Ball Joint Failure - no longer theoretical

[jaw777]

I built the 818s, chassis number 244, using a 2002 WRX Wagon donor. I have maybe a couple hundred miles on my car and was just getting the engine running better.

I read the two other forums on this topic and they are stale so that is why I started a new one.

I was driving (street) and pushing the car a little bit. Up until now it handled fantastically. I took a 90 degree right turn from good concrete pavement to good concrete pavement and pushed it pretty hard. If I had to estimate the speed I would say I was approaching 30 mph. The lower ball joint pulled out of the knuckle and "sat down" on the wheel. I wouldn't have pushed the car if anyone was around so the other lane was open.

BTW - having one of these towed is a real ***** when the suspension is sitting even lower than designed. Fortunately the car is light so I could lift the front end enough to get the tow truck's dolly hardware under the car.

The pinch bolt seemed to be torqued correctly. The bolt does not seem to be damaged in any way. I still have to take the ball joint out of the control arm, but it does not appear damaged, other than the crown nut riding on the rim.

So has anyone had this problem in the past?

I don't recall if that lower ball joint was a stock part or issued by FFR. It was new because I did not reuse an old ball joint. If issued by FFR does anyone have the part no.?

I am hoping someone has come up with a better design, because I have to tell you I am concerned about putting this on a track. Or I am hoping to discover how I failed to put this together correctly so when I get it right I don't have this issue.

[AZPete]

When you get it apart can you post a photo or two? I'm also concerned. You wrote that were making a right turn but which ball joint let loose? Right or left?

[jaw777]

Sorry, I forgot the detail on the wheel. It was the driver's side (left).

I have attached the pics as requested. Upon further review it does appear that the bolt is somewhat deformed. The bolt holds the ball joint in and therefore it must have had to deform for it to come out. Or the knuckle I guess. It is a 10.9 hardness bolt.

579425794357944

[ssssly]

It doesn't look like that ball joint failed. It looks like it wasn't tight enough in the hub and just fell out.

The bolt looks like the one that tightens the ball joint housing.

If the ball joint and that bolt were installed properly, the ball joint would not be able to come out of the hub unless it snapped or severely bent that bolt.

I am guessing you did not seat the ball joint all the way into the hub and were just holding it in from the pressure of the bolt.

The bolt should not be installed when seating the ball joint.

When seated all the way in, when you install the clamping bolt, the clamping bolt will pass through the indent in the ball joint. Not only providing clamping pressure on the ball joint to hold it in place. But also locking it into the hub.

[jaw777]

Since the pinch bolt holds the ball joint in it would seem to help if the non threaded section of the bolt (I am sure there is a technical term for this) extended across the ball joint. It seems that would be less likely to deform. Unless you had it specifically made for the application I suspect you would have to start with a longer bolt and then trim it if it conflicts with anything else.

A nut on the end of the longer bolt might allow for some over-torquing, but I don't consider that a solution. Squeezing the knuckle more might add a little bit of help, but to be honest the whole design concerns me now.

[jaw777]

ssssly - Thanks. Obviously I can't tell at this point whether or not I seated the ball joint all the way or not.

Based on what I can see at this point it would seem that I couldn't get the pinch bolt in unless the ball joint was seated. I believe the bolt fits in the ridge of the ball joint. Can you confirm this?

If you think the ball joint is still good I'll try to reseat it and put a new bolt in. Not sure how I might want to retest it though.

I consider myself lucky to only be out one rim and a tow bill at this point.

[jaw777]

So after looking at the other side it would appear that I may not have seated the ball joint because it is coming out of its knuckle as well. (or it came out partially during the same event) However, when I reseat the current ball joint it does not take much effort to get it into the knuckle. I can't imagine that I wouldn't have seated it properly. It was a new Raybestos ball joint. I assume Moog would be better, but I'm not sure that changes the situation.

The ball joint is in tension and held in place by the knuckle in compression. The torque is 37 ft.lbs. That is not a lot.

Unless there is other advice I will reseat both ball joints, use new bolts and pins, and drive it around a while and just not push it as hard and see if they stay seated.

[RM1SepEx]

Big safety issue, if in doubt replace the ball joints. You obviously had the bolt loosely in the hole and didn't seat the ball joint fully as it is impossible for it to come out when the bolt slides through the groove in the ball joint. I'd replace the arms too as I'm guessing that the loads deformed the hole, that's why it seats loosely. I just passed 1300 miles on mine today. No issues, I replaced everything with new and am using STI aluminum arms.

