Bearings with diameters between 36 and 169 in. typically cost a lot and have long lead times, forcing end users to keep spare bearings on hand or face extensive downtime should a bearing fail. Therefore, it can make sense to replace a bearing before it degrades. Bearings removed this way often make good candidates for repair or refurbishment. But what qualifies a bearing as repairable? How are repairs carried out? And what kind of results should users expect?

Upon receiving a bearing for repair, our company evaluates it and gives the customer a detailed report that suggests one of several approaches. These range from cleaning, verifying clearances, and relubricating the bearing, to replacing major components such as inner or outer races and rolling elements.

After cleaning the bearing, we first check the raceway’s integrity by looking for excessive wear or damage. We also measure the surface hardness and perform advanced ultrasonic backscattering to determine the hardened-case depth. These measurements are critical when we must regrind the raceway because enough of the hardened case must remain to support the anticipated ball and roller loads.

When customers provide precise loads, we can simply calculate the needed minimum effective-case depth (ECD), usually measured as the depth to 50 HRC equivalent. But most of the time, the applied loading and duty cycles are unknowns. In these cases, we approximate the ECD as 11% of the rolling-element diameter. This design rule comes from data we collected from extensive bearing tests and analytical studies. When the ECD meets or exceeds this minimum, the case should withstand the stresses occurring below the raceway surface.

In certain instances, we examine the bearing’s overall configuration, especially when its declining performance seems to come from inadequate design. We then suggest characteristics such as preload, breakaway torque levels, or radial or axial play. This information comes from analytical tools including solid modeling, finite-element analysis, and internal computational programs which model stresses and bearing life as a function of application parameters. When we must replace components, we can do turning, gear cutting, grinding, and induction hardening in-house.

Bearing clean-up and verification operations that don’t affect the hardened case depth should provide a bearing with a useful life equal to that of a new bearing. Bearings needing more-advanced repairs will gain a useful service life close to that of a new bearing. At 40 to 70% of the cost of a new bearing and at 25 to 50% of the lead time, a refurbished bearing can keep end users from spending thousands of dollars and wasting months of downtime.

Machine Design