×
INTELLIGENT WORK FORUMS
FOR ENGINEERING PROFESSIONALS

Log In

Come Join Us!

Are you an
Engineering professional?
Join Eng-Tips Forums!
  • Talk With Other Members
  • Be Notified Of Responses
    To Your Posts
  • Keyword Search
  • One-Click Access To Your
    Favorite Forums
  • Automated Signatures
    On Your Posts
  • Best Of All, It's Free!

*Eng-Tips's functionality depends on members receiving e-mail. By joining you are opting in to receive e-mail.

Posting Guidelines

Promoting, selling, recruiting, coursework and thesis posting is forbidden.

Students Click Here

Roll center migration

Roll center migration

Roll center migration

(OP)
Hello,

I was wondering how roll center migration affects suspension performance during roll movement.

I have carried static suspension calculation, and I did not see any releveance of non migration roll center suspension, but my calculation are static not dynamic.

Do someone had experience on no roll center migration suspension and how it performs?

Regards

RE: Roll center migration

Millions of cars have been built with MacPherson front suspension which inherently has enormous instant center migration. They work well enough for most people.

The only type of independent suspension that does not have migration relative to the bodyshell are the swing-axle-type (including Ford Twin-I-Beam) or single trailing or semitrailing arm type, that have a single defined pivot axis fixed to the bodyshell. They all have bad side effects. The "cure" (to the imaginary/nonexistent "problem") is worse than the "disease" (that the instant centers are not fixed in space relative to the bodyshell).

The only types of suspension that do not have migration relative to ground level are certain rigid-beam-axle types, depending on the design of the mechanism that locates the axle side-to-side.

What imaginary problem are you trying to solve by not having instant centers that move around?

RE: Roll center migration

(OP)
Mac pherson is a cheap solution for production car, wich you are right work well for most people.

I was dealing about upper and lower link wich are use on race car.

On this configuration, if the link are not too short, it is possible to design a suspension with no roll center migration (without computer).

Most of book carry calculation with roll center, as this point is often moving, this could be a problem, and calculation went wrong or need further itterations.

Some author mention a good suspension is when the distance of roll center to the center of gravity, do not vary.

I was only asking if some people on this forum have already test or experience suspension with upper and lower link with no roll center migration relative to center of gravity?

RE: Roll center migration

The instant-center of an upper and lower A-arm suspension will most certainly move relative to the bodyshell or center of gravity, especially in two-wheel bump. If it's in bump (or if someone has lowered the suspension!) the upper and lower arms will both be pointing and intersecting at a different spot than they will at nominal ride height. The ONLY way the instant center stays in the same exact spot relative to the center of gravity is if the suspension physically has its single pivot axis there. That's Ford Twin-I-Beam, or pure trailing or semitrailing arms (or its extreme variant, swing-axle). An upper and lower A-arm design cannot do this ... unless the A-arms physically are extended across the car to meet in the same physical point ... which is swing-axle geometry (or Twin-I-Beam which is the same concept).

In dynamic situations you really need to be using force-based analysis, not necessarily focusing on the "roll center" which (for most independent suspension designs) is an imaginary construct anyhow. What's happening at the outside wheel matters far more than what's happening at the inside wheel because there's much more force on the outside wheel.

Upper and lower A-arm suspensions allow much more design flexibility in terms of camber gain, anti-squat, etc and they remove some practical geometry limitations imposed by MacPherson, but the instant center motion is a small part of this.

It will help to know what sort of vehicle you are talking about here. For example, Formula 1 cars don't have much camber gain with suspension movement and the instant centers are probably close to ground level far on the opposite side of the vehicle ... but it doesn't matter, because the suspension is so stiff that it hardly moves, and it doesn't have to move much because they're designed for use on smooth tarmac, so they have extremely high spring and damping rates. If you are talking about a World Rally car, or perhaps a desert-racing truck that is expected to go over uneven terrain at high speed ... you have a much greater challenge at hand, because now it has to function over a wide range of movement.

