Roll center and CofG...rate of roll or force of roll

[meb]

I am under the impression that the farther the roll centers are from the centers of gravity - in a dynamic condition...real world - the more forcefull the roll. On the other hand, if the distance between RC and the CofG is reduced, so is the force. But, is the rate of roll faster with a shorter lever arm and slower with a longer lever arm? There must exist a sweet spot within every vehicle where the force and rate of weight transfer is ideally suited to very fast driving.

[TEAM78]

Originally Posted by meb

I am under the impression that the farther the roll centers are from the centers of gravity - in a dynamic condition...real world - the more forcefull the roll. On the other hand, if the distance between RC and the CofG is reduced, so is the force. But, is the rate of roll faster with a shorter lever arm and slower with a longer lever arm? There must exist a sweet spot within every vehicle where the force and rate of weight transfer is ideally suited to very fast driving.

the further apart the roll centre is from the CG (when facing the front of the vehicle) the greater the rolling moment is.
I think you got confused and using the term force when its a torque.

The rate of roll is dependent largely upon the damper settings and is not really looked upon being dependent upon the distance between CG and RC. the distance largely influences the roll angle taken up for a given lateral g.

[quote]There must exist a sweet spot within every vehicle where the force and rate of weight transfer is ideally suited to very fast driving [quote]
This doesnt make sense to me unless im misreading it,the sweet spot is dependent upon a huge number of parameters and not just weight transfer.
I reconmend you to work from the tyre back into the chassis, learn tyre load sensitivity. work on each senario such as steady state cornering, the transients and linear acceleration and braking. how each parameter affacts the vehicle for each senario. e.g the dampers are very much influencial during the transiants but not at steady state senarios, this of course ignoring surface variations.
Its a massive subject which few people fully grasp and I certainly dont know near enough, so best take one step at a time, weight transfer is a good place to start though along with undertanding the basic functions of the tyre.
p.s: Merry Christmas everyone

[ubrben]

Originally Posted by meb

I am under the impression that the farther the roll centers are from the centers of gravity - in a dynamic condition...real world - the more forcefull the roll.

With a larger roll moment the moment of inertia of the sprung mass about the roll axis will go up (parallel axis theorum) this will slow roll rate and require stiffer roll damping to adequately damp roll without overshoots, further reducing response.

You need to consider the ratio of elastic to inelastic load transfer IMO because if you have a low CG and wide track you might get adequate response without running a lot of anti-roll percentage.

The idea of one "sweet spot" is confusing. Sweet spot for what? The driver's perception of performance, what the tyres need...?

The only way of proceeding is to consider some high level metrics like lateral load transfer distribution, ratio of elastic to inelastic load transfer, damping ratios in heave, pitch, roll then test see what you feel about the car then change one and see how it feels then.

There isn't a correct roll moment.

Ben

[meb]

My amateur statis is by now apparent . I'm a weekend warrior on the track and I like to know why and how things work...I cannot accept that they just do. So my uneducated background on occasion is prone to asking some really dumb questions.

...the sweet spot...torque is a much better word than force. My vision, which may be a disability, of the relationship between the CofG and roll centers and their migration is fundamental to how a car handles. I may have read into a reply about roll centers and this prompted the sweet spot question. Let me re-phrase. As a car turns, a torque is transmitted about the roll centers to the CofG, correct? If this is so, then the length of the invisible lever arm determines the amount of torque transmitted to the CofG? Does the length of the lever arm also independently affect the rate of roll? In this question, I'm concentrating literally on the relationship between the CofG and RC - meaning, what will the effects of a short or long lever be with regard to torque and rate of roll without consideration given to damping and springs rates? I'm looking for the most fundamental physical actions that occur, not components that control these.

I too use my dampers to control transitions thru turns. These transitions can either scare me or build confidence.

How would I use this information? I'm simply trying to establish an ideal relationship between track width modifications, ride height, and, spring rates for my 2005 JCW Mini cooper. The damping, swaybar adjustments and camber/toe adjustments I see as fine tuning.

If this helps, the tires are Toyo RA1s - 205/40/17. I would have preferred 16" wheels, but the 12.4" front rotors will not clear a smaller diameter wheel.

Thank you for your help,

Michael

[trikesrule]

Mini's - do they run independant suspension alround?.......trikes

[meb]

Yes, Mac strut up front and a multi-link setup in the rear; two links per side (upper and lower control arms) and one trailing arm/hub. These rear links, or upper and lower control arms are quite long; their inner pivots points are seperated by only a couple of inches. Mine are adjustable with spherical rod ends. The struts up front are adjustable coilovers and the upper perches are a pillow ball setup with up to -2.5 degrees of camber adjustment.

