Sheared hub shaft, check the photo!

If you are interested, perhaps I can help with the failure analysis, reverse engineering of a new shaft and manufacture of a new shaft. I'm a metallurgical consultant and have spent my 26 year career performing numerous failure analyses, reverse engineering and related materials engineering issues. More info on my credentials can be viewed at my website www.tiptonmetallurgical.com.

Based on evidence from the photo the primary fracture mode is clearly bending fatigue. Multiple fatigue origins are evident over a large arc of the shaft circumference. It looks like fatigue may actually extend around the entire circumference but the light reflection in the photo makes it difficult for me to draw a definitive conclusion. This is important only with regards to determining whether the bending mode was unidirectional or reverse bending. The % of the fracture attributed to final overload suggest a low to moderate mean static stress. The fatigue mode is most likely high cycle fatigue which means the alternating stresses were below the material's yield strength.

Before manufacturing a new shaft it would be important to know the shaft material, yield and ultimate strength of the shaft, heat treat condition and whether surface modifications were employed such as surface hardening, shot peening, etc.

There are material enhancements and an array of surface modification techniques that can significantly increase the fatigue resistance of a new shaft. I have machine shops near by that can manufacture this shafting. Their base business is land based gas turbines and critical components for the nuclear industry.

If I can be of help just let me know.

Tony

 

Great advice...and offer!

 

Had some free time today to look at this photo again and just wanted to point out a couple of features.

The initial fatigue crack had indeed been growing for quite some time although it is not possible to put an estimate on when it started. It had reached it's critical size and was just waiting for the right loading to cause the final fracture. I can not tell from the picture if the final fracture is torsional shear overload or a tensile failure due to bending overload; not really important.

As I mentioned in my previous post the fatigue progression is due to bending fatigue. This implies small alternating loads superimposed on whatever normal mean stress. There are only a couple of root causes for this senario, one is shaft misalignment and the other is shaft imbalance. Vibration due to resonance or a forced excitation does not line up with the evidence. It is recommended that the source of the bending or imbalance be identified before installing the new shaft or it will just happen again, albeit perhaps not for awhile.

I would recommend inspecting the other shaft although it doesn't follow that if one shaft is misaligned the other would be also. Magnetic particle method is fine if you want to take the other shaft off however a simpler method would be a straightforward penetrant inspection. This could be done with the shaft on assuming the area of interest is accessible. Since these cracks begin on the surface penetrant inspection is adequate.

I'm intrigued by your comment that these types of failure have been discussed before. Is there really a history of fatigue failures associated which these shafts? I have not heard of such. The normal design margin for rotating shafting is huge and alternating stresses are trival.

It would be interesting to know what changes/modifications cause these shafts to become prone to fatigue failure. A survey of failure history and photos of the failures would be a simple first step in forming an understanding.

Just a thought.

Tony

 

Bigger modern tires and wheels have altered the stresses from design loads....

One reason I keep the OEM sizes on my three examples...unless I want to go pimpin' downtown, of course!

 

Hi Tony,

The failures I have heard of have all been with track usage, more specifically with much higher grip racing tires. The vast majority of owners do not track their cars, some do with stret tires, but a few with the need for speed have installed slicks and probably doubled the loading on the shafts.

I would be very interested in your thoughts on the relative effectiveness of various crack detection techniques, and more importantly, the efectiveness of shot pening or other stress relieving measures for the stock shaft.

I plan to go to better tires and continue to track the car and would really like to avoid the wheel passing the car syndrome - really.

If it takes new custom shafts, i will do it.

Thanks,
chris

 

Excellent information .... It's good to see someone with high level technical knowledge contributing (don't go too technical on me , I ain't that bright)

Could the fracture have been started by a car running on poor quality / damaged / worn wheel bearings ? ( I am making a big assumption that the car in question has had previous owners ) ...... we really never know what has happened to these cars in the past no matter how well meaning the intentions of the previous owners.
Some people will drive a car till it drops ....

I can't see misalignment of bearings being the problem really. . . . . . please see attached photo of the rear wheel bearing hub ...this is a big solid bit of metal so it would be difficult to install the axle if they were misaligned. I recently installed new SKF bearings and all surfaces were a light press fit. (hand powered press ....never go powered hydraulic)

Could a fracture have started through a mechanic being extremely rough removing a wheel bearing from the axle? My outer bearings were quite tight. I spent days making a special bearing puller to ensure these bearings were removed with no force applied other than straight up the axis (the bought pullers couldn't access under the outer edge of the bearing).

