Though more details are required( photos/spectrum etc) but I suspect sudden fluctuation in load or process parameters may have dislocated bearing in the housing causing high friction / temperature resulted in bearing seizure. 

Also check the greasing schedule for the bearings. 

Hi,

more info: Motor 140 HP & 1200 rpm

the catastrophic failure happened many times , only with the 21316 E/c3 ( 20 rollers) , so as soon as the BPFO appear, some hours later and it is died !

Could the number of rollers { 21316E/c3 (20Rollers) & 21316 CC/c3 , (19 Rollers)}  play a role in the carrying load capacity and skidding characteristics which leads to fast failure ?

thx

Pizo

Did you measure belt tension in N? And compare it with the bearing load limit?

hi

i did not measure and compare, bu it is a good direction to search, but which load exactly should i compare ?

Fap=maximum permissible axial load [kN]

Fr=radial load [kN]

P=equivalent dynamic bearing load [kN]

P0=equivalent static bearing load [kN]

Pm=equivalent minimum load [kN]

****i did not work with these calculations since 20 years ....out of my scope, but i will do my best go through...

Regards

Pizo

You should use some pro tension meter like Trummeter, which has its own calculator for suitable belt tension in N or in Hz. Or measure the force directly at the installation. I think the measured belt force on the Trummeter means just force on one side of the belt so you should multiple it roughly by 2 (or consider angles for more precise value). For the new belt installation, the tension value should also be multiplied by 1,3.

When you get this value, it should be reasonably lower than the radial load bearing values.