RE: Roll center migration

(OP)
Brian,

I prefer to join you a picture of the suspension I mention (sport car , formula ford, etc). It is 2d example, not 3d, but a good start.

Formula 1 is not interesting, because wishbone are placed for aero reason (mainly at the front).

You mention instant center movement, but I never mention IC movement, I understand your thinking: for no roll center movement you need no IC movement.

I only mention distance from roll center to COG keep the same.

If you have an example of suspension to submit me, and if you leave me one point free, I will be happy to show you that it is possible to limt near minimum , roll center movement regarding COG.

To minimize roll center migration relative to COG, it is the same construction problem as having no bumsteer or no cvd plunge, it is a 3 link compatibility problem.

To my knowledge, Instant center is an important point, this position deals with track variation and camber gain, the tangent of this point deals also with camber gain (length ratio between lower and upper links).

RE: Roll center migration

(OP)

RE: Roll center migration

I don't have a method of creating an illustration and posting it here.

Now ... Take that diagram and draw in the instant-centers. I'm going to guess that the instant-center of each side (where the projections of the control arms intersect) is about 1 vehicle-width on the opposite side and close to ground level.

Then ... Lower the bodyshell by (let's say) 50mm and re-draw your instant-centers. They'll be below ground level.

Then ... Go back to nominal ride height and re-draw with 5 degrees of body roll. It's quite possible that the instant-center of the outer wheel could end up pretty close to ground level. I don't know where the inner one will end up but given that the arms will be going closer to parallel, it's likely to be far off to the opposite side of the car. But we don't care much, because most of the weight is on the outer wheel, not the inner one.

So, you have fairly low vertical motion of the instant-center of the outer wheel (the one that matters) in roll (and we don't much care what the inner one does), but you have considerable motion in two-wheel bump (it goes below ground level).

The reason I refuse to extend the discussion to the "roll center" is that it's an artificial construct because the concept doesn't account for most of the forces involve the outer wheel rather than the inner one.

Now ... Explain to the rest of us what it is that you are actually trying to accomplish. "The big picture".

Circuit-racing cars have instant-centers close to ground level at nominal ride height (not necessarily at other ride heights - but these cars always carry the same well-defined payload) and with minimal vertical motion in roll. "So what".

This type of car normally has really stiff springing and dampers. The suspension hardly moves, anyhow. If you hardly let it move, it hardly matters what the instant-centers do. Circuit-racing cars of this type use that suspension layout because it works with the packaging and it works with the aerodynamics, and the geometry is good enough. (MacPherson would not be a good choice!)

It gets MUCH more complicated when you go desert-racing and you actually need large suspension movement. It also gets MUCH more complicated when you try to develop a road car that is expected to have good steering characteristics and good grip and good ride quality and do so whether the driver is alone on board or is carrying 3 passengers. And then you end up with MacPherson front and twist-beam rear anyhow because the bean counters dictated thus ...

RE: Roll center migration

(OP)
Brian,

My initial post was regarding if anyone has tried a suspension with no Rc-COG variation, and what was the result?

You post, by trying to explain me that on Upper and lower A-arm suspensions, achieve no Rc-COG variation is not possible because Ic move.

I reply: it is possible to minimize RC-COG, and ask you to submit an example, so I can show it is possible, you reply : I don't have a method of creating an illustration and posting it here.

So what I say is: I have a method to construct a twin arm suspension wich minimize RC-COG distance, I posted below an example:
upper figure are +-50mm travel
middle figure is static
lower figure is 5° roll

same wich Ic drawn

Ic is moving but RC-COG alter of 3mm

The big picture is: I create a program calculate every force on suspension element under lateral and vertical acceleration, wich give me a good idea of what happen in a steady state corner (roll).
I find a construction to design parallel twin suspension with small RC-COG distance variation. I did not see any interest of this with my program, but I have the feeling that it could be a plus in dynamic, so I ask if somebody had always tried a twin arm suspension with small RC-COG variation, and if yes what was the result?