Curious to hear much more ..............
Image Unavailable, Please Login

 

I forgot ........... (told you I wasn't that bright)

Vibration ! When I first got my car my wheels were running 1/16" off centre ...yes 1/8" total run out. It can happen guys. Some idiot had repetitively painted the wheels on my car , it was obvious that the wheels had been standing vertically while the paint dried. This left a huge amount of paint on one side of where the wheel bolt meets the seat in the wheel.

And yes , the car vibrated badly ........ if this was left for long periods could this affect an axle ?

Alternately could a damaged / badly worn CV joint inflict enough vibration to affect an axle ....... ?

I have given these 2 examples because I think they are distinctly different types of vibration ............. I think

 

o.k.; time to chime in
I was thinking over and over this issue. And the question which remains unanswered for me, is why the shaft broke BETWEEN the bearings, where bending stress is much less than outside the bearings. Most failed shafts I have seen or heard of broke outside the bearings directly at the wheel flange.
Last year I also had suspect shafts. See that thread
http://ferrarichat.com/forum/showthread.php?t=186447

Within this thread post 61 shows a picture, which shows, what can happen if someone mixes up the front and rear wheel bolts. If someone uses an impact driver to tighten those it will apply an incredible longitudinal load onto the bearings and shaft. Who knows about the long term effect.

Best Regards from Germany

Martin

 

Chris,

Regarding non-destructive test methods or crack detection techniques. The common methods are liquid penetrant, magnetic particle, eddy current, ultrasonic and radiographic. The first three are used to identify surface defects and the last two are volumetric methods, i.e. they are used to identify internal defects. They each have advantages/disadvantages and limitations. The method chosen for a given component is generally based on the type of defect you are looking for, the expected defect orientation, material being tested, accessability, geometric considerations, number of like pieces being tested and of course cost. Here would be my recommendations for the subject shaft.

For used shafts-since we are looking for surface oriented cracks you have a choice of liquid penetrant or magnetic particle; Magnaflux is a tradename for the company that manufactures magnetic particle equipment. You could perform the liquid penetrant inspection yourself as it is pretty straightforward and the materials can be bought right off the web fairly cheaply. You will want to go with Zyglo ZL-56 water washable fluorescent penetrant. Simply wipe the penetrant on, let it sit for awhile, wash off with water and hold under a black light. This grade of Zyglo has very high sensitivity and is used for the most critical applications. I used to use it for inspection of jet engine components. You will be able to identify cracks as small as 0.010" deep. You can also have the shafts magnetic particle inspected but you will have to have that done by someone else. The sensitivity of magnetic particle inspection is not any better than liquid penetrant although it does have the capability of finding defects that are slightly subsurface. The cracks you are looking for will be open to the surface so this is not really a big plus. If you do want to use this method be sure to ask for testing facility to use a "wet magnetic particle method" and tell them you are looking for circumferentially oriented cracks, this way they will know what direction the magnetic flux field should be oriented.

For a new shaft-Request that the shaft material be ultrasonic inspected while in the barstock or forged round condition. This will insure that there are no internal defects. Request them to scan both axially and circumferentially using a llogitudinal beam transducer. After machining the shaft may be wet magnetic particle inspected. The magnetic particle is chosen over liquid penetrant for new shafts because I'm interested not so much in identifying cracks as I am defects that might be slightly below the surface. There is one last thing you need to specify to the manufacture of the new shaft and that is acceptance criteria. I can offer some guidance on reasonable acceptability standards if you are interested. Send me an email and I'll let you know under separate cover.

Now for shot peening. Shot peening imparts compressive residual stresses to the surface and these compressive stresses extend into the subsurface for up to 0.014" depending on peening intensity. It is beneficial for increasing the bending/torsional fatigue strength and stress corrosion resistance of all metallic materials; fatigue strength increases of 30-40% is common. It will not increase the monotonic strength of materials, improve axial fatigue strength of materials or improve fatigue strength when alternating stresses exceed the yield strength of the material. Simply put I would shot peen these shafts period. It is the lowest cost method of improving fatigue resistance. Ask the supplier to peen to an intensity of 6-14A using cast steel shot or stainless steel cut wire. Shot peening does deform the surface and the diameter will increase slightly, maybe 0.001-0.002", when measuring across the high spots. For this reason request the supplier to hard mask the spline. Now I do not know how tight the bearing fits are on the shaft but if they are very tight then also ask the peening supplier to hard mask the bearing journals and to "feather" the peening as he approaches the masked areas to prevent a sharp demarcation between the peened and unpeened areas; he should understand what you are requesting.


Hope I haven't rambled too much and this is of some help to you.

Tony

 

I'll apologise in advance for adding another complexity to the issue .... Sorry !

