franck84

Completly agree with you !!! I leave in the south of France and the winter is not very hard so we can ride around in many tracks...
I Will leave you more pictures when the gt3 Will be on 4 wheels !!!

911rox

Franck, the procedure does not require you to check torque after each session at the track, just to install the wheels as indicated before going to the track... I did 500km on track last Thursday and did not touch the wheels once accept for tyre pressure...

Did the expert test the tension on the remaining 3 wheels to determine if they were over torqued?? Did he explain how the aftermarket wheels contributed to your failure?

Regards Chris

911rox

Bump!

Franck, did Porsche give provide any feedback on the likely cause of your failure??

I believe an inspection was scheduled on the 18th??

medpilot105

Yeah. Definitely interested to know what Porsche said.

911rox

Don't want to be jumping to conclusions but considering the silence, most likely what we've previously had formally confirmed... an overtorque or undertorque situation...

GTRR

I like the look of the CL's but this thread just convinced me to keep the 5 lugs! Getting GMG 130's (non-CL!)

911rox

If it ain't broke, don't fix it... Stick with what you've got

Perfectionist

This is my first post here. I am considering buying a 997 Turbo S and discovered info about potential centerlock hub failures. I am also a corning G force fanatic so I started looking into Porsche centerlock design and reading about the rare but serious failures of them. I don't intend to present myself as a "know it all" here but, on the subject of fastener torquing, I have a lot of knowledge. Somebody in this thread nailed the problem. The wheel and hub connection is what carries the load. The hub itself is only there to make sure that the wheel is tightly connected to the hub. In other words the centerlock hub itself is not designed to carry cornering loads. The friction between the wheel and hub mating surfaces carry all loads. All the hub does is keep the wheel and hub together in this setup. This sounds simple in theory. I am going to qualify this statement later. Let's consider the threaded hub as a bolt and what happens when a bolt is torqued is the same whether it's a centerlock hub or a connecting rod bolt. When a bolt is torqued it stretches and bolts (not torque to yield bolts) are elastic. The bolt is designed to always want to return to it's original length and so it holds two parts together with a pulling force. When bolts are properly torqued they get longer. Stretch a rubber band and it always wants to return to it's original length and a bolt acts the same way, the properly torqued centerlock bolt is always clamping the wheel to the hub with a force, it's not just there in a static state. Problems arise with over and under torquing. If a bolt is under torqued the two clamped parts (wheel and hub) will begin to move against each other and the load begins to be carried by the hub rather than the hub and wheel interface. If that goes on long enough then the hub will shear and probably at the root of a thread. (I don't know how many words I am allowed here but if I run out of space I will continue in another post). The part of a bolt that is most prone to failure is the root of a thread. Normally there is a sharp angle there and that is not a good thing for bolts that are subjected to shearing forces (wheel and hub sliding against each other from improper torque). If a hub is going to break, it's going to break at the root of a thread because of improper clamping force. Over torquing stretches the bolt beyond it's elastic design point and the two clamped parts can begin to move against each other because the bolt is not elastic anymore and under torquing does the same thing although the bolt is not overstretched. Either way the bolt can fail by shearing. High performance connecting rod bolts such as those made by ARP have rolled threads rather than cut threads so the roots of the ARP threads have a small radius and that radius helps to prevent failure at that weak point. I don't know how Porsche hub threads are machined but I bet they are cut threads. Maybe the factory race hubs are made differently. Here is a statement that you can take to the bank. You cannot trust torque wrenches and especially torque multipliers to be accurate.
To be continued:

2XIPA

Interesting analysis. I'm looking forward to your comments to come.

FFaust

Good first post.

Interestingly, you speak of rolled v. cut thread: Many people, myself included, have suffered from the shearing/breaking off or wheel studs (in the 5-lug setup). This happened to me 3 times last year, even though the studs were fairly new.

While speaking with the vendor, I was told that a new product was forthcoming (early 2012), but this time, the studs would feature rolled threads. Haven't broken one since.

