migrating roll centers

[1975BMW2002]

I am doing the calculations on roll center movement in my suspension design. I have a plot of how it moes through the suspension movement. I have made a few assumptions... first is that the inside arm picks up the same amount that the outside arm is pressed down. and that the amount that the chassis lifts or drops is insignificant since it is only 16 inches wide at the mounting points.

And now for the questions...
I would like to know if people think that these are valid assumptions.

and I would like to konw what is considered acceptable for roll center movement. I have it so that it moves less than a quarter inch up and down, but quite a ways from center at full travel. how far from center is acceptable before drivers start to complain about funny handling?

Bill
University of Maryland FSAE

[1975BMW2002]

I am doing the calculations on roll center movement in my suspension design. I have a plot of how it moes through the suspension movement. I have made a few assumptions... first is that the inside arm picks up the same amount that the outside arm is pressed down. and that the amount that the chassis lifts or drops is insignificant since it is only 16 inches wide at the mounting points.

And now for the questions...
I would like to know if people think that these are valid assumptions.

and I would like to konw what is considered acceptable for roll center movement. I have it so that it moves less than a quarter inch up and down, but quite a ways from center at full travel. how far from center is acceptable before drivers start to complain about funny handling?

Bill
University of Maryland FSAE

[Frank]

try this..

chassis in front view

http://www.users.bigpond.net.au/franko/sus.jpg

A/B = 1.5

and see what happens http://fsae.com/groupee_common/emoticons/icon_wink.gif

Frank

[Sam Zimmerman]

.

[Sam Zimmerman]

See http://fsae.com/eve/forums?a=tpc&s=763607348&f=125607348&m=4166078664. I think the assumptions that you have made are very bad. It seems to me, however, that many FSAE judges make the same assumptions and look at kinematic based roll centers. Think about roll centers from a common sense point of view and you will clearly see that your suspension is no where near symmetric in the turns.

But of course all of this is after a night of drinking so I may think differently later. http://fsae.com/groupee_common/emoticons/icon_smile.gif

Sam Zimmerman
Vandals Racing (http://www.uidaho.edu/~racing)

[1975BMW2002]

I had already assumed a frame setup similar to the drawing above. the important measurement left out is how long the arms should be. this determines quite a bit. if the top arm is about 10 inches long and the bottom is 14, it matters how far apart the arms are located in the verical sense at both ends--the upright and the frame. I have figured out a geometry that gives me a roll center about 3 inches above ground and a steep camber curve. it looks as though i could almost eliminate ARBs it is so steep. the tradeoff is that my roll center moves about 7 inches to the outside in one inch of suspension travel up on one side and one inch down on the other. it stays pretty much level the entire time though. I'm wondering if this migration is going to be enough to make a driver uneasy.

Looking at the geometries of past cars that we have laying around, and of a friend's formula ford and superV, it seems that the roll centers of those vehicles move very far laterally, and rise or fall a little.

I'm still confused on what is acceptable.

Bill

[Kevin Hall]

The tires will not accept steep camber curves. I once had a first year try to tell me that he solved all of the problems outlined by Smith and Milliken......and then I tried to explain to him that we didn't want 15 F'in degrees of camber.....it's not your 76 Buick, it's got some racing slicks under it.....think 2 deg of camber...

Just my rant for the day

Kevin Hall
University of Saskatchewan
'03-'04 Team Director

[Frank]

kinematic based roll centers

the A/B ratio wil cause lateral migration in roll

to large a ratio RC ->
to small a ratio RC <-

steering will move it one way too

1.5 was a guess at where to start

[1975BMW2002]

I have figured out what each aspect of changing the suspension does to how the roll center behaves through movement. I've analyzed ratios of bottom arm length to top, frame widths, ratio of the vertical distance between the two inboard points versus outboard points, and more.

I am not really asking the question of what happens when I do what, or even what anyone believes is a perfect design. I only want to know what my goal in the design is.

I am looking for guidance along the lines of "the roll center should not move more than X amount horizonatlly. It should/should not cross centerline through movement. it should stay above/below ground. it should move downward/upward through the suspenion's movement. it should start above/below ground because...

This is the kind of wisdom that I have not found anywhere on this list, or in Carroll Smith's books, but am in need of.

Thank you for your help.

Bill

[This message was edited by 1975BMW2002 on November 03, 2003 at 02:44 PM.]

