LCA bolt failure?

Jim K
Verde
Verde
Posts: 1708
Joined: Tue Nov 30, 2004 12:10 am
Location: Athens,Greece

LCA bolt failure?

Post by Jim K » Thu Nov 13, 2014 11:03 am

This is something I've never heard happen before and I would like to hear from someone who knows their ...nuts and bolts -material science-wise. Problem: During hard downhill brake testing on my 75 1.8T with the big series Brembo 4-pots, I heard a metallic noise under the front of the car, and a split-second instability. Conditions: very grippy clean pavement, 888 tires. I released the pedal and stopped slowly with no mishap. After a quick look, I went home and had a better look: The rh LCA/chassis front 12mm bolt had lost its nut/washer and the lh one was ready to go -self-locking nuts used. Just the front ones, hmmm. I remember checking all of them with a 19mm wrench before last trackday, 3 wks ago. On the way to the track, I did ~8-10 progressively harder stops on the highway, the last one from 180kph -->40kph. I was very impressed with the deceleration, no comparison to my old brakes with the smaller Brembo red 4-pot calipers, also with 310x28mm diameter discs. However, my current DS2500 pads have ~40% bigger area than the former set. I can see why the front LCA bolts bear most of the load during decel, but stripping threads? :shock: I have used Allen-head 8.8grade bolts, same grade as the 19mm hex-head oem ones (the reason for Allen-heads is no clearance for installing oem bolt heads when using 27.3mm torsion bars). My question to you materials guys: is a 10.9 or even 12.9 grade suitable/advisable for this application or not?
I will stress again, I am talking about eye-popping deceleration with no tire squeal!
Jim K.
kevin
Verde
Verde
Posts: 2762
Joined: Sat Jan 22, 2005 9:09 am
Location: Esher, UK

Re: LCA bolt failure?

Post by kevin » Thu Nov 13, 2014 2:49 pm

Jim I have had these bolts come loose quite often on my old track car which ran 28 mm torsion bars . I always checked before each race and nipped them up . One thing is for sure mine could never come out as they were hex head . I had no trouble putting bars in with these bolts as I put always had the front in first . I have always been advised anything over 10.9 is to brittle .
What pitch thread are you using as you need a 1mm or metric fine .
Thing is the bolt can strip if it come loose first as their is no loading on the one face of the thread . When its in it lose state then it can strip .
Probably no one has experienced this coming out as they don't use Allen heads . Are you still using the large washers ?
If you want to keep Alan head bolts then use two nuts to lock bolt but second nut must be half the thickness . If their both same thickness then the pull load of one cancels the push load of the other .
What's condition of lower arm bushes ? If no good it will act as jerking action under sudden braking . This can deflect box section on chassis .., maybe :?:
Jim as you said you need a nut and bolt expert
Jim K
Verde
Verde
Posts: 1708
Joined: Tue Nov 30, 2004 12:10 am
Location: Athens,Greece

Re: LCA bolt failure?

Post by Jim K » Thu Nov 13, 2014 7:37 pm

Thanks Kevin, good points! I can't imagine any difference between hex/Allen head, but I use coarse thread... I suspect there's something to that...and yes, of course I use the large washers and the (oem) bushes are 1yr old.
Jim K.
PS: 8x15" Compos???
User avatar
Maurizio
Verde
Verde
Posts: 679
Joined: Wed Dec 01, 2004 4:49 am
Location: the Netherlands, 153.1km from the N'ring :-)

Re: LCA bolt failure?

Post by Maurizio » Fri Nov 14, 2014 3:10 am

Stripping threads is strange...
Normally a nut - bolt combo is designed to snap the bolt.

So first I would guess, the nut is not just 8.8 quality.
And I always was under the impression the alfa bolts for this one are 10.9's
I know several guys running slicks, never had problems and they are using 10.9 ...

As for metric fine thread: the inner core diameter is larger , so yes a M12 fine can withstand more force in relation to M12 course. Is just a minimal difference and in practise 9 windings is max of working thread, with good quality of parts..

btw why is over 10.9 too brittle?

Makes no sense to me, bolt doesn't have same stretch % wrt a 8.8 bolt after fail, that's it.

Stretch after fail bolt
8.8 ==> 12%
10.9 ==> 9%
12.9 ==> 8%

But the Rp0.2, where you want to stay under is the factor to work with
8.8 bolt ==> Rp0.2 = 640 [N/mm^2]
10.9 bolt => Rp0.2 = 940 [N/mm^2]
12.9 bolt => Rp0.2 = 1100 [N/mm^2]

-------------------------
So lets take the more scientific approach and look how much force that bolt needs to cope with... and see where we stand when making some assumptions.

So Jim/Others, (I'm lazy) and pass some questions:

- R888, max G forces during decelleration (lets assume 1.5-2G?)
- Weight car
- Distance outer edge wheel to chassis member
- Distance between the two bolts in the LCA
- Length clamped part between the bolt and nut
Banned.. ? ;-) Daily donky.. ==> BMW 118d Image
E36M3 (3.0) Ringtool :twisted: ==> definitely BANNED!