Perfectionist

I used to work in quality control for a major production engine rebuilder and proper fastener torque was very important to us, so much so that we had a 15,000 dollar computerized torque wrench tester to assure the accuracy of our torque wrenches. All of our torque wrenches were checked weekly and the computer recorded the accuracy of the wrench and stored it by it's serial number. If the torque wrench began to show a pattern of inaccuracy it was either replaced with a new one or sent back to the factory for calibration. How many techs do you know that have their torque wrenches checked on a regular basis? It gets worse. All quality torque wrenches have an accuracy rating and all are not equal. The only type of torque wrench I will own is one that meets aircraft and military specs. Those are the most accurate. Within the mil-spec types there are some that are rated more accurate than others. I am a Snap On believer and of the Snap On wrenches, the most accurate are the electronic models and the old style ones with a round dial and a needle that indicates torque. As far as I am concerned, the "clickers", the micrometer adjust types, are unacceptable due to a high inaccuracy rating. And if somebody thinks that Macs or Matcos are better. all of these and Snap On are made by one of two companies, to the specs. of the tool company. There are two ways to torque a centerlock hub. One is to use a "normal" torque wrench and a torque multiplier and the other is to use a long torque wrench without a multiplier. The stated accuracy tolerance of the Snap On multiplier that is used for a fastener that torques in the 450-500 lb. range is 10%. In other words, if a hub is torqued to 500 lbs., it is possible that is is only being torqued to 450 lbs., with a new multiplier. Add the accuracy rating of the torque wrench to that which might have a 5% tolerance brand new and you have a potential problem. The only way to be assured that the torque wrench/multiplier setup you have is accurate is to check them with an accuracy checker and make a chart listing the true combined readings. Sending the tools back would not work because they can be off a lot and still be within factory specs. The other way to torque the hubs is to use a Snap On TechWrench, (digital with a 2% accuracy rating that will handle 60-600 lbs.). This is 4 feet long and costs over 1,300.00. If I am stuck with centerlocks I am not going to trust any technician to torque the wheels on my car unless he is using this tool, no matter what his reputation is. I am such a fanatic that I have taken my calibrated torque wrench with me when I have had tires mounted on 5 lug cars. I have no use for those I call "The Air Wrench Boys". Being particular and careful has always paid off for me. From experience I have found that the little unnoticed things will many times come back and bite you. Most bolt torque ratings are linked to the use of a certain type of lubricant. Usually something that torques to the specs of a centerlock hub will require anti-seize and not just any anti-seize. This is to prevent thread galling (metal to metal contact). Galling will cause a torque wrench to indicate correct torque even though the bolt is actually not torqued enough because galling causes the threads to lock together before the bolt is stretched enough.
My personal conclusion is that I would not want centerlocks, and, if I just had to have a Turbo S with them because of a rare color or option package, I would be ready to convert them to five lug wheels and hubs. If 5 lugs are good enough for almost all other high end high performance cars they are good enough for me.

Perfectionist

When I was building high RPM race engines the parts that worried me the most were the connecting rod bolts and the valve springs. That was long ago and before the great technology of today. In that era there were no rolled threads on connecting rod bolts, and even though most of the time, if you used the best parts and torqued them right, not many rod bolts broke but some did and then all hell would break loose. As the years rolled by and RPMs increased the weakest links in engines began to "rear their ugly heads". When rolled threads appeared bolt breakage was greatly lessened but not eliminated. That technology came from the aircraft industry. Hub stud breakage has always been a concern of NASCAR builders. They use the finest steels and rolled threads (i think including inside the nuts). All of the studs are of large diameter and they go all the way through the lug nuts. The same setup is used on hard launching drag cars. When the top NASCAR teams get boxes of studs and nuts, they inspect the thread finish with a magnifier, check the pitch on them, hardness test them, and sometimes X Ray them looking for hidden flaws. There are no greater fanatics than the top NASCAR teams. I am going to attempt to attach a pic of studs on a NASCAR Craftsman truck. Notice the radius on the ends for guiding the nut on quickly.

Bob Rouleau

Perfectionist - many thanks for the informative posts!

Best,

Gofishracing

I do not think it is humanly possible to over torque the CL. You would need a multiplier and a pipe. How many ft lbs do we estimate it would take to damage CL's? 1000ft lb? The problem goes very deep.

Perfectionist

The only way to know whether something threaded has been over torqued is to perform some kind of structural analysis test on it such as magnetic particle inspection or X Ray inspection. Yes, over torquing may not be the problem. Under torquing definitely could be the problem. It could also be the thread design, wall thickness of the hubs, or the metallurgy of the hubs. When Porsche issued the service bulletin, I think that they specified higher torque specs for the new parts for some reason. Sometimes highly stressed parts that are made of the finest steels and appear to be strong enough for the job can still fail from slight improper torque. State of the art engine connecting rods nuts or bolts are not designed to be torqued with a torque wrench. They are tightened with a tool called a Connecting Rod Bolt Stretch Gauge. It is a c shaped aluminum tool that has a dial indicator on one end and a point on the other end. The indicator point goes into a machined depression on the bolt head and point on the other end of the gauge goes into a depression on the other end of the bolt. The nut or bolt is tightened with a box end wrench. When the specified stretch for that particular bolt is reached, the optimum clamping for that bolt is attained. As long as the proper lube is used this is a dead nuts accurate way to do it. With an American V8 rod bolt the stretch is around .005-.007. Even though the centerlock hubs appear massive, they are still going to stretch when massive torque is applied to them and the integrity of the wheel/hub union depends on the elastic properties of the threads and that depends on proper torque. Somewhere some engineer probably determined what he thought the torque of the hub nuts should be using some kind of hub stretch measuring device and maybe he got it right but with little margin for error.