The lower control arm is positioned via a ball joint in the forward portion of the control arm and a dual compound poly bushing in the rear attachment point. The rear bushing is not stock. The control arm does not have a lot of compliance, in other words.

[trikesrule]

Meb you have the ability to perform Toe-in / out changes as well as Caster changes with those spherical rod ends. But as too altering roll centre on the rear I don't see that you easily can unless I missed something. With this teams cars if I lower the roll centre I get more loading on the outside wheels (increase roll). If I raise the roll centre I get more bite off the inside wheel(decrease roll). I'm talking rear wheel drive. The rear arms being so long will slow down reactions in heave happening at the back of the car when compared to the same car with shorter arms. Not a bad thing if your racing on a hilly circuit.........trikes

[trikesrule]

My view on r/c vs C of G is the C of G transmits a torque load thru the r/c'. Basicly I just work with what I've got to provide the driver with a balanced car. Does your car feel balanced? And yes a shorter arm reacts quicker than a longer one....trikes

[Lukin]

Originally Posted by trikesrule

With this teams cars if I lower the roll centre I get more loading on the outside wheels (increase roll). If I raise the roll centre I get more bite off the inside wheel(decrease roll).

I might be missing something here. Lowering the RC (assuming all other things remain constant) will reduce the load transfer on that axle in steady state cornering and therefore reduce the load on the outside wheel and increase the load on the inside wheel. It will increase roll yes, but load the inner and outer tyres in an opposite fashion to the way it appears are you described it. Look at a anti-roll bar; it reduces roll yet increases the load transfer at that end of the car.

I've never really referred to RC's in terms of torque arms or anything but as Ben said it, more a ratio of RC:CoG. The problem with trading elastic for geometric load transfer is the way and speed which they act (which you guys have touched on). Raising the RC at one end will unload the tyre quicker (I think Ben said somewhere it unloads at unsprung mass natural frequency (wheelshop), which makes sense. Elastic load trasnfers at the sprung mass natural frequency). As a result a RC change won't make a big change in steady state but will make a big difference at the initial turn and as the steering inputs are changing.

[trikesrule]

Hi Lukin re load transfer. It works just like I said at the track. Got the dust collectors to prove it. How much load if measured doesn't mean much to me. It's the ability to change / adapt the car to the changing conditions to keep the car at it's peak. I think we are talking about the same thing in a different language. If I lower the r/c the c of g rolls over the lower r/c easier. K.I.S.S works for me....trikes

[Goran Malmberg]

Originally Posted by meb

IThere must exist a sweet spot within every vehicle where the force and rate of weight transfer is ideally suited to very fast driving.

I may not answer the question right off what you ask for, I am just adding a few words to the debate.
There are a number of circumstances other than the "roll action" of the chassis that govern the geometry of the A-arms and thereby the Rc. Like If the car has much downforce, parallell to ground A-arms is often used to lessen the cambercompensation. More parallell to ground A-arms also makes it possible to let the intersection of the forcelines to follow load transfer precentage from side to side, thereby even out the influence of the jacking effect.
Roll is a way of having the car sprunged sideways therby adding to tiregrip,
(but of course roll also create camber problems). So, low CGH is imperative.

Myself I like close to ground Rc, longer and quite parallell A-arms. That is to keep the car more "Zero" (a word of my own for a car that has less inbuilt geomerty effects) fo an easier track tuning-setup situation. I also speak for less scrub and SAI numbers. Roll is then taken care of by springs, antirollbars and shocks. And, as allready mentioned, a low Rc.

Right now I am constructing a new "Nordic Supercar" based on the mentioned theory. I may put out som images later on when production has come that fare.
http://www.nordicsupercar.com/

Goran Malmberg

[ubrben]

Originally Posted by Goran Malmberg

I may not answer the question right off what you ask for, I am just adding a few words to the debate.
There are a number of circumstances other than the "roll action" of the chassis that govern the geometry of the A-arms and thereby the Rc. Like If the car has much downforce, parallell to ground A-arms is often used to lessen the cambercompensation. More parallell to ground A-arms also makes it possible to let the intersection of the forcelines to follow load transfer precentage from side to side, thereby even out the influence of the jacking effect.
Roll is a way of having the car sprunged sideways therby adding to tiregrip,
(but of course roll also create camber problems). So, low CGH is imperative.

Myself I like close to ground Rc, longer and quite parallell A-arms. That is to keep the car more "Zero" (a word of my own for a car that has less inbuilt geomerty effects) fo an easier track tuning-setup situation. I also speak for less scrub and SAI numbers. Roll is then taken care of by springs, antirollbars and shocks. And, as allready mentioned, a low Rc.

