine face concentric with the driven shaft) eliminates the torque arm entirely — the reaction goes through the flange. That’s the cleanest solution for new designs. For existing hollow-shaft installs, the bushed torque arm is the fix.
A torque arm carries half a ton of reaction force and, installed rigid, turns it into a bending load on the housing. Rubber bushings at both ends let the housing breathe without shedding the load. Check every hollow-shaft gearmotor in the plant — solid bolt-through arms are cracked housings waiting to happen. The failed reducer wasn’t under-sized. It was over-constrained.
