Troubleshooting Shaft Wear

Hello Forum,

I'm having problems with a rotating shaft which is wearing excessively under its ball bearing support. The shaft is a cantilever layout with two ball bearings, one 6000 bearing in the middle and one 688 bearing at the end opposite the load. The wear is occurring under the 6000 (larger) bearing. The Housing is split with a 6000 bearing seated in one half and a 688 bearing seated in the other half. The shaft is stepped down for the 688 bearing which has a smaller diameter. I have used similar components with a sleeve bearing in place of the smaller bearing before without any problems. Here are some of my thoughts:

Concentricity of the shaft not high enough. During development we had some shafts with large eccentricities between two diameters and the shaft wore out under the larger bearing way too fast. We have changed the manufacturing procedure to have higher concentricity and were originally satisfied with the improvement.

Fit between shaft and Bearing. We have a sliding fit to ease assembly, and this same fit has been used before without problems. Ideally a tighter fit is needed.

Alignment of bearings in housings. Due to the split housing the two bearings may not be perfectly aligned. Since the housings are reinforced nylon they have some ability to self align.

Heat treating the shaft. Until now we have used 36SMnPb14 without heat treatment. Again this has never been a problem. I'm sure heat treatment would help prolong the life of the shaft.

I'm just hoping someone might be able to suggest why this design causes wear sooner than the previous designs, and which improvements will give the most bang for the buck.

Thanks

Look very closely at the worn areas. Analysis can tell you a lot. Is it from abrasion (relative sliding motion) or impact? Which areas are worn and which aren't? Why? See any fractures or cracks? A little detective work can tell you a lot.

My gut tells me that you have a problem with shaft flexing and inadequate bearing alignment. Even if the bearings are in absolutely perfect alignment (which is not possible), they still need to tolerate shaft flexure without causing damage. How much is the shaft deflecting? Once it deflects at all under load if the bearings don't flex with it you have a problem.

I received some parts from the field today and took a good look. The wear is from abrasion and shows small groove-like rings around the shaft. The wear is always stronger on the edge of the bearing on the side where the radial load is applied. The the wear tapers out moving away from load. The first part might tell me that the abrasion is due to creep because of improper fit. The second part points to what you said concerning deflection.

The shaft is 60mm long. Load is distributed between x=0 and x=20mm, Larger bearing is at x=30mm and smaller bearing is at x=58mm. I measure 0.3mm deflection at x=0. Shaft is ø10mm

Since the housing is reinforced Nylon I would assume it allows enough flex, but perhaps not equally. What I don't follow is if the shaft deflects and the bearing doesn't orient itself to this movement, then why is the wear not symmetrical under the bearing? I would assume misalignment would cause both edges of the inside ring to wear against the shaft.

Sounds like you are on a good path. If you can post a drawing showing geometry, loading, and support, you might get some helpful feedback.

[ATTACH=CONFIG]1466[/ATTACH]

was a little off with the dimensions before

I would look the tolerance structures on the shaft and verify conformance with ABEC requirements. Also, you might consider case hardening requirement for the bearing surfaces: See:

[U][B][URL="http://www.engineersedge.com/materials/carbon_steel_typical_heat_treatments_13249.htm"][COLOR=#000000][FONT=monospace]Carbon Steel Typical Heat Treatments Case Hardening Grade Table Chart[/FONT][/COLOR] [/URL]

[URL="http://www.engineersedge.com/materials/heat_treating_steel_13256.htm"][COLOR=#000000][FONT=monospace]Heat Treating Steel Time vs Temperature and Case Harden Thickness[/FONT][/COLOR][/URL][/B][/U]

The shaft wear tells us that the shaft is turning inside the ball bearing race instead of turning with the race. As jboggs indicates, perfect alignment is not possible and you may have some flexural issues.

Try supporting the load with a roller bearing instead of a ball bearing (which allows the shaft to move a bit axially) and position the shaft with a press fit into your ball bearing on the end furthest from the load. Generally a roller bearing can support a larger load for the size.

This is a bit late to be asking, but what is the source of the lateral load on the shaft?

I don't see what is driving the rotation of the shaft, so is the block on the drawing at the end of the shaft a pulley or gear?

[QUOTE=JAlberts;11460]This is a bit late to be asking, but what is the source of the lateral load on the shaft?

I don't see what is driving the rotation of the shaft, so is the block on the drawing at the end of the shaft a pulley or gear?[/QUOTE]

Thanks for everyone's input here. In the drawing there is black spot which is a straight knurl where a nylon gear is pressed onto the shaft. This gear is driving the shaft. On the left end there is a roll onto which the radial and torsional loads are being applied. Bearings are supported by a two-part plastic housing, and axial movement is restricted with a circlip on the outside of the housing, and the nylon gear butting against the bearing on the inside.

The original tolerances were f7 which gave a sliding fit, I am trying j6 which gives a press fit. So far about 500 hrs and still looking good. I noticed that the end deflected 0.3mm with the old tolerances and only 0.1mm with the new ones. I have found that total runout on the shaft is within 0.02mm.

I agree that roller bearings support higher loads for the size, but I think my problem is with movement not load, because the bearing only dies after excessive wear is present. Also we use the same bearing with the old tolerances in another product which is at least 10 years old with no problems. Only difference was a 10mm sleeve bearing instead of 8mm ball bearing on the back. For this reason if I were to change bearings I think I would consider changing the back one to 10mm to eliminate concentricity problems.