Anti-Roll Bar Rates

[IsheeM]

I apologize if this has already been discussed in depth, but i searched the forum and found nothing.

What rates (lbf*ft/deg) do most teams use for their anti-roll bars? Is it justifiable to have an ARB if the rate is 8 lbf*ft/deg?

Thanks,

[fsae racer]

It depends what contribution that makes to your total roll stiffness or front/rear roll stiffness, where ever it may be. However, most likely it will not be possible to have it provide a worthwhile contribution as the required motion ratio would result in plastically deformed bars.

[DanO]

That is somthing you have to design in.

There is no (lbf-ft/deg) "number" that we could give that could help. You need to determine the roll stiffness you want to achieve and look at where the bar is connected and calculate it that way.

[IsheeM]

I believe Denny mentioned once there is a good ratio between front and rear roll stiffness. If I didnt just completely make this up, could someone sound in as to what this ratio should possibly be and how they came up with this answer.

Thanks,

[Denny Trimble]

It depends on a lot of variables (almost all of them), and can even be adjusted day-to-day to suit conditions at a given track. For our previous cars, with a 45:55 F:R weight distribution, 55:45 F;R roll stiffness distribution was in the middle of the ballpark.

No theory there, just trial and error. But if we had good tire data... http://fsae.com/groupee_common/emoticons/icon_smile.gif

Oh, and here's a previous discussion of the same topic:
http://fsae.com/eve/forums?a=tpc&s=763607348&f=125607348&m=3226059164&r=7686059164#7686059164

[IsheeM]

Denny,
Thanks for the help. We currently have a FR51%:RR49% roll stiffness distribution with FR45%:RR55% weight distribution, which doesnt seem too far off what others have calculated. However in the thread you linked to, there was mention (as you hinted above) that if we had better tire data we could find a better solution instead of so much trial and error. Can you explain how tire data would help resolve this matter.

Thanks,

[Akos]

Like people have said previously, there is no fixed correct ratio of front to rear roll stiffness distribution.

Depending on your tires there is however two distinct ratios I have always ended up with (given a mid engined car with roughly 40:60 weight distribution).

If the front tires can generate a lot of camber trust, and the proper amount of negative camber can be dialled in, the ratio tends to about about 85-95% front stiffness. Essentially this unloads the inside front tire. With this type of setup if the front is not stiff enough the camber thrust generated by the inside wheel will actually hurt the overall front end grip, since the force is in the wrong direction. This type of setup has a couple of nice benefits including very good power down and fast transient response.

If the front tires can't be cambered, then both front wheels are needed and can be used for grip. This is the case where you end up with textbook quoted 55-65% front.

In terms of overall roll stiffness, my personal view is if you can see the car roll, than the roll bars are not stiff enough. This applies to uneven tracks. If the car is dancing around the track, the problem is generally with the springs and the shocks not a stiff bar.

Cheers,

[Denny Trimble]

Michael,
If we had better tire data, we would know more about the load sensitivity, camber thrust, and optimum slip angles at certain normal loads. An enterprising student could put together a simulation / optimization program to look at various vehicle states, and get answers for balance, total g's, braking and driving capacity, ideal torque bias ratio... the list goes on.

[ethanL007]

Denny, thats good stuff. I laughed for the first time this week.

I don't think there should be any more new posts on vehicle dynamics without first checking and signing up for:
http://fsae.com/eve/forums?a=tpc&s=763607348&f=125607348&m=88810382911
Because teams without data end up asking the teams with data for information, and any information will be far more valuable, and advanced, after the tire data comes in, and in a round about way is benefiting the teams without a share in the tire pool.

Finals cynicism is setting in.

