Rear anti-roll bar hub brackets re-design with simulation (picture and video heavy)

I decided to fit a new passenger side rear hub after four wheel bearing changes, a driveshaft change, ball joint changes and a track rod end change, all of which have left me with yet another wobbly wheel.



I try to reduce the weight of parts if possible when I refit/replace.

Some you can whittle, some you can't, so I do what I can, and I looked at my anti-roll bar bracket, especially as I will be adding some mass by using rose joint-based links, as a candidate for weight loss.



I looked at the old bracket to see where improvement could be made. This is the part:





When I aligned it on a surface I saw that it was slightly bent and twisted.









I decided to look at the movement of the hub and anti roll bar to see how the lads were imposed, as although the twist was small it was unlikely to have come form the factory like that.

Mindful of Mr Destroyer's mention of double shear modification for his brackets I sat down at the hub and jacked it up and down to see what was happening.



I created, from looking at the kinematics of the bar/link/hub an approximate vector of how the loads were applied as the hub was lifted by the bar's operation and settled on a 2 second application, and an estimation of the weight transfer of the car to simulate loading on a fast corner.

I used this as non-linear x,y,z loads:





I then made up a CAD model (Solidworks if you are interested) of the bracket:





Click to on to see video:







This was then put into a virtual test rig that used the mechanical properties of the various components:





and I ran some FEA analysis on it. I am a very low-level FEA user - CFD is my specialty. Bolt tension, for example, is not included, only that the bolts are rigid in the hub.



Assuming you can see the below video you are watching exaggerated movement - FEA does this so you can see where movement is taking place, it is not intended to show the real magnitude of the movement;



Click on video:





As the result is coherent with what should be taking place, and how a used bracket looks, it got me thinking that instead of trying to seek weight loss I should design a new bracket.



i created a number of variants of a single design, which would meet the criteria of:

It must fit the hub.

It must keep the link position.

It must clear the top arm during suspension movement.

It must deflect less.

It must weight 10% less or better (the standard steel part is 145g)

It must be able to be made from and on things I have access to.



The first design was:



The CAD package, when given T6 aluminum as the working material calculated 123.6g (13% saving)



Click on to see video







A few holes were added(?):



17% saving.



Click on to see video



The holes made the longitudinal web too soft.





The holes were countersunk for weight:



17.6% saving, although the previous result had made it redundant.





The hole nearest the link was removed:



16.8% saving.





The longitudinal web was made deeper:



12.4% saving.



Click on to see video







The web was chamfered for weight as it was not clear if it was adding anything to stiffness:



15% saving.





As mentioned, the videos show exaggerated movement, so the numerical results were plotted:





As the aluminium designs were shown as noticeably better, here is the close comparison:





After analysing the deflections,and considering that the major contributior to anti-roll bar lag is the y axis, followed by z, then x the winner was the 5th design (deep longitudinal web, no chamfer).





Normally at this point in a thread I show a picture of a shiny new part, but I have had time problems and not been able to find time to make them up, but as I am waiting for a plane I thought I would use the time to write the thread.



I'll update it when metal has been cut.

from the last news http://ift.tt/154yuIp

via IFTTT