My dad used to be with technical services with BP , Castrol and Shell. The first thing he noted when looking at the shaft photos was the dark brown area at the break. He has seen breaks occur because corrosion has provided a nice starting point .... it is not the cause of a break , it's providing a crack somewhere to start.
It may be interesting to clean the shaft at the break and for the 6mm or so that is dark brown and give it the Zyglo test. Maybe there is a ring of extremely small pinholes in the area ......... does this sound feasible Tony ? Could this have started the crack ?

If the area between the axle and the bearing spacer had no form of protection from rust ( that is -- bone bry ) and moisture has got in maybe it provided some small damage to enable the failure to occur decades later. As Martin said , this is a very odd place for the break to occur.

Dad has a question about the axle photo , , the whole axle looks a little blue / gold coloured , , was it heated at any time after or before the break ??

It sounds like the car has been cared for very well , , and still is , maybe better than most so this is not a neglect problem.

If corrosion is the actual starting point we can all benefit from knowing simply to make sure there is plenty of grease (or anti-seaze) on axles and spacers when bearing replacement is done.
........................... and if we don't our freshly overhauled cars will give trouble for sure in 2039 ! You have been warned !

 

Yes, absolutly

 

I am very happy for you Steve that you are okay and secondly that no extra damage to your car.

 

Makes you wonder how long it was cracked prior to the failure looking at those pics. Also how many other unsuspecting 308 owners are out there in the same boat? I bet lots.

 

Please do not spread any panic. I had my shafts crack tested last year and my car has 72kmls. I have the factory option 7" / 8" wheels and drive my car hard occasionally.
Result of the crack test was, that they are still o.k.
I still suspect, that the case we discuss here is not a 'normal' thing which will occur inevitably, but that the shaft got a damage some time in its past, what induced that failure.

Best Regards from Germany

Martin

 

Tony, that is exactly where you would expect the failure I think. That bearings allow a small amount rotation parallel to the axis which means alignment is less critical....it also means that the load applied by the side loading the wheel can be transmitted through the bearing and is then concenstrated by the necked down diameter causing a fatigure failure. It's a design flaw. If the shaft where held at the large diameter and then necked down at the inner bearing the there would be no stress concentration and failure would not occur.

 

I'll second that !

I think the bearings (SKF) which are high quality but a very basic design will fail long before an axle ....... full test results will be an excellent source of information.

 

sorry, but no. NO design flaw. With this bearing design the load is very well separated regarding axial load ( inner bearing ) and radial load ( outer bearing ). With your thoughts the main radial AND the main axial load would force the outer bearing.

The dimensions of the shafts are also very well designed and on that spot, where the shaft broke, there's just axial and torque stress but no bending stress.
Therefore I stay with my theory, that the shaft was either pre-damaged or that there was a material fault.

Best Regards from Germany

Martin

 

Nope.

I agree with you that the bearing separation is just fine, but it has nothing to do with the failure or the condition I'm talking about. The way the 308 hub/axle are designed corning type load turn the hub into a fatigue fixture and fail the axle in exactly the position you would expect to see a problem. It is a poor design. Run an quick FEA and you you see I am correct.

 

Its to make people think. Its not normal but its happened before so why not be aware of the possibility, especially if one tracks a 308? Thats where they seem to fail.

 

Here’s a very quick FEA (finite element)analysis….it shows that the load applied doing cornering bends the shaft just inside the outer bearing which the shaft at exactly the point the shaft in question failed. The bending then causes fatigue because the shaft is rotating which leads to failure. Design flaw…or more accurately a design which cannot handle the loads applied by modern rubber out on a race track.
Image Unavailable, Please Login

 

Yes, it's sure, that these are the most critical spots which will fail first. But how do you know which loads really occur while driving ? And the difference between stock wheels and sticky tyres ?
I think, the most difficult thing is to determine which loads really occur ? O.k., while doing the FEA I can slowly increase the loads till I get critical tensile figures. But what are the figures of ' real life ' ?
I also had the idea of a FEA in my 'other thread', but I gave up after a discussion with Sean ( luckydynes ) about the unknown loads which occur while driving. In the end I would get a nice, colorful image, but are these figures realistic ? Not worth the time and possibly a can of worms.
Based on experience with mundane cars I still think, that some time in the past a kind of a hard impact pre-damaged the shaft. For instance direct contact of the rim with a curb. I have seen wheel bearings which have been destroyed immediately after hitting a curb with the rim without the soft tire rubber.
Steve, do you know whether your car some times needed a new rim on that side ?

BTW; I'm a freelance mechanical engineer and I'm insulted, when a customer claims a design flaw, when a part designed by me fails after 30 years. O.k.; I hope I'm retired, when the first parts I have designed are getting 30 years old

Best Regards from Germany

Martin