Good reading while you're in the bathroom...
http://www.theherd.com/articles/swaybar.html

 

Note that it's talking about the Impala..... Although I see the f-car mentioned, and can't really figure out why.....

 

At one time he recommended the fbody front bar for the impala.

http://www.theherd.com/articles/fcarbar.html

 

I just thought it was a good discussion regarding sway bar design. I understand that the 2nd gen front bar fits the Impala. But the basic theory holds irregardless.

Here's another link:
http://www.theherd.com/articles/swaybar2.html

 

question.. do the bends in our stock front bar REALLY make a difference in stiffness? I mean, since the pivot points are the same, (ie, each bend has an equal and oppisite "mirror image" bend) I would think they would cancel each other out.

I am sure there are some difference due to the bar not twisting on a single plane, but maybe not enough to make a diff?

I need to figure out how stiff I need to go on my custom straight front bar :-/

 

Yes. Every bend adds more flexibility than does a straight portion of equal length. How much more depends on the bend radius and bar OD (and ID, if tubular). Since flexibilities add rather than canceling each other, you don't recover stiffness with symmetry. What you do get from the "mirror image" design is an anti-symmetrical deflected shape as opposed to an asymmetrical one. Also note that an asymmetrical deflected shape implies coupling between the roll and two-wheel bump modes.

Norm

 

flexibilty where? in the bar itself or the suspension travel. i may be confused but wouldnt a bar thats formed be stronger than a bar thats straight? an example of what im thinking about is a peice or metal. if its flat its wobbly and bends easy. put a slight crease(form) in it and its many times stronger. would the same not hold true for a bar?

and since we are on the subject of sway bars flexing. a customer wanted bigger bars on his 01 vette. so we put bigger addco bars on the car. the rear bar weighted maybe 3 or 4 pounds(with the insulators). stand it on end and it didnt take much to get it to flex. Trey

 

I've never been fond of Addco products.

 

oh crap thats the stock rear bar i was talking about. the addco piece was much thicker and stronger. sorry.

 

As I used the term 'flexibility' above, it is a structural property that can be viewed as being the reciprocal of stiffness. I did not intend it in the sense that something 'flexes' easily or not. That's just a rough description of the magnitude. Everything has flexibility, from a flat strip of aluminum foil to a one-foot cubic block of steel (to give a wide-ish range).

And stiffness is not something that has a direct correlation with strength, either, so please try to keep them separate.

As you have already found out, antiroll bars can have enough flexibility such that it is noticeable without instrumentation. My point in the earlier post was that if you took that same piece of bar stock and bent it differently that you'd have a different amount of free end deflection under the same load even if you kept the 'arms' the same length and supported it in the same places.

I hope that the Addco piece was also stiffer (less flexible).

Norm

 

oh yeah the addco bars were pretty big.

and yes i think i understand what your saying. thanks for further explaining it. Trey

 

Norm-
Any suggestions on how to determine a starting point for going to a straight bar?

I want a straight bar, equal to a ~somewhere around a 34-35mm(max) standard front bar with bends.

 

Finite element analysis?

I've tinkered with this problem a little bit using a small pipe stress analysis software, which is supposed to evaluate and use bend flexibilities. Without being entirely happy with the results, I might add, so it's not even on my back burner any more. It might just be that my model was too crude, it might be some assumptions made in developing the software, or even some unstated modeling requirement that wasn't met.

It would be nice if one could apply some percentage reduction in overall stiffness for each bend (or symmetrical pair of bends, for that matter), but it's not going to be that simple.

Norm