24 replies to this topic

[Jean Fred]

Hello I have some questions about roll center location an swing arm on modern racing cars like sport prototypes (radical osella dallara norma), F3 or F1.

-Is there any diferencies about roll center location according to the rules, the tyres characteristics, wheight, or running strategies in theses three classes?

-What strategy must a designer adopt to choose the roll center location. Is the only goal to keep it as immobile as possible during bumping or rolling?

- Is it better to locate it near the ground and stiff anti roll bars because the lateral forces applied on the gravity center increase roll moment due to the distance between roll center and gravity center? or to keep it near the center of gravity to avoid rolling?

- do you know some books or internet links that really teach about that? I have already red
«Competition Car Suspension » (Allan Staniforth)
« How to make your car handle » (Fred Phun)
« Chassis engineering » (Herb Adams)
« Race car chassis design and construction » (Forbes Aird)
« Race car engineering and mechanics » (Paul von Valkenburgh)
some of theses talk a little about roll center and swing arms but it has brought me more questions than answers... so I am looking for something more specific. I need teaching about that.

Excuse me for my English I am french

Thank you for helping me

[Antoine]

Hello Jean Fred,

I don't think that there's one solution for one car! and the roll center height of all the car you talk about were defined years after years by studiing all the other parameters and trying to find the good way to increase the performance

in F3 or F1 we move the roll center when we need, so there's not only one ! You can see the front suspension of Mac Laren or the last one of toyota which are a lot different of the other team!

About the book take a look at the RCDV.

[Jean Fred]

OK I agree with you but what about tendenties?

ten years ago the roll centers were very far to what we can see today

[Greg Locock]

-Is there any diferencies about roll center location according to the rules, the tyres characteristics, wheight, or running strategies in theses three classes?

It's a tunable parameter, not defined in the rules


-What strategy must a designer adopt to choose the roll center location. Is the only goal to keep it as immobile as possible during bumping or rolling?

The geometric roll centre controls how quickly forces at the contact patch get transmitted into the body - the higher the roll centre the greater proportion of roll is reacted by the arms as opposed to the springs, a/r bars, and shocks.

- Is it better to locate it near the ground and stiff anti roll bars because the lateral forces applied on the gravity center increase roll moment due to the distance between roll center and gravity center? or to keep it near the center of gravity to avoid rolling?

In practice it tends to be near ground level. It is not uncommon for small cars to have it below ground level at the front.

- do you know some books or internet links that really teach about that?

Not really. Milliken and Milliken is at least correct, but doesn't go much into the details of why you would choose a particular setup. I think they include it in their very useful tuning matrix. Watch out for the other books, many authors have written rather regrettable things about roll centres. The most common misconception is that the vehicle rolls about the roll axis defined by the two roll centres. It does not.

One other problem is that the geometric roll centre (GRC) moves in a different way to the force based roll centre (FBRC), and the latter, while harder to work out, seems to have greater physical meaning, since it directly predicts the weight transfer.

You should also be aware that the GRC is not an isolated parameter. If you change the RCH on a double wishbone suspension then you will also change the camber gain and track gain, both of which are important in their own right, typically.

If you wander over to my website I have a spreadsheet, SLARCK, that I use for setting up hardpoints which calculates the GRC, for the simplest case where front and rear body side hardpoints are coincident in front view.

www.geocities.com/greglocock

look on the downloads page. I have extended it quite a lot lately so that it runs as an optimiser, but I'm not distributing that version.

Alternatively if you can find a program called wishbone.bas, that uses a full 3d model of the suspension. I have checked and it gives the same results as SLARCK, and is generally a far more useful program, although it has the user interface from hell.

[Fat Boy]

Originally posted by Greg Locock
The geometric roll centre controls how quickly forces at the contact patch get transmitted into the body - the higher the roll centre the greater proportion of roll is reacted by the arms as opposed to the springs, a/r bars, and shocks.

To comment on this thought and your observation on how RCH has changed, I think it goes back to the tires we are dealing with now as opposed to 10 years ago. If a tire can accept very quick loadings, then you can run higher roll centers with softer bars and springs. The loads taken though the wishbones happen very fast. Loads through the springs and bars are relatively slow. If your tire cannot accept being loaded quickly, then it's best to run low roll centers and proportionally carry more of the rolling loads in the springs/bars.

I find roll center changes to make their biggest effects in transients. When everything is said and done, they don't seem to effect steady state nearly as much. Since most of what we deal with is transient, though, a roll center change can be very effective.

