Roll Gradient

[Moop]

Hey guys,

I was randomly thinking about my car's roll gradient today and I realized that I don't really know where the give and take is in adjusting roll gradient. I also realized that I don't know why we run the gradient we do other than because it's close to a number mentioned in RCVD or because it's "what we've always run"(which I don't think is a very good justification, haha).

Having a lower roll gradient obviously lets you reduce the amount of roll and hence camber change you get in corners(potentially covering up poor kinematics that have too much I guess, haha). I would also think that it makes your car more responsive.

I know absolutely nothing about NVH, but I'd think that running a stiffer roll gradient would be harsher on the driver, and obviously there's some benefit to keeping your driver happy, reasonably comfortable and not breaking his/her neck.

I get the feeling that I'm missing some other benefit of running a softer roll gradient. Does anybody know if I'm missing anything here?

[JDS]

Choosing a roll gradient is dependent on several factors and really there is no one correct answer. The two biggest factors to me would be the kinematics (instantaneous center mostly) and tire data for inclination angle. The first step for me is to look at the tire data and determine the optimum inclination angle for a range of load and slip angles. Then run weight transfer calculations to determine load at a range of expected lateral accelerations. Determine the roll angle based on your spring setup and check with your kinematics to see where your camber would be at those lateral accelerations. If the camber is non-ideal based on your tire data, you can either adjust your spring setup to get more or less roll, or you could try a different IC that will get you to where you want to be with camber. Milliken talks about designing with desired roll gradient in mind, but to me, the roll gradient is an after effect of the above mentioned trial and error method. Hope this helps.

[exFSAE]

I would argue that in a SLA suspension, looking at cornering only, any amount of body roll is always going to be detrimental to some extent to your mid-corner tire inclination. Never helps anyway. But that's beside the point.

As you add stiffness to your suspension you approach rigid-body, 'go-kart' behavior. There will be less lag for the car to 'take a set', though there is a point of diminishing returns.

Bear in mind, one of the whole reasons for having a suspension is to isolate road disturbances from the sprung body. On aero cars, there's a trade off of springrate for keeping your downforce generators in the orientation and height you want, against keeping the suspension compliant and the tires in good contact with the track.

FSAE is obviously a bit of a different animal, some teams not using aero, and the competition surfaces now being pretty damn flat.

Some other things to keep in mind... as you run stiffer suspension, your dampers will do less and less as they get little travel and generate little velocity. There are also trick little things you can do, with roll-steer for example, which you then eliminate the option of by locking out the suspension with a near-zero roll gradient.

Not to mention that with a very stiff suspension on a fairly compliant car (as FSAE vehicles tend to be), your vehicle balance is going to be affected in large part by compliance rates rather than spring and bar rates - which you probably don't want from a tuning standpoint.

Just some stuff to consider. If you want to see what it really does, get a high-stiffness and low-stiffness setup and track test them both.

[ZAMR]

Like exFSAE said, your frame's torsional stiffness will become more and more important with stiffer suspension settings. At a certain point you will not be able to control your car's lateral load transfer distribution as effectively. Additionally there are negative transient effects from having widely different roll frequencies front to rear.

Also keep in mind the stress placed on vital components especially the anti-roll bars.

On dampers:

If you increase your roll frequency without correspondingly increasing your ride frequency, the dampers will provide very different damping ratios for ride and roll. To properly dampen your car in roll with a high roll frequency and moderate ride frequency, your ride damping will be overdamped which can have effects on road surface induced TLV. That being said, having a high ride frequency also negatively effects TLV.

So when damping comes into the picture everything gets a lot more complicated. Honestly analyzing the tradeoff between increased frequencies v. increased TLV v. compliance issues is beyond the scope of FSAE, which is why we have "guidelines."

[flavorPacket]

Originally posted by ZAMR:
If you increase your roll frequency without correspondingly increasing your ride frequency, the dampers will provide very different damping ratios for ride and roll. To properly dampen your car in roll with a high roll frequency and moderate ride frequency, your ride damping will be overdamped which can have effects on road surface induced TLV. That being said, having a high ride frequency also negatively effects TLV.

This is assuming completely linear dampers...which I would submit are not suitable for FSAE cars for this very reason, among others.

[exFSAE]

In any event, at this level, it is interesting stuff and good to learn but probably not worth spending too much time on.

1. Build rigid, reliable car that doesn't look like it has the force/deflection curve of fresh mozzarella.

2. Get your drivers to operate outside the linear range of the tires.

3. Don't do your reports the night before they are due, nor in a manner wordy enough that it rivals The Iliad in length.

4. Top 10!

Better though to ask these type of technical questions I suppose and actually be thinking at a high level, rather than - "How do I design brake? Iz urgent."

[buggaero]

I would like to mention that i'm still learning about the art of Roll gradient tuning.
Don't feel like experienced enough to write something here... but IMO, bringing in the Anti-dive and ARB in the scene, it changes alot in the whole picture,a totally new script.

In very concise terms, i find that we don't utilize our generated power to its max end with a stiffer roll and waste it more than we should with a softer one (considering Non Aero in both).
So it turns into finding the best compromise.