[jaw777]

I'm going to order the Moog ball joints and see how they fit. I don't think the control arms are the problem. The knuckles may be bad and that will really suck. I could see one getting rounded out or have an unnoticeable crack, but not both. I suppose it is possible that I did not tighten them down properly. Maybe just missed them (twice).

[Frank818]

It's easy to not seat the ball joint completely, it takes some efforts to push it in completely, depending on the rust inside the knuckle. Hopefully you'll fix your problem.
After totally bending 8.8 bolts while the car rolled 0in, I changed for 12.9 bolts and that's where I discovered that the joints weren't seated properly and I torqued too much.

Don't forget your cotter pins on the castle nuts.

[Junty]

Given the catastrophic event that could likely occur with failure or improper fitment. Why not utilise the precision of OEM (Subaru) parts? I personally would never depend upon non OEM parts for such!

[Bob_n_Cincy]

I have about 1500 track miles and 1500 street miles on our 818. These were new ball joints the came with my aluminum arms.
I have noticed that the ball joint has pulled out about 1/16" to 3/32" by eyeball.
I check the torque on the bolt but didn't do anything else.
Bob

57971

[ssssly]

Since the pinch bolt holds the ball joint in it would seem to help if the non threaded section of the bolt (I am sure there is a technical term for this) extended across the ball joint. It seems that would be less likely to deform. Unless you had it specifically made for the application I suspect you would have to start with a longer bolt and then trim it if it conflicts with anything else.

A nut on the end of the longer bolt might allow for some over-torquing, but I don't consider that a solution. Squeezing the knuckle more might add a little bit of help, but to be honest the whole design concerns me now.

Unfortunately because of the low profile of the top lip of the ball joint, it is totally possible to not seat it all the way in, install the pinch bolt around the bottom side of the ball joint and think all is well.

The ball joint in your picture looks almost new. If it was the ball joint that had failed it would look like a grenade went off inside it.

Moog ball joints are great, so they will do you well, but I don't think your ball joint failed. I run the whiteline ball joints myself. Ne real reason, just always have (maybe I like the yellow boots).

I have rarely seen the OEM balljoints fail themselves. Never seen a Moog or a Whiteline fail.

Clean out the inside of the hub where the balljoint seats really well with a rotary wire brush on a drill. Grease the hell out of the hole, seat the balljoint by hand, and then smack it a few times with a plastic dead blow.

Thread the pinch in just until the threads engage and then try to pull it out.

If it is seated correctly, you will never be able to pull it out by hand.

[STiPWRD]

Jaw, could you post a picture of your ball joint installed in the spindle?

+1 for thoroughly cleaning out the bottom of the spindle where the ball joint goes in. I usually grease that area to prevent rust and help the ball joint slide in. It may take a bit of force to properly seat the ball joint in the spindle before locking it in with the shoulder bolt. But once that bolt is in, you shouldn't be able to pull it apart.

[jaw777]

I took a look at it again this morning after my reinstallation. ssssly is correct that it is not completely seated. I will remove the bolt, clean it and drive it in. It currently looks like Bob's picture above with more of the joint showing. I suppose the concern is that if his is working its way out that others could as well. I will post a before and after picture.

I guess after I get them seated correctly I will just put it up on jacks and look at it every couple hundred miles until I get comfortable.

My mistake was obviously a newbie mistake, but a serious one that could have resulted more consequences than I experienced. I don't recall any emphasis in my manual (been a long time) on the installation of the ball joint, but I would have to rate this as a relatively easy mistake to make with potentially serious consequences.

[Mechie3]

I don't think they would work themselves out with that transverse bolt. In the OEM application the clearance between the bolt and the ball joint is biased in compression thus driving the ball joint in. In the FFR application it's biasing the clearance towards the bottom driving it out somewhat, but it really shouldn't be able to go beyond that.

[jaw777]

I have attached pics of my ball joints now that I think I have them seated and torqued to spec. The passenger side joint had to be driven in. The driver's side went right in and had to be held up into the knuckle until the bolt tightened the joint. I am concerned about that, but it was very clean, probably a result of the sandblasting. Interestingly enough, since the driver's side did not have to be driven in I was able to see how much play was between where the joint would have been "set" and where the bolt becomes the force holding the joint in. The "play" looked exactly like Bob's.

At this point I am guessing that somehow I did not get the joint set and that was the source of my problems. I am not sure how I could have missed that, but I don't see anyway for a seated joint to come out without destroying the bolt or the ball joint itself.