Right now I am constructing a new "Nordic Supercar" based on the mentioned theory. I may put out som images later on when production has come that fare.
http://www.nordicsupercar.com/

Goran Malmberg

I really don't understand why you're so obsessed by having no geometry effects. It seems you only reason is that they make it easier to understand. Forgive me, but that doesn't seem a very good reason.

Damian Harty makes some very good points about geometry effects in that they don't fundamentally alter the vehicle characteristics, but they can dramatically alter the perception to the driver. Therefore it seems odd to remove much of that opportunity by dumping any geometrical effects.

Ben

[Goran Malmberg]

Originally Posted by ubrben

I really don't understand why you're so obsessed by having no geometry effects. It seems you only reason is that they make it easier to understand. Forgive me, but that doesn't seem a very good reason.

Damian Harty makes some very good points about geometry effects in that they don't fundamentally alter the vehicle characteristics, but they can dramatically alter the perception to the driver. Therefore it seems odd to remove much of that opportunity by dumping any geometrical effects.

Ben

Actually I use the "Zero" car example as a way to explain the complex situation of many geometry angles involved in a race car. If we theoretically start off with a "Zero" car, we may then add just those components we do need, (for the driver or the car). Then those components will be balanced to the situation, no more no less a no unnecessary.
Goran Malmberg

[AU N EGL]

Originally Posted by meb

I am under the impression that the farther the roll centers are from the centers of gravity - in a dynamic condition...real world - the more forcefull the roll. On the other hand, if the distance between RC and the CofG is reduced, so is the force. But, is the rate of roll faster with a shorter lever arm and slower with a longer lever arm? There must exist a sweet spot within every vehicle where the force and rate of weight transfer is ideally suited to very fast driving.

This may help you slightly

http://www.racelinecentral.com/RacingSetupGuide.html

It is set up for NASCAR. However, make some wording changes for road racing, left AND right turns is not difficult to do.

Here is a quote from part of the information:

Where to Begin
Before attempting any chassis adjustments it is important for you as a driver to know & understand what your car is doing on the track. The only way to understand how your car is reacting is through seat time. The more laps you turn, or practice you get; the better off you will be in deciding what your trouble points are on the track.
Learn how to hit the same line lap after lap with the default setups packaged with NASCAR Racing before attempting any other adjustments. The biggest mistake rookie drivers make, are trying to adjust a chassis for what they believe is an ill handling car. I get so many emails from drivers that think the default easy, intermediate or fast setups that come with the sim are terrible because they cannot drive them. You must become comfortable with these default setups before you should even look into tinkering with any adjustments in the garage. Begin with the easy setup, then graduate to the intermediate setup and lastly the fast setup. Only when you can run clean consistent laps with the easy setup, should you move up to the intermediate setup. The same goes with the intermediate setup before graduating to the fast. Only when you can run competitively with the fast setup against the AI at 97% without spinning out, should you attempt to make any adjustments in the garage. There are nothing wrong with the default setups in NASCAR Racing. If you cannot drive them properly it is because you simply do not have enough seat time and or experience.
I can't stress enough the importance of adjusting only ONE component at a time. Making more than one change is nonproductive because you won't be able to determine what change made the car better or worse. Be sure to have plenty of paper & pencil on hand. It is critical to take notes after every adjustment you make. With all the adjustments available to make, it's real easy to forget what adjustment you made 3 practice sessions ago. Record keeping is important. If you do go the wrong way with an adjustment you can always set it back to where it was before making the change by simply referring back to your notes. By tracking & logging information, it also allows you to refer back to them to see how you progressed to cure your handling problem. These notes could be useful for setups at tracks with similar configurations which can turn out to be a real time saver. By keeping track of adjustments, you will be able to refer back to your notes to see what worked & what didn't. This will help you decide what to adjust if a similar situation arises at another track.

[meb]

Originally Posted by trikesrule

Hi Lukin re load transfer. It works just like I said at the track. Got the dust collectors to prove it. How much load if measured doesn't mean much to me. It's the ability to change / adapt the car to the changing conditions to keep the car at it's peak. I think we are talking about the same thing in a different language. If I lower the r/c the c of g rolls over the lower r/c easier. K.I.S.S works for me....trikes

Yes, the last sentence feels intuitively correct to me. But, something Lukin wrote is new to me; a car will roll more with a lowered roll center - all other things equal - but the load or weight transfer (what unit of measurement, pounds?) will be reduced...and perhaps proportioned a bit better between the inside and outside tires. The second part is new to me, but helsp me to understand the intricate dynamics at work.

So then, if weight transfers more readily - rolls over the RC - but is distributed more qeaully between the inside and outside tire, what linds of decisions should one begin to make with regard to spring rates? And part of my goal is to increase track by 15mm per side up front and 10mm per side in the rear.

Thanks for your help, this is great stuff!

