Spiral Fracture of Hex shaft 1

[ewalk]

I have a 30mm hex shaft turning at 650 rpm. The torque varies from about 339 ft-lbs to 1357 ft-lbs. Normal running torque should be about 450.

The shaft fails at the lowest torque. The latest failure is continuous spiral with fractures about 1/4 inch apart.

The material is DIN 42CrMo4 or AISI (41400).

I do not believe there is any heat treat on the shaft as I can easily cut it with the bandsaw.

 

[JCJackson]

Is there any components mounted to the shaft in the fracture zone? If there is/are, is there any "coco" staining around the fracture zone. A helical fractute indicates the shaft failed in shear/torsion.

 

[CoryPad]

A helical fracture indicates a BRITTLE torsion failure. This would lead one to believe the shaft was heat treated. Otherwise, a ductile torsion failure would be flat with evidence of rotational strain.

Regards,

Cory

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[rb1957]

an interesting mix of units ...

a solid hex shaft 30mm, would have an area at least 500mm^2 (0.8in^2), and J > 40000mm^4 (0.05in^4)

A torque of 339 ft.lbs = 4070 in.lbs would produce a shear stress of something like 41ksi which should be manageable; mind you your maximum torque would produce a shear stress more like 160ksi, which is probably OTT !

How quickly is the torque applied ?

Can you check the heat treat with an impact test ?

 

[sreid]

For 4140 in the normal condition, isn't the yield stress below 41ksi?

 

[unclesyd]

The shaft could be preheat treated 4140 which is normally around 28-32 Rc which can easily be machined.

I think JCJackson is hinting at a possible answer. I would tend to look for torsional fatigue and sometimes a clue is the staining on the shaft as mentioned.

Can you describe where the apparent ends of the cracks are?

 

[desertfox]

Hi Dragoservice

If your shaft material is weaker in tension than in shear which is common in brittle materials then the fracture you will see is the one you have described.
However you can see the same fracture in ductile materials
which has been subjected to a unidirectional repeated torsion and the reason for this is that a shaft in torsion
as tensile and compressive stresses known as principle stresses which run at an angles of 45 degrees to the applied shear stress, over a period of time the shaft develops fatigue cracks which can only grow perpendicular to
the principle tensile stress and hence the helical fracture.

According to my calculations at 1357lbs-ft the max shear stress on the hex shaft is 86181.7lbs/in^2.

How often are the shafts failing?

regards

desertfox

 

[ewalk]

Thank you for the great feed back.

My failure rate is about .5% at the current rate I'm guessing it is not a matter of if, but when.

I have failures at several loads.

Is there a way for me to estimate hardness? I do not have access to a rackwell tester.

 

[rb1957]

thx ewalk ...

i thought you had something closer to 100% failure rate,

you said that they fail at the low torque load ... are the pieces being tested (controlled loads) up to your high load (which i think is going to seriously over stress your steel) ? maybe the failure initiates at these high loads and only becomes evident when the test has gotten to a low load ??

or are the pieces in service (uncontrolled loads) ? if your high load is reasonably achieved in-service, i think there's a basic strength problem. and maybe there's a loading concern also ?

as for your steel, if you can't do a rockwell test, a shop (maybe a small-scale manufacturer) nearby should be able to help ... it's easy

 

[Cockroach]

CoryPad nailed it...brittle failure by torsion. Good job, one of those purple stars on the way.

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[desertfox]

Hi ewalk

How long does a shaft last when you replace it and how many running hours does it get? looking at some yield stress figures for 4140 and even at 86000lb/in^2
you only exceed the yield stress for one group.
It maybe you have a fatigue problem which is causing the problem.

regards

desertfox

 

[JsTyLz]

For any cyclic torsional loading, durability/fatigue always drives the design size/shape and material selection.

 

[ewalk]

Thanks for the help...

This is a GREAT group.

These parts are failing at about 500 hrs. I don't have good data.

I can not see the typical signs because the parts are rusty or dirty by the time I get them.

