Gear failure analysis

Cryogenic treatment for the gears will help solve this. You may have to releif the leading edge of the teeth, as well. Redline oil helps. Sounds like your syncromesh is out of alignment. Check the thrust washers. Bushings and bearings may be out of spec also. Harmonic vibration may be of a concern. Did you modify the ballancers? Is there a pully with a bad bearing within the drive belt assembly? A bad tensioner pully bearing will destroy transmission bearings, at frequency, and cause gear wear and failure. My worn tensioner pully caused the tailshaft bushing to wear in my truck transmission. Do you have more information?

There are an awful lot of assumptions in MOTO's response; they may well be correct, but if I were in your position, I would start from more basic investigation.

Does the fracture face show 'beach' markings? i.e. sequential small crack increments, as the crack progressed across the tooth-root? If so, then yes, it is a fatigue failure - if not, then more likely an overloading situation/incorrect HT/incorrect material/misalignment/etc. Is there any sign of overheating? eg blueing of contact faces? Do contact faces show signs of scuffing?

What is the service/application history? Is this a new design or an established component and application that normally survives well?

If you care to send good close-up photograph to my private email address. I m ight be able to offer more advice - not through open forum.

You said endurance testing, that speaks of fatigue to me. If it's high-cycle you might not see the characteristic marks except perhaps with a good microscope. I'd also be looking very carefully for any residual hobbing marks which could completely knacker your fatigue life. I'd also want to be sure there are no burrs or chips on the edges, chamfer the tooth edges if they aren't already. 'Barrelling' might help to remove burrs, in the extreem, shot-peen the things! (ok, probably not practical on a gear with fine teeth)..

If the endurance testing hasn't thrown up tooth wear as a major problem, how about reducing the hardness?

Someone mentioned incorrect material... It would probably be worth the money to send the failed part out to get the material inspected. Your supplier might have messed something up. If this problem is isolated to a single supplier or batch of parts, that would be a big indication.

If you have visual access to the part while it is being tested, you can try watching it with a strobe. That should help show you if you have any misalignment, meshing, or other problems while running that would increase the force on the teeth.

If this is an automotive application the gears are most likley ground. You might want to check for grinding cracks. This is easy to do using dye penetrate. . I would also do a metallurgical analysis, perhaps you will find non-metalic inclusions.

It might be just a case of exceeding the gear's designed capacity.