Porsche 944 turbo ball joints

The problem

First, the front control arms' weak spot is the ball joint. This plastic-lined, glued-in joint fails. The problem with the original design is that the spring loaded ball has moreforce locating it against the plastic cup than the original ball joints usedby Porsche themselves up until this aluminium based design.

This additional load wears out the plastic prematurely and shortens the life(hence they do not last as long as the old type).

Once the plastic housing is worn the forces on the ball joint (shock loadingdon't forget - many times more than normal loading) bangs through theplastic (which has no bending strength) and forces against the aluminium
which cannot restrain the loads. This opens up the aluminium housing untilthe plastic breaks and the ball joint is useless.

944 Ball joint failure diagnosis

Signs of failure include clunking while turning or over uneven surfaces, and steering vibration or drift. Easiest way to check for health is with front wheels off the ground. Either shake the whole wheel assembly, or use a pry bar [gently] to wriggle the ball joint around. There should be nearly no movement other than normal pivot. The boots should also be inspected for cracks, which can lead to premature failure. This unit is one of the most important safety checks on the entire vehicle... for if it fails, the wheel is now free to move within the wheel well... or not. Naturally driving the car hard increases the wear rate.

944 Ball jointRepair options

There are some options when considering what to do:

A repair kit: either with original parts or uprated items that were used on the Cup cars
Refurbished arms: there are 2 suppliers in the UK

Aftermarket uprated arms from Blazaak, Charlie, Rennbay and Fabcar, e.g

Secondhand parts (a big risk IMHO)
New replacement parts from an OPC (over400 each)

See here for supplier details

I contacted both the refurbishers (by email and fax) and received the following from Hartech. He has responded to some comments that were made on the email list:

The reason for failure is simple the original method of maintaining contact between the ball and the plastic casing is flawed. The spring is like putting a spring on the other side of a big end cap to keep the cap and shell in contact with the crankshaft (with no clearance and the shell being pushed onto the crankshaft journal by the strength of the spring). If the spring was strong enough to do this then the big end shell would only last minutes, as the loads imparted through it to the crankshaft journal would wear it out prematurely. Similarly with a ball joint, if a spring is strong enough to prevent the suspension loads from forcing the plastic casing apart then it produces too much pressure on the bearing face for it to last as long as one fitted with perfect clearance but no contact loads. A better solution is to do a similar thing to an actual big end in which a minute clearance between bearing faces reduces contact load and leaves room for lubrication without pinching the bearing. Everyone knows how pinching any plain bearing face too tightly reduces its service life and the spring does just this.

Scottish repair has an additional greasing facility to help lessen friction. This is not the right interpretation because it can only lessen friction if there was not enough grease present in the other solution. Our ball joint is grease filled and encapsulated and protected by the rubber cap. It has just as much grease as you could get in by pumping grease in forever. Too much grease is also a bad thing and anyway, only after perhaps 15 years could there be a benefit from re-greasing the cup by which time the rubber would have perished or split anyway and if the ball joint at that point is still OK you cannot obtain another rubber cap and would need to replace it so in my view the extra grease nipple achieves nothing

Engineering of the moving parts (the ball joint and housing) is manufactured for Porsche in bulk by a huge ball joint manufacturer to great precision and quality control. Competitors appear to have been trying to re-machine the actual housing and inserts in a small batch production environment. Which is likely to be of the higher production quality standards?

Both ball joints have a steel ball with a plastic separator housing in which it sits and works. On the outside of the plastic in our case - is a sturdy steel machined casing (exactly the same as he has in the ball joint on his older designed pre 86 944 and how long has that lasted?), The original Porsche solution on post 86 models (and other refurbishers) uses the machined out - curved aluminium face of the wishbone to support the plastic insert. The plastic is only 1.5 mm thick and cannot resist loads from distorting it so any loads from the ball bear directly on a small area of the aluminium casting.

The loads in our insert bear on plastic as well but then on the precision-machined steel and impart them to the wishbone across the whole area of the insert in a flat - curved contact area many times greater than the other solution. Not only is the steel better able to resist any damage from this loading and support the plastic by keeping its shape longer, but the accuracy and surface finish is also much superior to that achievable with cast aluminium. Furthermore the very quality of steel compared to cast aluminium is of a much higher level due to the difficulties in controlling the casting process.

