Weird ARB

[M3Shark]

I have been researching Anti-roll bar design for our upcoming years car and cam across a design that i have not seen before. Cal Poly Pomona among a couple of other teams i have seen use it on their rear. and i think i see why, at least for us it would be a lot easier to package than a standard u-type bar.

You can see the bar i am talking about here on Pomona's car: http://picasaweb.google.com/gi...#5221671463799226082 (http://picasaweb.google.com/gielamonster/BestOfFSAEWest#5221671463799226082)

I think i understand the theory on how it works, but how do you do the roll resistance calculations for this type of bar. Are their any disadvantages to this type of bar over the typical u-type bar? Any literature floating around somewhere on it?

Thank you,

M3Shark

[J. Vinella]

This type of bar is know as a T-type bar, I think you can see why.

It works on the same principal as a U-bar, a member in torsion activated by lever arms. Typically teams use a "blade" design for the lever arms. The blade can be rotated from a soft to a stiff position, one downside it the progression from soft to stiff is not linear, think a tan(x) function from (-pi/2 to pi/2) but sideways. There is really no change in stiffness at the top and bottom. You end up more or less three different adjustments. 0, 45, 90 degrees. The major advantage is that you do not change link length like with a standard u-bar and you do not change the motion ration. You can almost keep things in plane if you really try hard. You can see what I am talking about in the rear of our '05 car. The ARB is in the 45 degree orientation.

http://students.washington.edu/dennyt/fsae/cnc/wc_oncar.jpg

Cheers,

[RacingManiac]

we've had some trouble with T-types bar in the past, especially if they are of a longer bar from the pivot to the lever arm(top of the T). The issue tend to be they might end up doing more bending than torsion, which does wierd thing to your rate, especially if rate purpose you end up with a small diameter, thin wall tube.

2007 car's front T-bar:
http://www.fsae.utoronto.ca/20...uction/Testing04.JPG (http://www.fsae.utoronto.ca/2007/pictures/2007%20Construction/Testing04.JPG)

2002's rear T:
http://www.fsae.utoronto.ca/2002/large/det033.jpg

[murpia]

T-bars are good for packaging but harder to calculate a rate for. Especially if the bending stiffness of the T arms is of the same order of magnitude as the torsion of the vertical of the T (in wheel rate).

If you can, do a physical test, maybe as part of an overall torsion test of the car.

Regards, Ian

[Wes Johnson]

This is off topic, but sweet mother check this out!!

http://picasaweb.google.com/gi...#5221671621844421714 (http://picasaweb.google.com/gielamonster/BestOfFSAEWest#5221671621844421714)

[PatClarke]

And, a couple of pictures later, check the effect of the 'snatch' loads when the rear wheel gets back on the ground!

Pat

[Brian McGreevy]

Our team used a T-bar ARB setup in 2007 and 2008 with much success. You can design-in a lot of adjustability into it and it allows for different kinds of packaging. We're going to a more conventional U-bar setup this year because of our packaging.

[L B0MB]

I dont mean the tread jack but, damn there were some ugly cars at that comp!

Ours wasnt pretty but it wasnt munted either http://fsae.com/groupee_common/emoticons/icon_smile.gif

[M3Shark]

so with out physical testing, is there any general formula for how much roll resistance the T-bar will provide.

I know there is one for a u-type bar( i have it some where, just not off the top of my head), but is there a similar on for a t-type bar? or has everyone who used them simply done physical testing to find out.

[RacingManiac]

For our case we just assumed for torsion without bending, but I think at some point it ceases to be accurate. If you have a short enough bar(or a big enough OD/L ratio), that assumption will be relatively accurate. The correct answer would be actually testing it.

[J. Vinella]

Originally posted by RacingManiac:
The correct answer would be actually testing it.

FEA will get you close enough for a first guess.

Another added benefit to the T-bar that often never gets mentioned is: it makes bearing really easy to package where doing something with low friction on a U-bar is not as easy and often I have found heavier.

[murpia]

Originally posted by M3Shark:
so with out physical testing, is there any general formula for how much roll resistance the T-bar will provide.

I know there is one for a u-type bar( i have it some where, just not off the top of my head), but is there a similar on for a t-type bar? or has everyone who used them simply done physical testing to find out.
Of course you can calculate the rate. The formula for the torsion rate of a bar or tube will be the same as used in the U-bar. As will the kinematics required to calculate how much bar link deflection is caused by wheel travel. The only difference will be that the bar torsion per link deflection is half that of the U-bar.

Regards, Ian

[M3Shark]

Originally posted by murpia:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by M3Shark:
so with out physical testing, is there any general formula for how much roll resistance the T-bar will provide.

I know there is one for a u-type bar( i have it some where, just not off the top of my head), but is there a similar on for a t-type bar? or has everyone who used them simply done physical testing to find out.
Of course you can calculate the rate. The formula for the torsion rate of a bar or tube will be the same as used in the U-bar. As will the kinematics required to calculate how much bar link deflection is caused by wheel travel. The only difference will be that the bar torsion per link deflection is half that of the U-bar.

Regards, Ian </div></BLOCKQUOTE>

Ian,
Sorry i am not quite following you on what the formula is.

So the roll bar resistance formula for a standard u-bar shape is:
[Bar linear Rate X (wheel movement X Bar pickup movement)^2 X Track^2 X Pi]/180

with bar linear rate equal to:
Angular rate/ (Lever length^2 X pi/180)

and angular rate(steel) =
[19,700 x (OD^4 - ID^4)]/Bar Length.

So where exactly am i dividing by two for the T-bar. And if it not to hard to explain, why am i doing that?

[salad]

A U-bar will twist twice as much as a T-bar, all else being equal. What it means is that for the same bar dimensions (OD, ID, and length, and all the other stuff), the U-bar will give twice the roll rate for the same torsional stiffness of the active bar.