Control arm replacement question

[nmlss2006]

I've been reading back and forth about front and rear control arm replacement and I was wondering: we all have read that 'control arms hardware should be tightened when the suspension is loaded and at right height'.

However, I'd like to know: specifically WHAT PARTS have to be tightened when the suspension is loaded and at ride height? I can see it for some, but not for others.

I remember reading a post from psjr (as usual ) that mentioned that the TIS said that in the rear, only the bolt that holds the integral link to the wheel hub has to be tightened with the suspension loaded. The rest, all the rest, could be tightened with the suspension in its fully extended position.

Is that correct? Also, any information about the front, while I go and take everything apart? First time on this particular car, so wish me luck. Armed with P'Blaster, citric acid, phosphoric acid (don't do this at home) and heat gun.. will report back.

I am also going to take the 'easy' way out and remove brake parts and brake dust shields to do the job. Fortunately Santa Claus was nice a few years ago and got me a shiny Maxjax...

[psjr]

Thanks to this post, I took a closer look at that post of mine, and realized it was not correct - I had only looked through the procedures and not the torque spec sheet that I posted. Basically, anything that is not a ball joint end should be tightened under load (as is intuitive).
I have updated my earlier post with that information:
https://www.bimmerforums.com/forum/s...4#post29876524

[nmlss2006]

You have just made me very sad. I wonder if I can approximate the height by putting a floor jack under the hub? Some of the bolts will *not* be reachable with the wheel on.

[ancho]

You have just made me very sad. I wonder if I can approximate the height by putting a floor jack under the hub? Some of the bolts will *not* be reachable with the wheel on.

You can use a jack to lift the corner of the suspension until the body begins to just slightly lift - that's going to be the fully loaded position for the suspension.

[nmlss2006]

Yes, that's what I meant. The problem is, I'd be doing this on a lift, so gauging the 'correct' height is going to take a little ingenuity, especially because my day could turn unpleasent in a hurry if the car decides to take an unexpected excursion. I think I'll have to play with it a bit and measure the hub height.... does anyone have the 'official per spec' ride heights handy? I'll look in the Bentley but I (stupidly) don't have a copy of the TIS handy since I'm no longer running vmware on any box.

psjr, thanks for the info, the only part of your post which I take a small amount of exception of is the 'as is intuitive' part. There is no reason why the bushings should not be made to rotate properly - and a lot of reasons why they should... but of course that would cost money and be useless in the first 50kmi of life of the car so... argh. Don't mind me. I am just having a bad E38 day after chasing a PS leak I CAN'T FIND all of yesterday.

[psjr]

does anyone have the 'official per spec' ride heights handy? I'll look in the Bentley but I (stupidly) don't have a copy of the TIS handy since I'm no longer running vmware on any box.

http://files.psjr.org/alignment/
has alignment specs, how to load to "normal position" and ride height specs.

[nmlss2006]

Thank you sir. I just figured out why there is a requirement for proper ride height before tightening. Unfortunately, the reason why THAT is done is neither clear nor satisfactory, but I am sure that having the torsion of the rubber contribute to the suspension control is key to the car's handling. There of course is no other easier way (...) that this could be accomplished.

Moving right along. in the same 'and now why didn't I think of doing it that way' kind of narrative, I would like to thank (...) the engineers who determined that it was a good idea to have the strut tube end need to be moved to get to the nuts for the control arm and the tension arm. Very wise.

Of course, there is the additional advantage that in order to get the required mobility and not damage the ball joint, it is best to remove the tie rod end. Except that it's almost impossible to do so without damaging the boot - and thus requiring a new tie rod end.

Or do I just disregard? I *really* don't like the angle that the tie rod ball joints are working at. And I'll have to come down at least 1" with the steering knuckle - and 2" would be better, since the control arm ball joints have the allen head spin stoppers, not the slots in the stem for a wrench that would actually give you a sporting chance without going completely nuts.

While I'm here! Is there ANOTHER trick to getting the nut and bolt for the tension strut? The stabiliser bar WILL have to be unbolted but even then it's going to be awfully interesting...

I assume that wisdom dictates (1) cracking the nuts on the ball joints at the steering knuckles (2) removing the subframe ends of the tension strut and control arm from the subframe (3) getting the ball joints out the rest of the way.

All this after removing BOTH the swaybar from the subframe AND the swaybar from the swaybar link.