AR 75 TS Ringtool '90, AR Spider 2000 veloce '79
Jim K
Verde
Verde
Posts: 1708
Joined: Tue Nov 30, 2004 12:10 am
Location: Athens,Greece

Re: LCA bolt failure?

Post by Jim K » Fri Nov 14, 2014 9:21 am

Ok Maurizio, in contrast to you, I will be more un -scientific: to enhance rigidity (and complicate matters) I have a flat bar 6mm thick connecting the (front) left/right LCA bolts -a few mm right under the oil pan. I understand as the chassis tries to flex during cornering/braking those bolts are subjected to additional tension than the one designed for. I am adding this info to help your calculations! 8)
I'll try finding answers to the measurements you ask.
Jim K.
User avatar
Maurizio
Verde
Verde
Posts: 679
Joined: Wed Dec 01, 2004 4:49 am
Location: the Netherlands, 153.1km from the N'ring :-)

Re: LCA bolt failure?

Post by Maurizio » Fri Nov 14, 2014 10:37 am

JimGreek wrote:to enhance rigidity (and complicate matters) I have a flat bar 6mm thick connecting the (front) left/right LCA bolts -a few mm right under the oil pan. I understand as the chassis tries to flex during cornering/braking those bolts are subjected to additional tension than the one designed for.
Scientific.... I knew that would get your attention :mrgreen:
Jim, I'm not a "nut case" expert, but i try to help :lol:

I disagree about the extra load due to tying the two chassis members together.
Yes you get more stiffness / less deflection in the horizontal plane of the suspension during braking / cornering. But the amount of deflection of the chassis member doesn't change the forces through to bolt(s).


btw some food for thought.
For who didn't know, nuts can also have strength classes.
8.8 bolt at least nut class 8
10.9 --> 10
12.9 --> 12

Strength M12 course / fine: core diam ==> 9.85 [mm] vs 10.47 [mm]
Which also give a difference in area (10.47^2/9.85^2) = 1.13 is 13% less strength in the bolt core diameter.

This doesn't help you, because you stripped the thread.
Which actually means that the nut was a better quality and not enough thread engaged.
Stripping thread goes linear as its the inner thread diameters difference.
10.47/9.85 = 1.06 so 6% less strength.
Banned.. ? ;-) Daily donky.. ==> BMW 118d Image
E36M3 (3.0) Ringtool :twisted: ==> definitely BANNED!

AR 75 TS Ringtool '90, AR Spider 2000 veloce '79
Jim K
Verde
Verde
Posts: 1708
Joined: Tue Nov 30, 2004 12:10 am
Location: Athens,Greece

Re: LCA bolt failure?

Post by Jim K » Fri Nov 14, 2014 10:57 am

Wait, wait... The bolt thread was stripped on both L/R bolts only in the area where the added stiffener was (ie ~6mm) and the rest was ok. On the right side though, it seems the self-locking nut threads disintegrated and the nut flew off. The plastic locking nut insert was still threaded on the bolt. Therefore, I must conclude there is movement of the stiffener that destroyed the bolt thread and pushed out the nut, even though it was clamped very tight! I will have to check if there are any grade marks on the locking nuts.
I shopped around for M12 15cm long Allen-head bolts today and for lengths over 12cm, they only come in 1.5/1.75mm pitch so no fine thread! :(
I will post pics of the bolts when I replace them.
Jim K.
User avatar
Maurizio
Verde
Verde
Posts: 679
Joined: Wed Dec 01, 2004 4:49 am
Location: the Netherlands, 153.1km from the N'ring :-)

Re: LCA bolt failure?

Post by Maurizio » Fri Nov 14, 2014 11:31 am

I'm not getting the stiffener construction...hence higher forces.
It would help to see some picture to understand it.
Banned.. ? ;-) Daily donky.. ==> BMW 118d Image
E36M3 (3.0) Ringtool :twisted: ==> definitely BANNED!

AR 75 TS Ringtool '90, AR Spider 2000 veloce '79
Jim K
Verde
Verde
Posts: 1708
Joined: Tue Nov 30, 2004 12:10 am
Location: Athens,Greece

Re: LCA bolt failure?

Post by Jim K » Fri Nov 14, 2014 12:15 pm

ok, when I lift it up for new bolts.
User avatar
75evo
Verde
Verde
Posts: 784
Joined: Tue Oct 11, 2005 8:56 am

Re: LCA bolt failure?

Post by 75evo » Fri Nov 14, 2014 11:50 pm

This is my very unscientific and simplistic understanding.