Of the books that you've mentioned, be very careful with the Herb Adams book. Some of the stuff in there is just plain wrong.

[Greg Locock]

I've been told that at least one of the Carroll Smith books is wrong, although he seems to have sorted it out in another one. Sorry, I don't know which is which.

[Deepak]

I started out reading a few of the books above but after a while they were confusing.

I just decided to get Milliken and started reading it. I find a lot of things easier to understand & decided to stick to this.

[Ben]

Originally posted by Greg Locock
I've been told that at least one of the Carroll Smith books is wrong, although he seems to have sorted it out in another one. Sorry, I don't know which is which.

Tune To Win has some utter rubbish about front and rear CG heights and the non-existant mass centroid axis. To his credit, on the one occasion I was lucky enough to meet Carroll before he died he told me that "a lot has changed in twenty years since I wrote those books". The fact remains the Tune to Win is wrong on roll centres but this showed me that Carroll was humble and always prepared to learn new things.

Ben

[Jean Fred]

"The geometric roll centre controls how quickly forces at the contact patch get transmitted into the body" -> because of the wheel instant center height?

the higher the roll centre the greater proportion of roll is reacted by the arms as opposed to the springs, a/r bars, and shocks." -> because of damping? again

Can you explain more in details please, I am not sure to understand why as you can read it.

"If a tire can accept very quick loadings, then you can run higher roll centers with softer bars and springs. The loads taken though the wishbones happen very fast. Loads through the springs and bars are relatively slow. If your tire cannot accept being loaded quickly, then it's best to run low roll centers and proportionally carry more of the rolling loads in the springs/bars."

What characteristic of the tyre do you talk about? lateral stiffnes, or vertical, or something else?

for Herb Adams book I noted also some contradictions but there are still some parts which are interesting.

what is the title of Milliken's book?
I found "Chassis Design: Principles of Analysis"
and
"Chassis Design" by F.Milliken and Douglas L.Miliken based on previously unpublished technical notes by Maurice Olley

Thank you very much for your answers it is very interesting for me.

[Supercar]

The slow or fast transmission of contact patch forces into the body has to do with jacking and with the delay in reaching the steady state cornering condition due to the roll motion of the car. If you have a 100% jacking, the RCH is as the C.G. of the car, then there will be no body roll due to the motion of suspension. The suspension will not move when the car starts cornering, only the tires will compress. The wishbones are pointing at the C.G. of the car. Any forces at the contact patch will get instantly transmitted into the body through those wishbones (once the tires are compressed). In this case the springs, bars and dampers do not see any additional load when the car is cornering. Everything is instantly transmitted through the wishbones.

If the roll center is at or below ground, then the wishbones alone cannot transmit all the tire load because they will want to fold up. The car will have to roll first and compress the springs, bars and dampers. That takes some time, hence the delay.

I will let FB explain the effect if tires and tire properties. Not sure I can do it well.

The book is "Race Car Vehicle Dynamics" by W. Milliken and D. Milliken.

Philip

[Lukin]

Originally posted by Fat Boy
To comment on this thought and your observation on how RCH has changed, I think it goes back to the tires we are dealing with now as opposed to 10 years ago. If a tire can accept very quick loadings, then you can run higher roll centers with softer bars and springs. The loads taken though the wishbones happen very fast. Loads through the springs and bars are relatively slow. If your tire cannot accept being loaded quickly, then it's best to run low roll centers and proportionally carry more of the rolling loads in the springs/bars.

I find roll center changes to make their biggest effects in transients. When everything is said and done, they don't seem to effect steady state nearly as much. Since most of what we deal with is transient, though, a roll center change can be very effective.

Great post. Rear Roll Centre height is good for quick tune ups on the taxi's (to borrow Greg's term). Going up on the RRC is good to cure a lazy car that has some understeer. The combination of slightly lower roll angle and quicker load transfer on the rear is a proven tuning tool with our cars. As you'd expect for a long high speed constant corner radius corner you don't want too high RC heights (front or rear) but on street circuits its very good; going up on the RRC also allows us to run softer rear springs and maintain a similar 'steady state balance', which is never a bad thing on street circuits. Also a general observation; green tyres can obviously take more RC without punishing the tyre.