The hoop stress created by shrinking an insert into an aluminium housing depend of the interference fit intended but in any case are much less than the point shock loading that would be transmitted by the ball through a plastic housing directly on to the aluminium. This argument is therefore futile. The wall thickness of the aluminium housing is too great to suffer any problems from a carefully controlled light interference fit.

Accuracy is essential and we machine our fit to within .01mm maximum to minimum variation.

The original or competitors solutions do not have a sacrificial seperator and I do not understand this concept. We must also remember that most of the car is built with steel and aluminium contact. Most of the engine, gearbox, steering rack, brakes, suspension etc they seem to work OK to me? We must also remember that our insert is isolated from atmosphere by the rubber seal and grease at one end and the epoxy filler at the other!

There were some suggestions that point contacts caused by minute machining differences between a jig bored hole and a precision machined steel diameter are a problem without commenting upon the issue of huge point loaded contact between the ball and the housing (through the flimsy plastic to the aluminium) in the other solutions which is infinitely higher. So if this is a criticism of our solution thinking the argument through would condemn the others completely.

The hoop stress resulting is infinitesimal and not of any concern and far less than the stress applied by the original or competitors replacements because their plastic inserts bear directly on the alloy housing.

(1) The ball joint shaft is secured to the McPherson strut and so is restrained in a plane along the centre of the strut (there can be no tilting loads applied).

(2) The ball (or insert) cannot be pulled through the top of the wishbone, as the ball diameter is bigger than the hole in the top of the wishbone.

(3) The ball (or insert) cannot be pushed through the bottom of the wishbone because it is bolted to the McPherson strut.

(4) There are no loads imparted to the ball joint or wishbone by the linear action of the suspension (as this is contained within the strut).

(5) The only suspension load imparted is through the anti roll bar when one wishbone is out of line with the other. This load apposes the spring in the std and competitors examples and the spring has to be strong enough to keep contact with the bearing surface despite the load on it.

(6) Steering loads are minimal in the context of shock loading through hitting bumps and obstacles.

(7) The main loads on the wishbone ball joint are imparted by restraining the wheel from moving backwards (through braking and when hitting pot holes and ruts). The other loads are lateral - through cornering but the main loads that are of concern are shock loading through hitting things that the car was not designed to do - imparted through the ball to the housing. Because the ball is curved this imparts vertical and horizontal components to the plastic insert and the wishbone housing (or insert housing).

(8) Shock loading results in much higher forces than normal use and plastic is not a good material to resist shock loading unless it is well supported on the other side (especially when it is only 1.5mm thick). Plastic is OK when it is supported by a medium that resists localised impact loading (in which precision machined steel is much better than machined cast aluminium).

(9) It is either the shock loading that pushes the ball into the aluminium) through the plastic) that creates a void so that there is no support for the plastic eventually breaking up the plastic, or it is the addition bearing load forced on the plastic by the spring, that wears it out prematurely. Both problems exist with the original design and other suppliers solutions but neither exists with ours.

(10) For the mileage expected of such old cars, the period of ownership and their life expectancy, there is nothing seriously wrong with most solutions. However before we felt it was necessary to manufacture our own, we tried others and found that some were too tight or lumpy as we received them, others squeaked through poor machining quality of the replacement plastic inserts and matching hole in the wishbone. Some were too tight, others too loose etc. Since we have been reconditioning, we have had all the oppositions refurbished wishbones returned to us for repair so clearly they do not last as long as the original Porsche wishbone. Meanwhile the ball joint that we use has historically lasted longer than the type used by Porsche in this wishbone and so far has not had a single reported failure. I can only conclude that our technical interpretation of the system as a whole, our reasons for replacing the joint with a more reliable one and our quality control are better than any one elses solution.

When we first carried out this exercise on the 944 wishbones, we concluded that the spring pressure was the main problem and added an unnecessary load to the plastic housing

The 968 incorporated 60% improvements over the older 944's and Porsche (in their wisdom) decided also to drop the spring and go for a fixed clearance as in the older ball joints (and the ones we use).