Edited to reduce grumpiness

[Frank]

Bars are used for two reasons:

Firstly, to limit roll; and the total acceptable roll is totally dependant on camber gain in roll. (much camber gain in roll when VSA is long)

Secondly, to create a form of adjustment in roll stiffness. (preferably driver adjustable)

The ratio of roll stiffness front to rear is predominantly dependant on tires and track surface. (on that day, at that time)

For a 50:50 weight split car, the roll resistance ratio (RRR) could be anywhere between 40:60 and 60:40, but a RRR close to weight split is a good start in fsae.

Remember that RRR is NOT really a metric. It can be defined in a number of ways. What is important is that RRR is a comparative number that describes the individual wheels weights for a given situation. Anyway, it's a great idea to always know what your number is at any particular time (which involves knowing your motion ratios ACURATELY, and ARB rates)

(Loosely) quoting Carroll Smith:

"On the racetrack, on the day, with race tires, and race setup. Set the (drivers adjustable) bars in their middle setting, and throw springs at the car until the driver reports a neutral car.

Then play with shocks, spring and bar rates (keeping the same RRR, and %critical damping in ride) (( ideally keep the same % critical damping in roll too, but this is difficult unless you manage to decouple the roll and ride damping)), camber, toe, ackerman, etc etc.

Repeat process of getting bars in middle setting, and that's it.

Don't be at all surprised if the car comes back from a race with front bar full soft, rear full hard, or vice versa.

Remember the change in tire condition and racetrack surface condition is in a constant state of flux. (especially if you abuse tires)"�

Me personally, I like stiffly sprung FSAE cars with the front considerably softer sprung than the rear. My (drivers adjustable) front ARB gives between 50-130 Nm/deg, which only offers about 6% change in RRR.

The front tires and springs contribute 185 Nm/deg.

If I had the chance to do it again I'd insist on front and rear bars (drivers adjustable blades both ends). The blades in their middle setting they'd offer about 40% of the total RRR, and 10% RRR change would be available to the driver.

Oh, and there's no point to driver's adjustable ARB's if the driver is incapable of knowing how they're being used (get another driver).

And importantly, don't allow any "slop"� in the ARB mechanism.

Regards

Frank

http://www.uq.edu.au/fsae/statics.xls

http://www.uq.edu.au/fsae/sae_a_2004/images/design_01.jpg

[FATMOUSE]

Alright, huge bump here.

When doing up the geometry for the 2008 CU car, in the absence of tyre data, I had the devil's own time in guessing what would be the best way to set the whole thing up w.r.t. roll.

We already had a set of bike shocks (Fox DHX3s) with some hefty 400lbs/in springs on there. After going through ride frequencies, wheel rates, etc. we ended up setting a motion ratio of 2 (.5" shock travel for 1" wheel travel).

So I ask myself, what to do about roll? The only data I had to go on was that somewhere in the region of -1 degrees of camber was where the Hoosier slicks we had were happiest. And this was just from half a post's content on this forum.

I also felt that we probably wouldn't have time to do an ARB setup properly for a pretty much first year team, so I just gave the car full camber compensation in roll. But then competition came around, which was the first time we really drove the car at any speed, and the thing barely rolled anyways.

So in come the questions:
1. I'm guessing this lack of roll is down to those hugely heavy springs in concert with that motion ratio. The picture I'm forming in my mind is that it's easy to rotate a door around the hinge when pushing the far edge, but it takes a lot more force to do so when pushing it close to the hinge. So basically the motion ratio gave a sane wheel rate in bump, but together with the heavy springs made it really stiff in roll. Is that a more or less sharp assumption, or am I completely wrong?
2. Let's assume that I'm correct. (If not, feel free to ignore the question). In that case, is it better to run really stiff springs with such a high motion ratio and get large roll stiffness with a reasonable wheel rate, and allow more camber change in roll? Does this have any adverse effects of things like weight transfer?

I already know tyre data would be of huge importance for making these decisions. I appreciate any thoughts, and would especially *love* people telling me that I'm completely wrong (that way we all learn more, natch).