We are isolating the shaft to limit any canatlivered bending load. The shaft will style have a slight bending load located close to a bearing.

If the shaft can handle the load,without a cantalivered bending load, then I maybe OK. If not, hopefully the life will increase t a point where my failure rate is much less than the current rate.

This shaft is driven by a 24 tooth # 80 chain sprocket, Is it possible that we are hitting a natural frequency which is contributing to the failure? If so, is there an alternate material that is less prone to this type of failure? Or will I need to play with loading to work around the frquency?

Thank you,
ewalk

 

[desertfox]

Hi ewalk

I was looking only at torsion of the shaft, is the shaft only supported at one end with a load that is acting at the free end to create a bending stress as well as torsion?
It will help greatly if you can describe how this shaft is exactly loaded and its length.
One thing you might be able to do is when you get a failed shaft try putting the two halfs together, if they will go
together it bit like a jigsaw then that indicates a brittle failure and if they don't fit together very well then its usually a sign of a ductile failure its not fool proof but it might give us a clue.
Can you find out from the people supplying you with replacement shafts exactly what condition the shaft is supplied in ie:- hardness, heat treatment etc

Regards

desertfox

regards

desertfox

 

[sreid]

Even if the sprocket is located between two bearings, a chain puts a one sided force on the shaft which has to be reacted by a bending force.

 

[desertfox]

Hi ewalk

In addition to my last request for info can you also tell us where the shaft is actually failing and in relation to the bearing you refer to. If you can post a sketch of the situation that would be very useful.
Lastly are these shafts in a corrosive and or high/low temperature enviroment.

regards

desertfox

 

[ewalk]

The shaft fails about 0.5 inches inside the bearing race. The bearingis race is 2 inches wide. The sprocket is opposite the failure at the bearing. The sprocket is supported between two bearings.

The shaft is in the weather, but not a corrosive enviornment. It could expirence some anarobic fermentation and related corrosion, but minor. Temperature is normal fall temps, 20F to 90F. no obvious correlation to temp of corrosion.

 

[sreid]

Let me make some comments that might help our problem solving.

Ewalk is Agricultural, and the machine is used out doors hence the returned parts are slightly rusted and dirty.

I'll guess that the Hex shaft is supported by Hex ID inner race bearings and the shaft is a slip fit in the bearing race (common in agricultural equipment).

I totally agree with fatigue failure but at 500 hours, the number of cycles may be low (low RPM shaft) indicating a high stress situation.

 

[desertfox]

Hi ewalk

If the shaft is failing inside the bearing housing I assume it is circular in section there what dia is it?
What fit do you have between bearing and shaft?
How long is the shaft between bearings and what distance is the sprocket from the bearing in which the failure occurs?
If I imagine your set up correctly you have a hexagonal shaft with a sprocket on it and it is positioned close to the bearing, were the failure occurs? is the shaft circular in section were the sprocket is attached?
Lots of questions I know but we need this information to try and help.
If you imagine the set up as a simply supported beam the maximum bending moment would occur under the sprocket and in addition if the sprocket is close to one bearing, that bearing via the shaft will have to react the majority of
the bending force on the shaft whilst the one further away
see's very little of the bending force.
By way of example imagine a simply supported 2m long beam with 100N sat in the middle, the reactions at each end of the bearing would be 50N, now move the 100N to within 0.25m of one of the bearings and the reaction increases in the bearing close to the sprocket from 50N to 87.5N and the bearing further away falls to 12.5N.
Now in addition you have the torque that your transmitting with the sprocket and shaft which increases the stress in the shaft due to the shearing reaction of the shaft.
You could try moving the sprocket out towards the shaft centre, this will even out the bearing reactions but the shaft will see more bending stress.
If you can give the info I request I or others might be able to help you further.

Regards

desertfox

 

[desertfox]

Hi ewalk

Forgot to ask did you contact your supplier of shafts to check material condition?

Regards

desertfox