The chassis stiffener is not allowing the frame rails to flex or spread as much. So now the same amount of force is absorbed by the both the front bolt head and nut. The weaker of the two will fail if the force exceeds either of the two's rated strength. My guess is the nuts were not of "men's strength" ( you can ask MD he is an expert on this :) )
kevin
Verde
Verde
Posts: 2762
Joined: Sat Jan 22, 2005 9:09 am
Location: Esher, UK

Re: LCA bolt failure?

Post by kevin » Sat Nov 15, 2014 5:30 am

Jim I think I understand your description of the chassis stiffener your using . I'm sure this plus a possible inferior bolt and large pitch is the problem . This chassis stiffener sounds like it will apply tension that is eccentric ie there is a smaller lever arm apply to the one side of the nut . This will strip it . Put new bolts in with out the stiffener and go test .
I use the sump guard that came on the SA cars as the brace , this bolts directly to underside chassis rails .
Hex heads bolts can't come out if they strip as the hex catches the lip of the lower arm .

Ps : those compos are there but we still can't find a road calliper ie one that has dust seals to fit the replacement rims . I will email you on lack of progress :wall:
Jim K
Verde
Verde
Posts: 1708
Joined: Tue Nov 30, 2004 12:10 am
Location: Athens,Greece

Re: LCA bolt failure?

Post by Jim K » Sat Nov 15, 2014 7:10 am

In essence, we are preventing the frame rails from wandering around and flexing as they please during cornering/braking... Doesn't say much for the chassis does it? :x Now I see why factory racers had the roll cage extending all the way there and around the nose... maybe I'm too fast for my car! :lol:
Here is a drawing of what there is. The stiffener is ~2cm below the pan to allow for ease of engine removal/refitting (tilting etc.)
Jim k.
Image
User avatar
Maurizio
Verde
Verde
Posts: 679
Joined: Wed Dec 01, 2004 4:49 am
Location: the Netherlands, 153.1km from the N'ring :-)

Re: LCA bolt failure?

Post by Maurizio » Sat Nov 15, 2014 9:05 am

So I did understand the stiffener story.
For the bolt forces, it doesn't matter if the stiffener is there.

I will explain, just run through the parts and try to draw the force lines.

Original construction:
The brake forces goes via the LCA through to bolt&nut into the chassis member.

With stiffener:
The brake forces goes via the LCA through to bolt&nut into the chassis member via stiffener to second chassis member.

==> So for both situations all the forces generated by braking go first though the bolt&nut combination.

If we interchange the chassis member by a theoretical spring. By adding the stiffener only the stiffness of that theoretical spring changes. Which minimizes the deflection, the force (still coming from the brakes) stays the same!

And then about the stiffener, we can debate on its efficiency.
Yes adding steel increases stiffness. But when you bend and create a L or even worse a U shape, the stiffness is only really effective along the fold.

Try at home with some paper, fold it and clamp it with your hands and apply a load :mrgreen:

You actually added 2 folds per side. So I´m convinced the piece of the stiffener under the nut isn´t going to strip the thread, because your strip is able to elongate without bending the piece under the nut.

Oooh man what a long story, everybody still awake :wall:
Resume: I still think your bolt was simply to weak wrt the generated brake forces.
Banned.. ? ;-) Daily donky.. ==> BMW 118d Image
E36M3 (3.0) Ringtool :twisted: ==> definitely BANNED!

AR 75 TS Ringtool '90, AR Spider 2000 veloce '79
Jim K
Verde
Verde
Posts: 1708
Joined: Tue Nov 30, 2004 12:10 am
Location: Athens,Greece

Re: LCA bolt failure?

Post by Jim K » Sat Nov 15, 2014 9:50 am

Nice going Maurizio!
I fully agree with your reasoning. While its true any bends lead to deformation (flex) under pull/compression of the two end surfaces, the thickness of the strip keeps flex to a minimum -which cannot be reduced further, as there is no room to make a proper 'rigid' connection. Maybe a thicker strip? :?
I will have more post-mortem info when the bolts are removed. The thread wear pattern may reveal more. For now, I am thinking of a proper practical way to rigidly connect the two rails. The fact that they are hollow does not help... I am open to suggestions for a solution with removable reinforcing parts.
Jim K.
User avatar
Maurizio
Verde
Verde
Posts: 679
Joined: Wed Dec 01, 2004 4:49 am
Location: the Netherlands, 153.1km from the N'ring :-)

Re: LCA bolt failure?

Post by Maurizio » Sat Nov 15, 2014 10:01 am

A thicker strip is not the way to go, its just adding a lot of weight and a minimal amount of stiffness. But as it is a turbo, you can compensate by adding 0.1 [bar(g)] boost extra :lol:

The stiffener is the most effective when, the load is a pure pull force through the material.
A hint, you need a fold in the direction where you need the stiffness :wink:
Banned.. ? ;-) Daily donky.. ==> BMW 118d Image
E36M3 (3.0) Ringtool :twisted: ==> definitely BANNED!

AR 75 TS Ringtool '90, AR Spider 2000 veloce '79
Post Reply