-Arjun Roy
Suspension und steering also
Columbia University Formula SAE '09

[Alan]

1. Wheel rates and roll rates are directly related by the equation roll rate = 1/2 * wheel rate * track width^2. Whether or not the resulting number would be considered stiff in roll depends on other things, for example CG height. The springs and MR you state give a wheel rate of 100 lb/in but the 400 lb/in spring is not what gives the car roll stiffness, it's the 100 lb/in wheel rate. So I think you are correct in what you are saying, but worded it awkward or I'm just not reading it right.

2. Not sure how to address this one since I'm not sure what it is you are asking. But really stiff springs with a high motion ratio will result in some roll stiffness which is not independent from the resulting wheel rate.

Not sure if that helped at all but if you have any questions, there are couple of really good sections in the Milliken book that go over these exact issues.

[FATMOUSE]

What gets me is this. Suppose there was a car with a beam axle, and one spring in the center attached to the chassis - surely such a car would have an appreciable wheel rate dependent on the spring rate, but have pretty much no roll stiffness whatsoever?

So does that roll rate formula you post apply in such a case?

[exFSAE]

Like a monoshock on a beam axle? Not sure why you'd want to use that.. trying to balance the chassis on basically a pivot axis that would just flop over.. but in that case yea you'd have no roll stiffness and the car would just flop to one side or the other. Though it also kinda depends on how you have it set up. If your spring was set up on a spherical joint at each axle that would be the case.

Alan's bit I would call the contribution of wheel rates of an independent suspension to roll stiffness.

In the scenario you provided its basically a 3rd spring or monoshock without anything else. Since the "independent" wheel rate is 0, the contribution to roll stiffness is 0.

Don't overthink the problem. And as Alan said, its not the spring rates that make any difference, its the wheel rates. If you don't have a baseline to work from or tire data, get a ballpark of where you want to be for ride rates. The OptimumG site and Milliken I think have some examples. Might even try the Claude advice of targeting your TLLTD a tick more front bias than your F/R load distribution. Buy a couple sets of springs.. one at your target, set or two stiffer, set or two softer. Have a couple roll bars to put on there. Test it. Zero in on your steady state balance, see if you want to go stiffer or softer, and go from there.

[murpia]

Alan's formula only applies to independent suspensions, not beam axles.

Just one thing to note about matching roll couple distribution to weight distribution (and aero distribution for that matter). Be careful that when matching percentages you match FRONT roll couple to REAR weight to REAR aero.

So for example a car might run:
45% F / 55% R weight distribution
43% F / 57% R aero distribution
57% F / 43% R roll couple distribution

If you misunderstood the concept and ran 43% F / 57% R roll couple I think that might make for an entertaining drive...

Regards, Ian

[Brian Evans]

1. I'm guessing this lack of roll is down to those hugely heavy springs in concert with that motion ratio. The picture I'm forming in my mind is that it's easy to rotate a door around the hinge when pushing the far edge, but it takes a lot more force to do so when pushing it close to the hinge. So basically the motion ratio gave a sane wheel rate in bump, but together with the heavy springs made it really stiff in roll. Is that a more or less sharp assumption, or am I completely wrong?


I don't understand why you think the wheel rate is somehow different if the wheel is loaded by corner weight, or roll weight transfer. Weight is weight, it matters not from where it derives. You had a corner rate of 100 lbs/in, which is probably slightly (or a lot) higher than your corner weight. That is a lot for a low/no downforce car. Probably no need for an ARB at all. Many very good cars don't use ARB's. You can get all the roll stiffness you need with springs and geometry. The only real advantage of an ARB is adjustability.

[FATMOUSE]

Originally posted by Brian Evans:
I don't understand why you think the wheel rate is somehow different if the wheel is loaded by corner weight, or roll weight transfer. Weight is weight, it matters not from where it derives.

That is what I figured initially, but got confused on when looking at a couple of sources that spoke of roll being resisted by spring rate, even though they mentioned wheel rate elsewhere.

Thanks for the responses, all.