We did not find this out until long after our re-design as 968's did not need ball joints replacing at that stage.

Now they do and we found out that Porsche must have come to the same conclusion as us some years before - adding further weight to our arguments.

However they still allow the plastic cup to sit inside the machined aluminium and not supported by steel (as in our case) so they still eventually need replacing - but it seems at higher mileages than the previous methods.


They explained that these BJs were originally fitted on Turbo Cup cars, and the aftermarket version was made by an outfit in Germany that went out of business. SSI acquired the rights to make them in the U.S. The balls are the same, but the races are brass, whereas the OEM ones are plastic.

I had them professionally installed in my S2 cabrio and the handling has improved dramatically

Basicly replacements for all the pieces that will wear out on the stock balljoint. You'll also be reusing the other non-wear parts.

First step is to remove your A-arm from the car.

Undo the sway-bar end-link. THen remove the front A-arm pivot bolt, be careful to not damage the steering-rack boots (install that #$&! bolt from the rear next time). The rear pivot should be removed by taking the entire caster-block off the car, leaving the pivot & caster-block on the A-arm itself (this will preserve your alignment settings).

Unclamp the spindle-clamp on the balljoint pin and completely remove the bolt. Reach in with a thin screwdriver from the top and tap out the pin. Now you've got the A-arm removed from the car.

To remove the old ball-joint, carefully remove the rubber dust-cover (don't damage the coil-spring, you'll reuse it). Then fit a LARGE socket over the balljoint. On the bottom side, put a small nut or socket on the bottom cover and squeeze the whole assembly in a vise like this:

Squeeze just enough so that there's a little gap between the circlip and the bottom cover. There's two ends on the circlip angled opposite ways. One end is angled so that when you pry it with a screwdriver, it'll grab and pull out a little. Use a second screwdriver to work your way around and free up that first end of the circlip. Grab with needle-nose pliers and "peel" the circlip off the groove.

Unclamp the vise, disassemble the balljoint and clean off the old grease, take notes on which order the parts came out because you'll be installing in the reverse order.

One thing I did that should help the next balljoint last longer is to install a Zerk fitting in the bottom cover-plate. Drill a 5mm hole in the center and tap for a 6x1mm Zerk grease-fitting. You may need to install the Zerk with a washer to prevent it from protruding out the inside surface.

Ok, time to install the new parts. The top cup has to be pressed into the A-arm. Find a socket that will barely fit into the balljoint tunnel in the A-arm; I found a Craftsman 22mm socket with 30mm outside-diameter fits snuggly. Then push in the top race hand-tight and make sure it's aligned squarely with the A-arm. Put the socket on top of it and squeeze the entire thing in a vise to push the top race all the way into the A-arm (use a block of wood on the upper side of the A-arm to prevent damage from the vice).

Grease the top of the new ball with some CV-joint grease; I've found the blue stuff packaged as "boat-trailer wheel-bearing grease" works extremely well. Push this all the way up into the upper cup in the arm, then follow with the brass lower cup. Then there's a thin spacer ring, followed by the conical spring. Then there's a small dime-sized spacer followed by the cover-plate. 

I found I couldn't squeeze the entire thing together tight enough to re-install the cir-clip with that little dime-sized spacer in there, so I left it out. Squeeze the entire balljoint assembly as shown in the picture above and re-install the cir-clip. I then covered up the cir-clip with some epoxy to make sure it won't go anywhere.

Hook up a grease-gun to the new Zerk grease-fitting you installed on the bottom, pump in some more grease so that it's flush with the upper level of the A-arm. Install the new dust cover boot. Push a small toothpick or precision screwdriver in between the rubber dust cover-boot and the balljoint pin to let the air escape and squirt in some more grease with the grease gun. When the grease starts to ooze out the top, remove the toothpick, clean everything off and you've got a brand new balljoint!

Is Sloan's kit the right one for a Turbo S?

You would need the balljoint that has a 19mm pin instead of the 17mm one. I think you can get those from him as well.