Did I miss anything?

[psjr]

You can get the traction strut bolts out by turning the wheels, so the sway bar doesn't need to be unbolted anywhere to do the arms IIRC.
Yes, remove the subframe ends of the arms before the ball joint ends.

[nmlss2006]

Further question... the traction strut chassis bolt is captive, right, there is no way it's coming out without removing the entire steering pantograph?

And it seems like you have to get the wheel angle EXACTLY RIGHT to have JUST ENOUGH clearance to get it JUST out of the way of the traction strut bushing.

What I think of this particular arrangement is left as an exercise for the reader but it follows very closely the other considerations on the actual workabilty of the suspension.

Yes, I'm grumpy. Bear with me. I found all sorts of posts on bimmerfest and elsewhere about people who have bent bolts and done ALL SORTS of funny things because this suspension works well when the components are new but that's where it stops: it's wear prone AND difficult to work on. Engineering choices, of course. Everything is a compromise.

BTW, I find the idea that someone could even PRETEND to be able to torque any of these bolts to anything better than +/-50% accuracy rather hilarious, but I will take pointers. Maybe I'm missing something again. At least the PS leak mysteriously vanished.

Oh and yes, while I'm in grumpy stream of consciousness mode: unless your E38 is less than 10 years old and has less than 80kmi on it.. or of course you've done it already within a comparable interval... replace the engine and transmission mounts. No, forget the suspension and the 22" rims. Replace the engine mounts and drive the car. The difference in precision and confidence is amazing, you lose the 'two phase steering' feeling entirely.

[canyncarvr]

Something I found useful for traction strut replacement: A set of LONG wrenches. Likely the only time I'll use them, but I would not have wanted to do the job without 'em. HF quality is plenty good enough! I see an 11 wrench set for $25. Use a 20% off coupon and have 'em in your hand for $20...if you happen to have a store nearby.

I think you're right. Torque specs aren't going to be possible without some esoteric tool. Get the feel of the value you're looking for by using a similar-length handled wrench on some easily accessible fastener like a wheel lug. That will get you pretty close.

Yes. You have to get the wheel angle 'just right' to get the bolt out of the way (not OUT, but out enough).

[nmlss2006]

So far so good... the only ADDITIONAL problem that I now have is how to get the tension strut bolts back in. WHY did they not put conical heads on the bolts?

Any ideas? Recommendations?

[canyncarvr]

This not helpful, much: It takes a bit of wiggling and fussing, but they'll go in.

So, my recommendation is to wiggle and fuss with it. It took a good bit more W&F than I figgered it would.

(based on my doing it..but only once)

Good luck.

[nmlss2006]

I wiggled and fussed quite a bit last night. Will do so some more this morning. I'm tempted to yank out the ball joint end, TBH. Get the frame end in, then straighten out the wheels and get the knuckle end in. Easier to.. wiggle and fuss that way.

[nmlss2006]

OK, so.

I will make a howto about this, because there's little information around and most of what is out there is actually not correct - or at least not the best way.

A few basic pointers:

Pre: car on lift, wheels pointing straight. I assume you're not touching the steering, though replacing the tie rod and the center link and thus having them out of the way does make things easier. Be mindful that you will be using all the wrenches that ISO says shouldn't exist, i.e. 16, 18, 21 and 22. You need at least 2x combination wrenches and one socket for each of these. Oh and hilarity will ensue when you realise that your german swaybar links for your german car... use a 5/8 nut. 16? Nope. You will need breaker bars as well as ratchets for the sockets.