RE: Roll center migration

Your first diagram isn't what I meant. You drew the left side 50mm into bump and the right side 50mm into rebound. What I meant was the situation where BOTH sides are 50mm into bump, or BOTH sides are 50mm into rebound. You should find that the former situation places the roll center below ground level, and the latter situation places the roll center above ground level.

Lots of race cars - practically all open-wheel race cars - have been designed with suspension generally like what you are illustrating, which has relatively low instant-center height movement in roll. For that matter, lots of street-driven cars are like that, too. And it's been done since the 1950s if not before.

RE: Roll center migration

A raising "roll center" results in progressive roll stiffness. Laterally it does not matter.

RE: Roll center migration

(OP)
Brian,

I m not dealing with Rc relative to the ground, I m dealing with RC to center of gravity COG distance, which is generally use to calculate roll moment in litterature.

'Lots of race cars - practically all open-wheel race cars - have been designed with suspension generally like what you are illustrating, which has relatively low instant-center height movement in roll. For that matter, lots of street-driven cars are like that, too. And it's been done since the 1950s if not before.'
I do not know what you want to explain, and I do not understand the sense, if you want to say that kind of suspension is well know, it is evidence.
what is more evident is to keep RC-COG variation small is less evident and most author mention the importance of limiting Rc movement regarding COG.

Buggar,

'A raising "roll center" results in progressive roll stiffness. Laterally it does not matter'.

Do you have experience or calculate this affirmation?

RE: Roll center migration

Draw the diagram that I suggested. You've drawn it at nominal ride height. Draw it again with the body 50mm below nominal ride height, and draw it again with the body 50mm above nominal ride height. With that suspension design, the roll center moves relative to the center of gravity ... it is not a fixed distance from the center of gravity. DRAW THE DIAGRAM.

If you don't know what I'm explaining ... I don't know what you are asking. I have a feeling that English is not your first language, and something is getting lost in translation.

"Do someone had experience on no roll center migration suspension and how it performs?" Well firstly, the roll center DOES move, so claiming that it does not move is not completely correct. It may move less than with other designs, but it does move, so that's part of the response to this question from your first post. If your question then means "Does someone have experience with suspension designs that have relatively low motion of the roll center relative to the center of gravity", given that this general suspension design has been in use for 70 years if not more, the answer would be "yes". And at this point, I don't know what you are asking.

RE: Roll center migration

Case in point was when we went from watts link beam axle to IRS. RCH vertical migration is roughly 0mm/mm jounce for the Watts, and say 0.5 mm/mm for the IRS. I've never found any physical meaning for lateral migration.

On a smoothish track or public road there was no difference in everyday handling. Sure on rough roads it was easy to pick (our proving ground's main access was a 5 mile gravel road).

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376: Eng-Tips.com Forum Policies http://eng-tips.com/market.cfm?

RE: Roll center migration

And I'm sure most of that rough-road difference would have been due to unsprung weight from the differential bouncing up and down together with the axle.

RE: Roll center migration

(OP)
Brian,

I DRAW the diagram as you asked
ride height -50 from static

ride height +50 from static

So Rc to COG distance alters by 1.4mm.

I did not DRAW the IC line, I can do but it will make small picture difficult to read.
I can sent you the file (dxf, dwg) and you be able to control by yourself that I did not cheat, and I know how to DRAW a roll center.

On my previous post, I draw one side +50 and the other -50, as far as it is symetrical, I put on one wiew 2 cases, may be it was not clearly understandable for you.

English is not my native tongue.


RE: Roll center migration

(OP)
Greg

'ase in point was when we went from watts link beam axle to IRS. RCH vertical migration is roughly 0mm/mm jounce for the Watts, and say 0.5 mm/mm for the IRS. I've never found any physical meaning for lateral migration.'

