thank you i understand that,the bars that im looking at are measured at say 400 with 5 degrees of twist ,thats why i mentioned 5 degrees

 

he's trying to get someone to do the math for him to compare against these swaybars (or similar).
http://www.1speedway.com/swaybar36.htm

seems just as easy to use the equations posted to back calc the stiffness of the bars in lb/deg.

 

1. Why not just measure it directly? The bar isn't straight so all the math in the world ins't really going to help you. All those bends mean basic torsion deflection equations no longer apply.

Pull the bar and mounts, bolt them up to something with one leg locked in position. Add weight and measure deflection.


Weight is weight. Until you are adding ballast back into the car, hollow swaybars absolutely are not a gimmick. It's 100% sound engineering principles.

Now, the fact there is only one hollow rear sway bar worth a **** on the market for the EVO is an issue. Also, you have no idea if the bar was truly designed correctly so you do get your desired rate. But if you are going to the extent of using something like griceiv posted, then hollow is the way to go.

FWIW, I've got something drawn up already using something similar to those speedway bars for the front. I'd like to go adjustable, but it triples the material costs. Use a bar that has lots of size options. There are bars out there that all use the same bearing mounting areas but then have different torsion areas in solid and hollow options.


Oh, and who makes aluminum sway bars? That sounds like a terrible idea. Or at the very least, something that will need regular replacement intervals.

 

To three decimal places, you are correct. But if you measure it the hard way and I measure it the easy way and I'll bet you they are within 5% of each other. And, if you allow me to guestimate the contributions of the non-straight bits, I'll bet I can get within 1%.

 

Be sure to note that they are almost definitely telling you pounds for 5* of twist of just the bar. To translate this into something relevant to your car, you need to know the motion ratio of the bar as installed on your car, plus the track of your car. By the time that all of that is taken into account, the actual contribution of the bar to your roll stiffness is going to be very different from 80#/degree.

 

Your guestimate is? 5% would likely be acceptable as the stack up of manufacturing tolerance and mount flex would easily reach that.

It is accounted for in your equations as noted, but it is surprising how much of the flex actually comes from bending the arms vs. torsion in the bar. Particularly when you have arms that aren't perpendicular to the torsion section.

 

Sorry. My post wasn't clear. What I meant was that, if you give me A, B, C, & D for the Fred Puhn formula, plus a picture of the bar, I can correct the values of B & C for the bar not being straight, just by eye, within about 5% of correct. But I don't have the values of A-D or a picture of the bar. The only Evo-related sway-bar data that I have are for the rear bar on an X. (Were they earlier in this thread?)

 

I've got the stock bars sitting in my basement. Post a diagram of the measurements you are after.

That post above is hard to decipher and the equation seems to be missing some values.

That said, I don't plan on measuring the stock rate any time soon to validate your numbers. I could model it and do a quick FEA study on it, but I don't care that much. Ha, swaybars are the last place I'll turn for tuning handling.

All I know, is that calculating the stiffness of the bar based on actual engineering would be a total *****. But hey, if you can guestimate within 5%, that would be great.

 

The four needed values are shown in Post #5 of this thread, but much better pictures and details are here https://www.evolutionm.net/forums/ev...imensions.html (even if that thread is about Evo Xs).

 

If we get within 5-10%, then that's more than sufficient to get the car in the right ballpark. The last 5-10% is probably going to be tuned by feel and trial-and-error, anyway. At that point, we're only interested in relative differences between bars, not the absolute rate.