1) Disconnect swaybar links at swaybar. If replacing, remove from strut and replace WITHOUT connecting to swaybar until the end of the procedure. Lift swaybar to gain access for (2) and following.
2) DO crack the nuts on the ball joints at the knuckle end
3) Right after this, undo the pinch bolts and lower the knuckles carefully about 2". Tighten pinch bolts back to avoid sadness.
4) At this point, remove nuts from ball joints at knuckle end
5) Use ball joint separator and remove ball joints from knucle end. NOTE THAT AT THIS POINT THE FRAME END OF ALL THE CONTROL ARMS HAS NOT BEEN TOUCHED.
6) AFTER freeing the ball joint ends of the control arms and tie rods (if doing those as well), loosen bolts for control arms.
7) Remove control arms
8) If applicable, loosen ball joints of center link and lift center link from idler arm + steering arm.
9) If applicable, remove tie rod ends and slide out center link.
10) If applicable, remove idler arm.
11) It MAY be easier to remove the swaybar brackets at the subframe to gain further room to work. That depends on what tools you have and whether you're touching the steering.
12) If steering is removed, remove traction strut bolts and traction struts. Be mindful that YOU WILL KNOW FRUSTRATION if you have anything to do with these bolts while the end of the traction strut is still in the knuckle so DO NOT REPEAT NOT do that.
13) If steering is not removed, you will have to move steering all to one side (via steering wheel) to gain access to the head of the bolts up in (12). Be mindful that the bolts will be captive. Also be mindful that steering needs to be at max otherwise you will not have enough clearance for the bolts. Remove traction struts.
14) Clean everything, including where old ball joints were. Be mindful of almost certain rust inside of swaybar brackets, both on brackets and on frame. Fix if needed.

When reinstalling, mostly reverse of removal but there are a few further considerations:

1) Read 12 above. Read it again. Read it a third time.
2) In consequence of (1), if the steering is out, first insert traction strut heads in subframe, thread bolts and put nut on bolt WITHOUT tightening to torque. If steering is not out, insert traction strut heads and slide the captive bolts in, then put nut on etc.
3) The ball joints have personalities. It pays to subsequently put in (or have in) the steering parts, then put the traction strut ball joints in first. Thread the bolts on them but do not tighten yet.
4) After this install lower control arms. You can put the ball joint end in first, thread but do not tighten the bolt, then put the subframe head into the subframe. Thread the bolts into subframe + arm but do not tighten.
5) Now the knuckles are fairly under control: tighten lower control arm ball joints while relying on the rest of the parts to help you hold the subframe still. These usually have allen slots in the ball joint stem so you are tightening the nut vs. force on the stem, less external help is needed.
6) Once this is done, undo pinch bolts, move knuckles back up to position being mindful to rotate the strut properly, put pinch bolts back in and tighten.
7) Now tighten traction strut ball joint nuts. Box wrench will be fine. If ball joint stems do not have allen socket (one of mine did and the other did not - both Lemforder, new, bought together from same supplier) use jack to compress ball joint to knuckle, thus preventing stem from turning.
8) Install swaybar links to end of swaybar. If you had previously removed the brackets for rust/to give yourself more room, you need to put them on before putting the links in. You can do this after (2) or do it right before putting the links in.

Finally, get wheels to ride height by preferred method and tighten subframe bolts for lower control arm and traction strut. Be mindful that this is not without risk if you're using a lift (or jack with jackstands). The only really safe way is to install the wheels and have a pit available.

Put wheels back on, ride five miles, go back on the lift and check that everything is tight, including wheel nuts. Raid the cellar for any and all available alcohol.

Deviation from this order will cause you to take a lot more time and use a lot more colorful metaphors. Ask me how I know.

[xraytube]

I have replaced all of control arms, front and year, and center link/tie rods over the past week or so. I did not replace the front traction struts as I did those 2.5 yrs ago. During the process I did find that the left front traction strut is failed. It is a Karlyn, lasted 2.5 yrs. Pretty easy to replace, so later this week when the new one arrives I will change it out. By far the hardest part was removing the bolt on the front rear control arm (edit: rear guide rod, the one that adjusts toe). This one fills with corrosion between the metal sleeve and the bolt. Total pain in the ass to remove. I did have rusted on nuts at the ball joint end. I split the nuts by drilling them. The only arm I tightened under load was the front lowers. I did it by simply putting a bottle jack under the disk and jacking it up to the approximate height that I measured prior to doing the work. I used the distance from the center of the hub to the fender vertical distance. In retrospect, loading the control arms prior to tightening has minimal impact on the ball joint ends. On the frame ends, you load the rubber bushing in torsion somewhat if you tighten unloaded. A few miles of driving on a bumpy road will take that torsion out in almost all cases. It is possible that it might not and you could get failure of the rubber bushing. Seems unlikely. Ultimately, the frame ends of the control arms need to be able to rotate anyway to allow the wheel to move up and down on the suspension.

I did my own alignment using the string method. Work astonishingly well and is quite easy to do.

