Failure Analysis Time!

[Dave_G]

Here's a pic of a drive spindle we made for one of our GT customers that failed recently. This part has a lot of time on it including the 24 hour race at Daytona last month. We have concluded what caused the failure but I thought it would be interesting to post it here to see if anyone notices the clues in the pics that would lead them to their conclusions as to why it failed.

Have at it!

Dave

"I know it all, but I can't remember most of it..."

 

[Dave_G]

Here's one more view.

Dave

"I know it all, but I can't remember most of it..."

 

[jeff]

I dunno. Is it something part related or was it caused by something else, like the carrier failed or the person missed a shift or the lug nuts were too tight? That starfish crack pattern is pretty cool. Was there residual stress created from cutting the splines? It definitely gave a little before it snapped - was there a turbo or nitro problem - like a sudden boost of power? There isn't any evidence it started out as just one crack and that it spread throughout causing ultimate failure. So I'm guessing some type of surface stress concentration in the spline valleys and it was all your fault.

Come on come on... I wanna know what broke it!

Aloha

 

[Tyson]

Was there just too much stress from it being on the inside wheel on a circle track? Um, excessive heat caused by braking. um. . .axle was put on the wrong side. jajaja. I dunno.

 

[ORI]

I would say it wasnt splined correctly? Also kind looks like Aluminum?

Off-Road Insider
I know ALL!!!!!!!!

 

[Dave_G]

I'll give one clue. Picture the splined part as being nothing more than a giant bolt. ;-)

Best,
Dave

"I know it all, but I can't remember most of it..."

 

[Dave_G]

Re: I would say it wasnt splined correctly? Also kind looks like Aluminum?

Nope. It was splined correctly. As for material it is made from 4340 steel and heat treated to 48Rc.

Dave

"I know it all, but I can't remember most of it..."

 

[Kritter]

Regardles of whether the splines were cut or forged the stress concentration in splines is quite moderate. The part was abused/over-used or not designed properly and the stress concentration due to the splines developed into small cracks from being over fatigued(too many cycles or too high of torque), that propogated until it could not hold together anymore and spun off, hence the center section has a different fracture pattern then the rest. The fracture pattern tells a story like a tree ring can tell age of a tree. The 24 hr of Daytona and being on a long time make it sound as if the part was not expected to go that long but somebody cheaped out on replacing it and now has to deal with a catastrophic failure. Since it is a race car part I would think that it is designed for a finite life and that life was exceeded from the looks of the pictures. One would think this would have been caught during NDT..

My thoughts anyways


Kris
<A target="_blank" HREF=http://www.animalhousedesigns.net>Animal House</A>

 

[Flea]

the heat treating looks like it might be deeper on one side than the other. maybe it the metal fatigued from constant heating and cooling? just a wild guess

GOD BLESS AMERICA! and hopefully my poor truck too.

 

[Dave_G]

Kris,
Your real close here so stop and ask yourself which way the part cycle` and why did it cycle that way? The other drive spindle which has just as much time on it is in perfect shape. Remember, this thing is nothing more than a big bolt. ;-)

Dave

"I know it all, but I can't remember most of it..."

 

[Dave_G]

Re: the heat treating looks like it might be deeper on one side than the other.

Nope. It's even.

Dave

"I know it all, but I can't remember most of it..."

 

[Kritter]

Ok I will add to above and see if I can get it

The high contact forces between the splined shaft and the coupling caused an (a) extreme tension
cycle(s) on the shaft when trying to displace in or out of the coupling(over a slight bump, lean, etc..). This high tension combined with all the small cracks that were propagating due to fatigue in bending and rotation, caused the material to prematurely "neck",dereasing the effective x-sectional area and causing it to fail as depicted. This shaft was designed for rotation to transfer torque, bending to support the car, and tension from when the shaft displaces in and out and the contact forces between the splines under rotation work against it. Maybe the tension from contact forces was not modeled right during design or adequate assumptions weren't made(doesnt work since the other side is ok), or lube ran low (sounds good to me) causing too much contact force which increases tension while the spline plunges in and out of the coupling, or...assumptions go on forever.