[Greg Locock]

nah, that was a 'Pat'ism. I still think they are fun to watch. I see from your sig you're admitting who you work with now!

well done

[Jean Fred]

Thank you very much for your answers,

"The most common misconception is that the vehicle rolls about the roll axis defined by the two roll centres. It does not." -> What defines the Roll axis, so?

What is the difference between Geometric Roll Center and Force Based Roll Center? I never heard about the latter.

What is a green tyre?

On F2005 & f2004 pictures i saw that upper front wishbone tend to be horizontal so there are a lot of chances that the swing arm is very long and roll center on the ground, F2001 2002 2003 wishbone seemed to be less horizontal so there is a tendency to put the geometric roll center on the ground, am I wrong? What makes toyota going against this apparent (for me) tendency?

[Greg Locock]

What defines the Roll axis, so?

Nothing simple, in general. For any particular set of forces and geometry there will, obviously, be a roll axis (instantaneously), but calculating where it is is first year mechanics, which I certainly cannot explain in a post like this. Try the first four chapters of Shigley and Uicker "Theory of Machines and Mechanisms" for a thorough description. You should be aware that the axis about which the body is rotating may not be the same as the axis about which it is accelerating.

"What is the difference between Geometric Roll Center and Force Based Roll Center? I never heard about the latter."

FBRC is defined as the effective point at which the lateral forces are applied to the sprung mass. If it were coincident with the CG then the body would not roll. As such it includes all sorts of second order effects. Even for quite common suspensions FBRC and GRC are in different locations, and may respond in different ways to changes in geometry. Initially I found this annoying and swore to use only FBRC, but I must admit, GRC has been a useful tool over the last few years, probably because my new suspension is effectively a double wishbone, hence its GRC behaves in a sensible fashion.

here's an old thread http://www.eng-tips....d=112722&page=2

and read the ortiz article that mcguire found in this thread

http://forums.autosp.....ht=roll cent*

You measure it by looking at the change in wheel vertical load as you roll the car.

[Jean Fred]

So you said that when we move up the RC, wishbone's flexibity alters less the ability of the geometry to rapidly transmitt loads to tyres because dampers are less loaded in roll motion and wishbones are more working in compression than torsion (if I have understood). and this is proportional until RCh = CG. but lateral weight transfert is more important.
excuse me I am french and translation is difficult for me

[Ray Bell]

Two questions on roll centres...

1. With a Watts Link, with the pivot on the axle, where is the roll centre?

2. (Lukin might appreciate this...) An item in a column of a magazine many long years ago... John Sheppard was the author... paraphrasing: "All this talk about roll centres! I was sitting in the john a couple of days ago and I realised I had the answer... there it was, the bit that's left over when all the paper's gone!"

[Lukin]

The pivot is usually pretty damn close to the actual RC, and it what we adjust, though it's not perfect. I forget whether its slightly above or below the pivot though. Im not much help sorry!

OT: Ray, Im back in Perth for a fortnight and saw the Veskanda yesterday. Unfortunately it's in pieces at the moment, which is a huge shame. http://members.iinet...jason/Veskanda/

[Ray Bell]

That's with the pivot mounted on the axle... how about on the chassis?

The Veskanda... it's certainly miles ahead of the old Bolwell that Bernie used to race...

[Lukin]

Originally posted by Ray Bell
That's with the pivot mounted on the axle... how about on the chassis?

We move the pivot point vertically and it's anchored to the chassis. Is that what your asking? Sorry mate, bit confused!

[Ray Bell]

I'm thinking of putting the pivot on the axle...

It's been done before.

[Lukin]

It might be, I have looked under there a thousand times but couldn't tell you for sure.

[Ray Bell]

I've had a lot of discussions over the optimum setup for these things... I recall Chris Dowd telling me about an early fifties Packard (yeah, big Yankee tank of a thing...) that had one, and it had its pivot mounted horizontally!

[Lukin]

Mac and Greg would be the two best blokes here to answer that one.

[Greg Locock]

A watts link with the pivot on the axle puts the GRC at the pivot, according to the simple diagram.

The reason is is that is the actual location where lateral forces are fed into the body, if you think about it.

However, I have a strong suspicion that that is not the full story if you incline the links. Ours only had parallel links, so I'm a bit hazy on that.

[Lukin]

Originally posted by Greg Locock
However, I have a strong suspicion that that is not the full story if you incline the links. Ours only had parallel links, so I'm a bit hazy on that.

If you draw a FBD with non-parallel bars, there will be a small couple at the pivot I would think, which would slightly offset the pivot?