Mark is spot on as always and in line with what I've speculated : this failure comes from a manufacturing defect and it will affect certain cars,not all...

 

You are always so kind my dear friend. Thank you. Mark

 

Just a quick observation on this failed GT3 tower picture. Unless there is an optical illusion, the wall thickness (cross section of the fracture) of the tower on the back side adjacent to the firewall appears extremely thin compared to the front section. It will be great if the guy who had the failure in his GT4 takes some wall thickness measurement all the way along the fractured surfaces (front, sides, back) and share the results with us. It should tell us if we have a fairly uniform casting. Mark

 

I think it's optical. If you follow along the fracture there are 2 additional areas which appear to be thin.

 

No sir, it has the exact same part for the shock tower as the other 981 and 991 cars, as least for GT4, Boxsters, Caymans, 991.1, 991.2, and GT3. But as you will see below, the GT3 and GT4 (maybe others) have a difference in the shock mounting itself (besides the rubber snubber on the non-GT cars) that I believe is a contributor to additional stress on the shock tower.

Quite possibly true, but I believe there are other contributing factors that would cause a random cast part with compromised integrity to be more likely to fail. See below.

First, a request: If anyone is aware of a non-GT 981 or 991 car that has this failure on the front of the car, please post or send me links or info. AFAIK, there have been none, but I haven't been extensively searching.

The first photo is my 981 Boxster S driver's side shock/spring, left side is front of the car. Note the shape and size of the aluminum mounting pad that mates the shock/spring to the shock tower. Note how close it is to the aft end of the shock tower, and it's overall size.

The second photo is my GT4, same side of car and orientation. Note the difference in size and shape of the mounting pad. If you zoom in on the aft (right) end of the pad you can see the flat machined surface of the shock tower casting left uncovered by the GT4 (and GT3) part, and you can get a feel for how far away this edge is from the aft end of the shock tower cavity. The aft end of the mounting pad is a straight line, the curve you see is part of the machined surface of the shock tower's mounting area. This machined area appears to be almost the same shape as the Boxster's larger shock mount.

The shape, size, and location of the GT3/GT4 part is causing more bending moment on the areas of the shock tower where ruptures have occurred. The further away from the vertical walls of the shock tower cavity the spring loads are applied, the more severe the stress is at the failure points we've seen. This is due to the longer lever arm and higher bending moments. Also, looking at the Boxster and GT4 springs and their location relative to the three mounting bolts in the shock tower, I think the GT4 strut is further aft, I suppose to provide more caster? If so, this shifts the center line of spring force to the aft end of the shock tower, relative to the Boxster strut. But the pad location and size relative to the non-GT part is the key focus for causing higher stress on top of the shock tower.

I am not saying this is the cause of any shock tower failures. I am saying there are more stresses on the upper shock tower structures with the GT3/GT4 shock mount.

The next three photos show the areas on the shock tower where the Clubsport brackets are installed, including the Clubsport gusset and white tube that extends forward to the outboard aft portion of the shock tower. After looking at my GT4, I think that white tube is inserted into a steel gusset that has a through bolt going vertically into the hole shown in the center of the 3rd photo, just left of the 90 degree bend in the aluminum air conditioning tube. This hole is part of the shock tower casting. And I'm supposing that steel gusset is also welded to the high strength steel that runs horizontally left and right just above that hole.

The 4th photo is just for orientation reference.

The 5th photo shows another bolt into the forward part of the shock tower casting holding some electrical connector in place. This bolt together with all the other bolts shown in the Clubsport photos in this thread could be used to assist holding down a metal hat machined out of steel or aluminum billet and bolted and bonded to the top of the shock tower for reinforcement. If someone so desired.

 

Well done Sir. Thank you so much for sharing the pictures. These are very telling photos. After reviewing the photos, I agree completely with you, the design and the shape of the flange of the GT4 shock could very well be a contributing factor.

I am really surprised the engineers did not properly utilize the entire seating area of the flange on shock tower. Had they done that, they would have reduced the overall stresses, but more importantly they would have allowed the stresses to distribute more evenly.