[1975BMW2002]

as for the camber curve that i was talking about i meant in relation to the chassis to offset roll. about .9 degrees camber for each degree of roll. this yields a little positive camber if no static camber is tuned in. but with camber induced by caster, it should never be positive.

in theory it should never go beyond 3 degrees of negative camber.

Bill

[Denny Trimble]

How much caster are you running in the rear of the car? http://fsae.com/groupee_common/emoticons/icon_wink.gif

How do you expect the dynamic camber to vary between low-speed and high-speed corners at the front of the car?

University of Washington Formula SAE ('98, '99, '03, '04)

[1975BMW2002]

I am not planning to run any caster at the rear. I am currently concerned with the front geometry. In the front I am planning on 7 degrees of caster, and about 1 degree of KPI.

I expect the high speed cornering to have close to zero camber since it will have little help from caster induced camber. the lower speed corners will get help from the higher steering angles inducing more camber. it is my understanding that the low speed corners are more important on the courses run at competition.

[Denny Trimble]

Bill,
I was kidding about caster in the rear. If you have enough roll-steer to worry about rear caster, that's another thread.

You can generally trade-off your camber change values in ride vs. roll. If your roll camber change rate is .9, I'd expect your ride camber change rate to be about .5 degrees per inch of bump travel. Instant trajectory can change this a little (see Claude Rouelle's seminar for more info).

To set your target for dynamic camber, you have to know what the tires like. I believe the sweet spot for Hoosier bias plies is around -1 degrees. Then, if you know your ride and roll spring rates, CG height, RC heights, anti-squat and so on, you can "simulate" the car in various phases of the corner, then plug the ride and roll values into your camber curves and study what the tires are doing around the course. Don't forget about structural deflections, which can double your camber change if you're not careful. Don't spend that extra week in front of the computer fighting for that last .1 degree of camber change, only to see it washed away in bearing or a-arm compliance.

Isn't it about time to start building? http://fsae.com/groupee_common/emoticons/icon_smile.gif Good luck.

University of Washington Formula SAE ('98, '99, '03, '04)

[clausen]

Hi there Bill

No, I think your assumptions are conflicting with your findings. For your chassis to have no change in height, it would follow that your outside suspension is compressing the same as the inside is extending. For one of these things to happen, the other must happen. This is fine, but what it does mean is that the car is rotating around a point on the centreline of the car, ie the roll centre is on the centreline. Therefore if your assumptions were valid then the roll centre would have zero lateral migration. As you know this doesnt happen.

In other words, if you had a lot of lateral migration so that at some stage the RC was in the centre of (say) the inboard wheel, then the car would be pivoting about that point and the inboard suspension wouldn't compress at all.

As far as acceptable limits go, my theory with RC height change is that it would be nice if it rose and fell at the same rate as the chassis (ie didnt move in relaiton to the chassis) so that the roll moment stayed constant. I'm not sure about laterally though. Some people say that as long as it stays within the track its ok. Mine only moves about 10cm I think.

Note that the big lateral migration numbers come as the RC goes through the ground plane, so it might be a good idea to try to ensure that this doesnt happen. I read a sportscar (lemans) suspension designer mention that this was one of his objectives in it's design.

Hope this helps

Regards

Paul Clausen
Uni of Adelaide

[1975BMW2002]

Paul,

You have an excellent point about the car rotating around the roll center, thereby dissolving my theory that I can assume that the suspension travel on one side is the negative of the travel on the other side. I had not considered that, even though it seems obvious. feel free to knock me over the head.

I had been operating under the assumption that vertical movement would be more important to control than lateral movement. I did some analysis on some cars that I know handle well, and the roll center seems to move quite a bit laterally, yet not vary more than an inch vertically.

Thank you for the insight,
Bill

[Frank]

the thing i find interesting when calculating kinematic roll centres is tyre deformation (neglected in SUS-A package)

WRT lateral roll centre movment..

if you have a very complient tyre (compared to springs) then you want the chassis more "rectangular" in front view as opposed to triangular.. ie A/B ratio nearer to 1 than 1.5

Frank

[1975BMW2002]

I checked into this. We are planning on going with 13 inch tires form goodyear. They are running over 500lb/in deflection at the kind of pressures we want. Our springs are much less than that. The current design I am going with has the upper arm about 65% of the length of the lower arm. It seems a little short, and if the suspension were to move much further than it's limited travel, it would go bad quickly. but for the movement I am putting it through, it seems fine.