You means rc migration relative to the ground, but relative to COG it is not constant in compression, rebound.

RE: Roll center migration

(OP)
Greg,

What did you experience or modeling in order to affirm roll center position has no importance?

RE: Roll center migration

Lateral geometric roll center location doesn't enter into the SAE's definition of suspension roll center.

Nor does lateral RC location enter into lateral load transfer summations done up as LLT through the roll centers + LLT through roll and suspension roll stiffnesses + LLT due to unsprung mass effects.

I think Brian is hinting at taking a lateral-anti approach, where the inboard and outboard force effects are neither equal nor symmetrical.


Norm

RE: Roll center migration

Yup, I just gave up trying to explain. In a real car with actual suspension travel and a non-zero center of gravity height, neither the geometry (due to suspension movement in roll) nor the force distribution (due to C of G height combined with front/rear load distribution due to all of the roll stiffness components) are symmetrical left to right, so what is the relevance of an imaginary geometrical intersection of two lines on a drawing?

The instant-center of each side can be a useful concept, because if you know or can estimate the left-to-right load distribution, you can work out the vertical components of the forces on each side separately - at least for steady-state cornering.

It will turn out that it more-or-less matters for the outside wheel, and it more-or-less doesn't matter for the inside wheel.

Hence why the millions of cars built with MacPherson front geometry work well enough even though the instant center in droop (inside wheel) does all sorts of screwy things. The weight isn't on that wheel, so it doesn't matter.

"Roll center too high at nominal ride height or in bump = BAD". That's about all that really matters.

RE: Roll center migration

(OP)

Brian,

You wrote: 'Well firstly, the roll center DOES move, so claiming that it does not move is not completely correct.', I spent some time to design with your request (+-50 mm, etc).
Do the image I posted convince you that is possible to limit this movement at the minimum?

I undertsand your meaning regarding with production cars wich mass are not centralize, so drawing a central point has no meaning, but on most race car, weight is central, or at least it can be set up with corner weight.

MacPherson work well enough on production cars, but mainly the use of MacPherson is relative to cost ,and the possibility to leave space for the engine, but it generates friction and camber gain are small. I never see a proper race car (signle seater or sport prototype) design with MacPherson.

For a twin arm suspension, regarding IC, vertical distance IC to the ground define the amount of track variation, you will get. Lateral position of IC define the camber gain, but if you draw the IC tangent (wich belong to upper and lower link ratio) it define also the camber gain linearity.

When you mention the useful concept of IC, do you refer to WC Mitchell FAP document?

Regards

RE: Roll center migration

Certainly you can minimise the roll center movement. You've shown that you have done so. My point is, "So what". It doesn't matter, as long as the instant center of the outside wheel when loaded up while rounding a corner isn't too high, and as long as it isn't too high when rolling down the road. Beyond that ... It doesn't matter.

I'm not talking about static location of the center of gravity. I'm talking about the instantaneous forces on the wheels while the car is rounding a corner. https://i.pinimg.com/originals/b1/6e/9e/b16e9e5195...

How relevant is the position of the instant-center of the left front suspension of this car at this particular moment in time? https://c8.alamy.com/comp/BAC3NC/classic-porsche-9...

In terms of the instant-center heights ... the only thing that matters, at that particular moment in time, is the geometry of the right front suspension. The left front suspension is - literally - just along for the ride. (and by the way, that car uses MacPherson front suspension!)

RE: Roll center migration

Bluefoxy,
Yes I have my "Chassis Roll Study" in different files all over my computer. I have some in pdf file format which this forum doesn't eat. I have posted some of my studies on this forum but have no idea where.

RE: Roll center migration

Just a bit more going on besides load transfer in cases where the 'roll center(s)' i.e. 'roll axis' shifts around during cornering. The sprung mass exhibits most of this motion which you can easily see (for example) if you watch our 'rail shipping simulator'. Each end of the vehicle moves laterally in accordance with the ground plane side forces.