I thought it might be due to a trigger happy pit crew spinning the nut on too tight, but after a Q&D calc.I came up with an outrageous amount of torque needed to cause a tension failure by itself assuming a 2 in. dia, and if it was going to fail due to overtightening it would have failed in the threads not the spline.

The splined end is all distorted from when the the shaft broke and pulled out of the coupling at a high RPM.

That's all I can think of this late without a picture of the assembly...especially considering it's spring break for me and I am not supposed to think about this type of stuff this week.

Are you telling us everything? Did this part fail in regular race conditons or is there more to this failure, like a spin out?


A question for you Dave,


What kind of a correlation can you give between HRC and Brinell? I do not own a materials handbook so I think in terms of Brinell since I can quantify properties from it instead of it being it's own scale like HRC.


Kris
<A target="_blank" HREF=http://www.animalhousedesigns.net>Animal House</A>

 

[BradM]

You're indicating a high tensile load but only the very center of the cross section resembles a tensile failure. There appears to me to be a distinct set of radial shear planes that indicate the failure due to excessive torque or possibly fatigue. Also, the deformation evident in the last 0.50 or so of the splines clearly indicates a torsional failure. If the part failed in tension, there would be more appearance of neck down in the fracture plane. I would say it started as a torsional failure and at some point the tensile load led to a yield of the center section and thus separation.

How about a broader view or an assembly pic?


"The only source of knowledge is experience." - Albert Einstein

 

[John Bitting]

Did the car hit the wall tearing the wheel off??

A man is not finished when he is defeated,
he is finished when he quits.
-- Richard M. Nixon

 

[Waldo]

I am going to hit the wall and tear MY hair out! This is too specialized for my knowledge (or lack thereof).

As my students would simply explain: "IT BROKE - THAT IS WHY IT FAILED!"

BRAAAAAAAAP!

 

[Jerry Zaiden]

Did the part go to mag befor every race? If not, there was a problem that was over looked. Maybe bad prep!

Try a better material; It might hold up longer!

 

[Dave_G]

Re: due to fatigue in bending and rotation

Kris,
You hit it right on the money. To put it simply, the torque on the nut that retains the entire assembly lost it's torque as the spacer between the inner bearing races failed. ( it was made out of s#it...we didn't make it. ) The bearing end play is determined by this spacer not by the nut adjustment like in a production car. There fore the effective spindle diameter is much larger than the splined part itself when torqued correctly. ( as an example, that's why it's a good idea to run a hardened precision spacer between the inner bearing races on a front spindle and torque it up tight. Makes the effective spindle diameter larger.) If the stack up torque goes away all the bending loads in the splined shaft get out of hand. As the shaft rotates it sees the bending load in a rotary motion as you have concluded. Other evidence of this failure is the internal spline is severely worn or tapered towards the inside end. ( I couldn't photo this). This car runs about 800 hp but it's not seeing jumps and whoop de doos like the off roaders.

The thread diameter on the end is a 1" thread and the spline diameter is about 1.210 As for Brinell correlation I don't have a conversion chart here at home but I recall that it will be in the low 400's. I'll see if I can find on online conversion for you. I also have a great web site for materials engineering but don't have it in my favorite places here at home so I'll have to wait till I get to work.

Speaking of work.... spring break? You letting cob webs invade you brain this week? If I was you I'd call some race shop and see if they'd let you come to work for a week so you can get some OJT on some race stuff. It would be better than hanging out with the women and drinking beer wouldn't it?.... NOT!!! ;-)

Good job on your analysis

Dave

"I know it all, but I can't remember most of it..."

 

[Dave_G]

Re: Did the part go to mag befor every race?

I don't know Jerry. I'm not in control of that aspect of the car. In my opinion it should have been but I simply don't know.

Dave

"I know it all, but I can't remember most of it..."

 

[Gabe Lara]

Hmmm, Maybe a new section for "Engineering, Testing and Implementation" on RDC?
Obviously with the information and theories in this analysis, proves it is well above the normal "Shop" topic.

 

[Dave_G]

Kris,
Check out http://www.matweb.com/searchindex.htm

It's a good source for material engineering data.


Dave

"I know it all, but I can't remember most of it..."