5800758008

[aquillen]

Just wondering if you happened to have ANY pix of your suspension area - you might be able to see if it was fully seated, just to really close the case. Sounds like you're pretty sure it was not down into the socket. As discussed, the bolt and the joint should have been chewed up pretty good to pull it "through" the interference they create with each other in proper assembly config.

FFR keeps revising the manual (I got version Q with my July delivery 818S). Send them a note and suggest they add a clear warning about this item. It happened to you so you carry the most weight on the suggestion.

[billjr212]

I'm not sure I agree with the assessment that FFR should somehow add this to the assembly manual. The failure point in this instance appears to be between 2 parts that would have been assembled on the donor car that FFR never tells the builder to separate in the first place (unless I'm missing it). Most do anyway given that this is a good point to perform this bit of routine(ish) maintenance. Starting down this course, we could essentially say they need to reproduce the entire relevant sections of the Subaru factory service manual into the assembly manual.

All that said, I think this thread is a valuable reminder and I'm glad things turned out safely for the OP. It is good to have another 818 in Illinois!

[longislandwrx]

Did you do the right thing and get OEM Subaru flange (pinch) bolts?

your original fastener was certainly not oem, your new ones look like they may but it's hard to tell.

do yourself a favor and get the Subaru ones, they are the proper length so you don't have thread on ball joint and plenty strong.

This was discussed in another thread. looking for it now.

[jaw777]

Long Island. I used 10.9 hardness bolts, but I did not realize there was a difference.

I will order them now and replace the bolt. Thx.

I haven't been able to identify the part no. Please let me know if you come across it.

[Frank818]

The failure point in this instance appears to be between 2 parts that would have been assembled on the donor car that FFR never tells the builder to separate in the first place (unless I'm missing it).

AFAIK they also never told they changed the suspension design in a way that the ball joints now face extension instead of compression and that these joints and Subaru design was made for compression, not extension. No matter if you take them apart or not it won't change anything, the suspension design on the 818 remains the same.

[ssssly]

I have seen a few Subaru guys do this.

Disassemble the hub and remove the ball joint.

Then reinstall the pinch screw first. You can seat the ball joint on top of the pinch and think it is in. Torque the pinch and drive away. But the ball joint is only being held in by the torque on the joint, while sitting on top of the pinch bolt, and not actually seated in the hub.

[Will D]

I haven't been able to identify the part no. Please let me know if you come across it.

I recall having to dig for this part number too - it wasn't in the parts diagrams where I was expecting to find it. Double check me, but I think this is it (from my build notes):
Pinch bolt: 901000060

[longislandwrx]

http://thefactoryfiveforum.com/showthread.php?13414-Frank-s-818R-build&p=230665&viewfull=1#post230665

Will is right, that is the correct part number, I think SOA identifies it as "bolt flange"

58026




I have seen a few Subaru guys do this.

Disassemble the hub and remove the ball joint.

Then reinstall the pinch screw first. You can seat the ball joint on top of the pinch and think it is in. Torque the pinch and drive away. But the ball joint is only being held in by the torque on the joint, while sitting on top of the pinch bolt, and not actually seated in the hub.


agreed, if you don't take a close look and see that its a captive ball joint you might assume that un torqued it slides right in to the socket.

Did you check the other side?

[jaw777]

Both done. Alignment redone. Rear control arms weren't quite level so I was experiencing some bump steer.

Bump steer issues are drastically reduced. I am considering the Whiteline ball joints and tie rod ends. Does anyone know if that is considered a reliable fix for the bump steer?

[ssssly]

The white lines are known effective on the GDB chassis.

For 818s it depends on how you have your suspension setup. R height, S height how low did you go?

So they may work for your issue, they may not.

There is a very thorough 818 bump steer thread in the R forum.

[frankc5r]

I am having a hard time getting my donor ball joint out of knuckle. Donor is 2002 WRX so could be pretty rusty. Tried PB and 50/50 acetone/ATF so far.

[Mechie3]

http://forums.nasioc.com/forums/showthread.php?t=1989958

I made my own. Comes out easy every time. I've shared mine with local guys often.

http://i139.photobucket.com/albums/q291/mechie3/a0a94f98.jpg


http://i139.photobucket.com/albums/q291/mechie3/dc77454b.jpg

http://i139.photobucket.com/albums/q291/mechie3/ff96ce1a.jpg

[frankc5r]

Thanks!! I will give it a try.

[Bob_n_Cincy]

I paid $79 for the snap-on tool
https://store.snapon.com/Ball-Joint-Sockets-Socket-Ball-Joint-Subaru-reg--P739363.aspx

After spending $300 of my time.