Tesla suspension failure causes, and durability testing?

[glbeaty]

Though there is a lot of FUD surrounding the company, there is also a lot of anecdotal evidence Teslas experience a high rate of suspension failure without a prior causal impact. Reported failures range from ball joint studs, to control arms, to spindles.

There are also stories of owners signing NDAs following these failures. Teslas also have a much higher fatalities per mile driven than other high-end luxury cars (though most of this could be due to owners using autopilot where they shouldn't).

Hypothesized causes for this include:

1) The use of extruded aluminum for some suspension links, with the links being loaded orthogonal to the extrusion axis.
2) Two-piece ball joints with a stud that seems to be (friction?) welded onto the ball. I'd never seen this before.
3) Hydrogen bubbles in cast aluminum parts.
4) The high mass of the cars (the base Model S weighs 4,647 lbs, with the heaviest X weighing 5,531.

For pictures of various links, head over to ebay.

What I'm wondering is:

A) Do these explanations make sense?
B) If I bought some links off ebay (Tesla does not sell new parts to the public), is there some place I could send them to have them put through some sort of standard automotive testing regimen? I'm guessing this would be tension-compression testing?
C) How often do these parts fail on other cars? My motorsport experience makes me believe these sorts of failures are extremely rare.

Thanks for any help.

Disclosure: I'm short Tesla's stock.

 

[GregLocock]

There are few useful generic road loads for vehicle testing. That is, we have to measure the loads directly, or using wheel force transducers, and then cascade them through the entire suspension. Then somebody else turns them into a load schedule for each component. My management would love it if we could somehow predict road loads for a new model without measuring them on a surrogate, in fact there is almost always some group of engineers somewhere around the world working on it. The last attempt I was involved in seems to have collapsed. We do have various ways of ballparking the equivalent static loads at the contact patch so that the initial design can be kicked off, but they aren't a whole lot more sophisticated than the traditional 3-2-1.



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[BrianPetersen]

(A) Who knows? The links show lots of photos of crashed cars and not many of broken parts. I did observe one rusty ball joint. That is indicative of a failed seal.

Tesla's suspension design involves lots of joints. Maybe it's better for ride and handling but it does add more failure points.

(B) I would think that the only type of test that would be informative would be a destructive test ... but there's a reason production parts don't get 100% destructive testing If you knew what the failure mode was, then maybe it would be possible to look for specific warning signs, but those photos look like all manner of breakage. If there are multiple failure modes then perhaps it is simply underdesigned.

(C) not often but it is not unheard of. Honda cars from the double wishbone era frequently had ball joint failures but almost always after high mileage and many years (I know someone that this happened to). One of the side effects of multilink suspension designs is that the weight of the car passes through at least one of those ball joints. That's not the case with MacPherson (which normally has a ball bearing inside the strut mount with only compressive load on it) Rear suspensions on road cars should not be using ball joints.

 

[BUGGAR]

Coincidentally, I'm designing some off-road control arms using water jet cut aluminum plate. I can get blanks cut and drilled for less that the cost of the Tesla bits I saw on e-bay. My regimen is build a control arm and use a huge lever to test to the "guestimated" footprint forces. If the control arm breaks, I order another one out of thicker plate. Then we go for a ride in the desert and try to break it out there. If it still breaks, it can be duplicated in steel plate. I used to use FEM and spend hours at the computer, but "test and break" is more fun and has to be done in the end anyway.
ps: It is highly variable but my control arms, cut and drilled are costing around US$25 per pound. I just send my dxf files to a shop that trucks them out to me next day. (I also have a sculpture garden full of my mistakes!)

 

[Tmoose]

I don't think double wishbone suspensions with the spring bearing on the lower wishbone don't necessarily HAVE to have the lower ball joint, or any ball joint, in compression.

http://corvettestory.com/1963_SAE-paper.php?page=12

Honda'c choice to do so probably had something to do with packaging a driveshaft in there. Although it looks like there is plenty of clearance to the wheel to keep the balljoint CL where it is, and make the balljoint in compression, and thus more fail-safe.

http://repairguide.autozone.com/zne.../i.ytimg.com/vi/8mpMRXWZpOY/maxresdefault.jpg

Yes, my daughter's Honda's ball joint separated shortly after a state inspection. Fortunately at very low speed and just a few miles from home.

 

[BrianPetersen]

https://electrek.co/2016/06/13/tesla-fale-complaints-suspension-nhtsa-keef-wivaneff/

Who is telling the truth?

 

[glbeaty]

Greg, thanks! Sounds like testing wouldn't be worthwhile. Do you have any comment on the use of two-piece ball joints and extruded aluminum control arms?

Brian, those complaints are made by non-owners (likely short sellers). I wouldn't call them fake, but they should be taken with a grain of salt. However there are spats of evidence of failures from Tesla owners, which often seem to quickly be taken down (lending support to the NDA anecdotes). I don't know the truth, which is why I was hoping testing might shed some light on this.

 

[SwinnyGG]

Control arms being cut from extruded profiles is not a new technique by any means. I can't speak directly to friction welding ball studs together, but I'm not surprised if that is how they are often made- it would be a LOT cheaper to cut the stud from suitably sized round stock, cut the ball, then weld them together, than it would be to cut the entire thing from billet.

I am very confident that Tesla did not decide to redesign the ball joint when they needed to buy some. They're not making ball joints in house, they're buying them from a specialist like everyone else does.

Also notice that the linked photo of a failed ball joint stud shows a failure outside the weld. Also notice that whoever commented that 'they know welding and it shouldn't be blue' is wrong, that part looks exactly how I would expect a round friction welded part to look.

With that said- I'm not saying Tesla's suspension components are faulty, and I'm not saying they are not faulty. All this post is meant to convey is that the techniques used to manufacture there parts are not abnormal.

Parts made using industry standard techniques can still fail if they are badly designed, or manufacturing quality is low. One or both of those might be at play here.

 

[glbeaty]

jgKRI, the little information I've been able to find on ball joint manufacturing indicates they're typically one-piece forgings. I have never personally seen a ball join which looked to be welded together. I would never run such a ball joint on one of my track cars, especially if it attached to the same link as the damper (and would thus bear the compression load of the damper bottoming out against the bump stop).

Tesla does a lot of things its own way, with extremely mixed results (they often come first in studies of customer satisfaction, but dead last in reliability). I would not be surprised if they're the only using ball joints like these on passenger cars. They're also very vertically integrated relative to other manufacturers. Though I would be surprised if they were making a part as commoditized as a ball joint in-house, I would not be surprised if they asked a supplier to make them ball joints unlike any others on the market.

 

[BrianPetersen]

I can tell you with certainty that Tesla is buying those ball joints (and other suspension parts) from a supplier.

I cannot speak for how the specifications and manufacturing methods for those ball joints were arrived at, how design loads were established, what sort of validation testing was done on them, etc.

 

[glbeaty]

Brian, that would make sense to me. All of the parts I've heard Tesla makes in-house aren't part of the suspension.

Here's a NYT article of a guy hitting a pothole in his Model S, and having the wheel come off. In order to get it covered under warranty he had to sign an NDA:

https://www.nytimes.com/2016/06/11/business/tesla-motors-model-s-suspension.html

Seems like the only way to arrive at a satisfactory conclusion here would be to study the failures, which I don't have the ability to do. Thanks everyone for the help.