Furthermore I really can't tell for sure from the pictures if the mating flange seating surface of the shock tower has a machined surface, or is it just a cast surface. If they used an investment casting to produce this part they could retain reasonably close torleances, and allow them to use the cast part as is without machining, which would most manufacturers do to cut cost. However the cast surface will never be as flat as a machined surface and normally that would not be a big deal if the seating area is large enough to distribute the additional loads of having some high points between the two mating surfaces. All I am saying if this surface is not machined, this could lead to yet larger problems since the current design does not allow full seating over the entire flange.

Again great observation. Well done. Mark

 

I am guessing that the failures we have heard about also had aggressive alignments. Would shifting the plate at the top of the spring also add to the stresses the casting sees?

 

i don't think it would make a difference within the movement of the slot towers. Adding shims does theoretically does lengthen the control arm and therefore increase the leverage. I'm not sure that would increase the likely hood of a failure? carl

 

What section of which highway is this so we can avoid it?

 

It was on 280, between the turnoff to 101 and the turnoff to 880. I have always been uncomfortable about that section, as it is bumpy. The left lane is the only one I have used.

 

Thanks for taking the time to get comparative pics. Interesting footprints.

I seem to remember seeing the same failure mode on a white 981 Cayman or Cayman S. Went over a curb, and it looked like a hard hit. Pretty sure I saw it here on RL. Did a fast search late last night, but couldn't find it; I'll look some more when I get a bit of time.

Looks like a stamped steel reinforcement—whether factory or aftermarket—that would spread the load on the bottom/inside of the tower might be possible...maybe even one that utilizes those holes to be bolted in and spread the load further. I don't love the idea of modding my car, but this is one mod I would do tomorrow if it was available and the price was anywhere near reasonable. The failure mode is one of those worst-case scenarios when it comes to peace of mind without it happening to you, actual repair costs if it does happen to you, damage to car, damage to residuals, hassle factor at time of failure and during repairs, etc.

 

I posted the link earlier in this thread...

 

ONe has to suppose Porsche fully engineered this but some aspect of either quality control OR the combination of the more agressive GT4 front end (spring rates, ride height, wheel size, brake size/weight, alignment, whatever) are contributing to the failures.

WIll be interesting to see if Porsche engages on this or if they divert to the comprehensive insurance.

 

From appearance with eyeball inspection, and a fingertip swipe of the exposed mating surface on the shock tower, I believe it is machined. Investment casting wouldn't leave the surface I see and feel by fingertip.

I forgot to mention in my post how the GT cars have a monoball mount at the top of the strut with no rubber bushing like is used on the non-GT cars. From the PET pictures, it also appears there is no rubber snubber at the top of the GT shock travel unlike the non-GT cars, but I have to believe there is some type of snubber to prevent harsh bottoming.

I'm looking at some of the aftermarket camber plates to see if they cover the same or larger area as the non-GT car OEM part. I see this as a reasonable bit of assurance the force is distributed as close to the edges of the inside of the shock tower as possible. It would still have a monoball mount with no rubber bushing though. The one made by Tarrett doesn't look to cover as much as non-GT OEM, but maybe more than GT OEM.

 

Moving the location of the installed camber plate will vary the stress distribution in the shock tower casting, but it's a pretty complex analysis. Would be best accomplished with a good Finite Element Analysis (FEA) tool using the actual dimensions, forces, and material properties.

Also occurred to me that Joe's car could have suffered an impact during post-factory handling before he picked up the car at the dealership. This weakened part could have gone un-noticed until the failure occurred.

And as has been said elsewhere in this thread the aluminum casting itself can have sub-par properties from manufacturing. You can find plenty to read about aluminum casting manufacturing process issues with an interweb search.

There are so many GT cars out there that have been thrashed under extreme conditions that I'm leaning on the side of sub-par castings or impact damage unknown to the owner making itself known with a total failure of the shock tower.

With the wheel removed, it would be pretty simple to perform dye-penetrant Non-Destructive Testing (NDT) on the inner surface of the shock tower. It would reveal any actual cracks in the tower. You can buy kits to do your own test. Caveat - with aluminum castings, I think the knit lines and porosity can possibly lead to confusing or false positive results with dye-penetrant NDT. By knit lines, I mean the areas where the molten aluminum comes together when being injected into the die and doesn't quite form a homogenous bond.

Or just stick you head under there with the wheel removed and a really bright light with some good eyeballs and see if you notice any cracks in the area where we've seen the fractures.