[M. Nader]

Ok, i would like to resurrect this one.

Roll gradient, kinematic changes, wheel rates and ARB tuning. regardless of the optimization criteria, has anyone run a stiff enough setup that the driver complained from it?
I am more interested in roll gradient than bump effects as no i am more specifically looking at ARB tuning thinking how far can i take it. no point designing the kinematics around no camber change in bump while in roll the camber goes up 5 degrees for example, or the opposite.

I want to find a common ground without having the driver assassinated by the car, so has anyone had a roll, gradient where the driver complained from? or one that the driver was exceptionally comfortable with?

[Buckingham]

I want to find a common ground without having the driver assassinated by the car, so has anyone had a roll, gradient where the driver complained from? or one that the driver was exceptionally comfortable with?

Don't paint yourself into a box. Ride and Roll are two different modes, which means you can design for them independently if you are clever enough.

There are multiple ways to have a high roll gradient AND soft ride rate (lots of jacking, Z's interconnected suspension, hydraulic circuitry, etc).

1. Choose a target roll gradient based on whatever criteria matters for you (aero, camber, transient response, etc).

2.Then choose target ride rates (depending on how rough anticipate your track to be).

3.Engineer a way to get both. http://fsae.com/groupee_common/emoticons/icon_smile.gif

[M. Nader]

Originally posted by Buckingham:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">I want to find a common ground without having the driver assassinated by the car, so has anyone had a roll, gradient where the driver complained from? or one that the driver was exceptionally comfortable with?

Don't paint yourself into a box. Ride and Roll are two different modes, which means you can design for them independently if you are clever enough.

There are multiple ways to have a high roll gradient AND soft ride rate (lots of jacking, Z's interconnected suspension, hydraulic circuitry, etc).

1. Choose a target roll gradient based on whatever criteria matters for you (aero, camber, transient response, etc).

2.Then choose target ride rates (depending on how rough anticipate your track to be).

3.Engineer a way to get both. http://fsae.com/groupee_common/emoticons/icon_smile.gif </div></BLOCKQUOTE>


Yes i know that, i am actually at step 3 now. (adjusting ARB and kinematics for roll)

But the main problem for me now is the driver, theoretically i can make it work with what i have. the rates and desired roll gradient.

but choosing a roll gradient is quite tough (this is not an FSAE application actually), i know i can improve on what i have if i choose another roll gradient, but i am not sure it will be good for the driver though. which is why i am looking for someone who actually tested something out as i couldn't find any ergonomic papers regarding driver comfort with roll gradient.

[Jay Lawrence]

I have driven FSAE cars on either end of the roll stiffness scale (unfortunately not the same car for direct back to back comparison, but still relevant I think). I found the soft car very difficult to drive because I did not feel in total control (I was being tossed about a fair bit and it was simply uncomfortable). I found the hard setup to be much nicer as my inputs were both easier to enact and to measure the result of (I found that the chassis was more 'transparent' in it's communication).

[M. Nader]

Originally posted by Jay Lawrence:
I have driven FSAE cars on either end of the roll stiffness scale (unfortunately not the same car for direct back to back comparison, but still relevant I think). I found the soft car very difficult to drive because I did not feel in total control (I was being tossed about a fair bit and it was simply uncomfortable). I found the hard setup to be much nicer as my inputs were both easier to enact and to measure the result of (I found that the chassis was more 'transparent' in it's communication).

Thank you, that's what i am looking for. can you give me some numbers for their roll gradients? soft and hard are quite vague and give a wide spectrum range

[Xavier ABRAM]

Don't forget to check your pre-loads as well... Sometimes you run so stiff that you need to have negative pre-loads in order to get that damper length you designed for in CAD.

This is not so much of a good idea because you get a F-word weird roll stiffness curve which is dependent of damper positions. And tuning handling with a suddenly dropping roll stiffness is maybe not as bad as steering play, but still not very funny.

The effect of compliance is a very good point BTW. Check it out too. Camber losses everywhere.

I'd say 1.0 to 1.4 deg/g... but it depends on so much things like aero-plateform, driver feeling, bumps on road, feasibility with springs and ARBs (lol), chassis stiffness, cg height, camber gains, ... on and on and on (all 31 bizillion^2 parameters of a car work together). Some run .8 other 1.6.

[M. Nader]

Originally posted by Xavier ABRAM:
Don't forget to check your pre-loads as well... Sometimes you run so stiff that you need to have negative pre-loads in order to get that damper length you designed for in CAD.

This is not so much of a good idea because you get a F-word weird roll stiffness curve which is dependent of damper positions. And tuning handling with a suddenly dropping roll stiffness is maybe not as bad as steering play, but still not very funny.

The effect of compliance is a very good point BTW. Check it out too. Camber losses everywhere.

I'd say 1.0 to 1.4 deg/g... but it depends on so much things like aero-plateform, driver feeling, bumps on road, feasibility with springs and ARBs (lol), chassis stiffness, cg height, camber gains, ... on and on and on (all 31 bizillion^2 parameters of a car work together). Some run .8 other 1.6.

Thank you that helped a lot.