[nmlss2006]

Well, yes: the ball joint ends are not affected by the geometry, OTOH the subframe ends are. And they do not rotate - by choice, if you look at the parts, the center of the bushing actually has teeth to bite into the steel of the frame and stay still. Do I think this is a good idea? See previous posts. But, BMW did do it this way and we're stuck with it.

[xraytube]

Well, looking at the front control arm(the one I did pre-load), it definitely rotates. It will track with the wheel as it moves up and down due to the road. You are correct, the metal toothed portion does not move. It is the outer portion that allows rotation.maxresdefault.jpg

[nmlss2006]

Hang on: you're saying that the bushing rotates within the arm's socket, so to speak?

[xraytube]

On the front lower control arm, the rotation is taken up by torsion of the rubber bushing. All the other control arms do not have "teeth" so they will seek a lowest strain state when at rest. Thermal cycles and vibration will facilitate the movement. Even the toothed front control arm over time will move with thermal cycles and vibration.

If you don't pre-load the front control arm, the life of the rubber portion will be lower. How much lower is open to debate. The shear strain is only about 2-3% if not pre-loaded. That is pretty small for synthetic rubbers and polyurethanes. For fatigue of the bushing the ideal state is to have zero shear strain at rest. For fatigue this would be a fully alternating strain state or R=-1 (ratio of the max to min strain). If you don't pre-load, the R value will be some positive value and lower fatigue life. There is a pile of margin in the bushing designs and other environmental factors kill the rubber faster than non-ideal pre-load. I pre-loaded because it was easy to do on the front. I did not preload the rear arms. Did not look necessary given the configuration of the bushing sleeves. On my car, only the sway bar bushings and traction strut bushing were bad. Otherwise all of the ball joints were toast which makes sense they take most of the motion.

[nmlss2006]

I understand what you're saying, I wish I were that certain about 1) the fact that the tension struts will actually rotate to the zero-preload position given time and that 2) the service life shortening caused by the preload is negligible.

Since I do not have data on (1) or (2), much as I agree with the theory, I bought a significant amount of tools to tighten with the wheel up to BMW spec ride height (per the file helpfully provided by psjr). As I mentioned in a number of places, I *did not like* doing this because the procedure means asking for trouble even with a good lift... but it's done. Tomorrow I'll start on the rear and we'll see how that goes. I suspect I will not have fun and I STRONGLY suspect that this time the entire rotor + dust shield etc is coming right off.

[racer2086]

None of the bushing ends of these arms move once they're tightened down -- with or without teeth. The bushing life will definitely be reduced if not tightened down under preload. The arms change a good 20 degrees or more going from unloaded to loaded suspension position, so that's how much those bushings would have to twist.

What I do to tighten down under load is I leave the bushing end loose of all arms and lower the car down and let the ride settle briefly. Then I put the front up on ramps, and tighten the bushings down. The ride height is pretty close at that point. I haven't done that with the rear yet...when I did the rear, I did the on the lift trick to lift the hub. Won't do that again.

[xraytube]

When you torque a bolt or nut down you create stretch in the bolt and compression of the interface you are bolting together. The amount of compression is a function of the bolt torque and most importantly the friction factor between male and female threads as well as the friction between the bolt head and nut head and the mating components. In my case, I really do nothing to address that friction. It matters because that friction strongly impacts the the bolt stretch and compression of the interface. In other words, a high friction factor will require a higher torque to achieve a specific bolt stretch. It is not a huge deal because the engineers have built in substantial margin into the design. It is not like these are moon missions where everything is running on the edge to save weight.

In the case of the front control arm with the teeth, those are sort of like a nordlock washer albeit a bad one. If you didn't clean up the interface or replace it (which you can't), the gripping action is compromised. Not completely, but certainly not like new. That one I pre-loaded. Given the vibration from daily driving and the odd pot hole hit at normal driving speeds, I would expect that interface to move. In the case of the rear guide rod (the toe adjusting one), I am pretty sure the sleeve was aluminum in the bushing with no teeth. I has all the same problems of the front control arm made only worse by a smooth interface as well as the differential thermal contraction of aluminum versus steel bolt. When it is cold, the aluminum will shrink more than the steel which reduces the bolt stretch and compression of the interface. Again due to vibration and bumps in the road it will move. it becomes a question of how much movement is bad. In reality, he design margin on the bushing is large on purpose.

The design engineers knew all of this. If you can preload the bushings, do it. If you can't, you will see reduced bushing life, but the ball joint will fail before the bushing regardless.