[meb]

...my frustration is this mode of communication isn't really free association...so the fluidity in my questions and thinking mis-steps, sorry.

Anyway, if the above is true, a lower RC proportions weight more evenly between the inside and outside tires. Increasing track, as is one of my goals, will lower the CofG. Will lowering the CofG also proportion weight transfer between the inside and outside tires more evenly? My intuition says yes. This appears obvious, but the more I learn, the more I learn there are many, many subtleties invloved in weight transfer.

The rear of the Mini feels very squirlly under threshold braking. This was a stock vise and this feeling has carried thru to the new setup. Grip levels are not especially great either. Responsivness and balance is not a problem; this car goes where pointed. Braking into turn 1 at Lime Rock and into turn 9 at Watkins Glen can be white knuckle events, in particular.

The dampers may prove to be a culprit in those transitional areas, but I want to make sure some of the fundamentals (RC and CofG relationships) are correct before changing these.

[ubrben]

Originally Posted by meb

Increasing track, as is one of my goals, will lower the CofG. Will lowering the CofG also proportion weight transfer between the inside and outside tires more evenly?

I don't know what that means. How can a wider track lower the CG?

Weight transfer is mass * lateral acceleration * (CG height / Track width)

You will a get a little bit of extra load transfer to the outside because the CG has physically moved sideways due to roll, but this term is small for race cars which should have low CGs and low roll angles.

A wider track or a lower CG are the only way of reducing load transfer.

Springs and bars and roll centre heights ONLY effect the front:rear distribution of the load transfer and the rate at which it occurs.

Ben

[Goran Malmberg]

Originally Posted by ubrben

Springs and bars and roll centre heights ONLY effect the front:rear distribution of the load transfer and the rate at which it occurs.

Ben

Let me just add that roll centre height is the separation point between the geometric and sprung part of the weight transfer. However, the "rollaxsis" could be tilted (lower rc in the front and higher at the rear) for the same geometric weight transfer thereby altering the front and rear distribution.
Goran Malmberg

[meb]

Goran,

and I believe that Collin Chapman invested a lot of time working this little arrangment out. If I remember from some reading a long time ago, he hated rear swaybars. His ideal was to arrange roll centers and weight transfer so well that a rear swaybar could be avoided. I don't know if he actually applied this ideal practically.

ubrben...

Thank you for pointing the obvious out. I agree the height of the CofG cannot be altered by track but that the load on the outside tire may be reduced - in my case - if track is increased...the car bites into turns obviously harder if the fanny dyno is to be trusted.

But something else you wrote about touched on one of my questions above. I understand that springs and swaybars, and perhaps dampers, affect the rate of load transfer. I also understand intuitively how roll center migration will affect load transfer. I'm a bit lost about how RC locations and migrations as a car rolls will affect the rate of roll. does this have to with the RC's location at any given time with repect to the CofG? I rationalize in my own brain an invisible lever arm from the RC acting on the CofG. As a car is turned, load is transfered to some suspension pivot points and the CofG. Is the rate and load affected by the changing length of this lever arm? Shorter = faster rate of load transfer, but less load and longer = slower rate of load transfer but more load? Do I have this correct?

I should invest in a kinematics program so I can begin to really see what happens. Learning the values of the forces might be a next step.

[ubrben]

Originally Posted by meb

Goran,

I'm a bit lost about how RC locations and migrations as a car rolls will affect the rate of roll. does this have to with the RC's location at any given time with repect to the CofG?

The kinematic roll centre isn't a centre of rotation for the sprung mass. I wouldn't worry about it. The kinematic RC height is a reasonable approximation of the ratio of elastic to inelastic load transfer but the lateral position is useless.

Ben

[meb]

Okay, more great info, thank you. But a question comes to mind - and I really appreciate your patience by the way. Can the RC move to a place that might cause jacking? I don't believe anything that I've done can ever cause this to happen, but I've thought about it.

[ubrben]

Originally Posted by meb

Okay, more great info, thank you. But a question comes to mind - and I really appreciate your patience by the way. Can the RC move to a place that might cause jacking? I don't believe anything that I've done can ever cause this to happen, but I've thought about it.

What do you mean by jacking?

Ben

[meb]

There is a point where the RC can move to which will actually lift the sprung mass. I think I read this in Millikin(sp) If I remember, and this is based upon mant specific variables, if the RC approaches a relationship near horizontal with the CofG, the sprung mass will actually rise. hope I got that right...

[meb]

...also, I assume that RC location moves as camber is altered with regard to a Mac strut??? The critical angle is determined by the 90 degree relationship to the strut axis??? and not simply the angle of the strut tower?

[EfiOz]

Yes, but it's some pretty screwy geometry to make that happen. Volkswagen Beetles and old MBenz's with the swing axle rear used to do this.