So, roll axis displacement laterally not only shifts the weight transfer trajectory, but it moves the nose and/or tail of the vehicle, often producing sensation of yaw angle or sideslip motion that may or may not be a real good cue to the driver.

Make up a soda straw model of the unsprung and sprung masses with the 'roll center' points connected to the unsprung axles via springs from a couple of writing pens. A hot glue gun is a handy way to put this together. Then push on the sprung C.G. and watch for yourself as the sprung mass orbits. This is why the front and rear motions need to be 'coordinated.

Makes for a great video, especially when Safety Lab targets are placed on the bumpers and photogrammetry analysis is done.

RE: Roll center migration

By logic, the height of the "roll center" is a measure of chassis roll stiffness when subject to lateral force. If the "roll center" is rising, the chassis roll resistance is tightening up.
(I dislike the term "roll center", but I realize what is meant).

RE: Roll center migration

Here's the hardware for this experiment. They usually put signals into it from actual rail car rides across the country. There are certain spots on some routes where enormous accelerations are recorded. (including sustained 9g steady state cornering forces. [Bet you can't figure out how this happens: Too many propeller heads playing with their 'unrepresentative' model trains at work].

RE: Roll center migration

There is a paper around where somebody calculated the geometric roll center height, and the force based roll centre height, as a vehicle went around the corner. They moved in different directions. GRCH is basically an instantaneous centre of rotation construct, it doesn't seem to me that that automatically translates into something drivers would care about.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376: Eng-Tips.com Forum Policies http://eng-tips.com/market.cfm?

RE: Roll center migration

(OP)
Brian,

On this particular case, I will be pleased if you could tell me more about the instantaneous forces on the wheels :

RE: Roll center migration

(OP)
cibachrome

'So, roll axis displacement laterally not only shifts the weight transfer trajectory, but it moves the nose and/or tail of the vehicle, often producing sensation of yaw angle or sideslip motion that may or may not be a real good cue to the driver.'

It sounds logical.

RE: Roll center migration

Quote (Bluefoxy)

Brian,
On this particular case, I will be pleased if you could tell me more about the instantaneous forces on the wheels :
I will have a guess and say at some point prior to the photo, a large roll moment was generated.

je suis charlie

RE: Roll center migration

It fell over. smile

RE: Roll center migration

The rear suspension is the notorious twist axle with the cross beam centroid and shear center facing downward and foreward a bit. That means it has -10% --> -12% rear roll steer (as in roll oversteer). It takes a mighty stiff tire to compensate for this at speed below 140kph. With those stiff tires, the roll and yawrate frequencies hooked up and the driver ran out of courage.

The camber of the front and rear wheels is permanently deformed due to the wheels sinking into the ground and the vehicle was tripped while going sideways. Look at the tracks. Loss of control due to the very high gain of the car, low understeer [less than 1.25 deg/g] and rain tread for wet traction. Axle loads > 5g (typical design max).

Then driver screwup! Just push it over and drive it away. Might want to clean the seats first...

RE: Roll center migration

The fact that macP front twistbeam rear can be made to work successfully is probably proof that in the vehicle dynamics world we spend a lot of time worrying over stuff that doesn't matter, or perhaps that we are such geniusses that we can make even bad solutions work. I wonder which it is.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376: Eng-Tips.com Forum Policies http://eng-tips.com/market.cfm?

RE: Roll center migration

(OP)
Greg,

MacPherson work well enough on production cars, but mainly the use of MacPherson is relative to cost ,and the possibility to leave space for the engine, but it generates friction and camber gain are small. I never see a proper race car (signle seater or sport prototype) design with MacPherson.

I created this thread regarding race car suspension, upper and lower arm, not MC Pherson. And ask if someone had ever try this kind of suspension with a roll center to distance remain constant and if they see any interest?

Brian mention: it does not matter because of assymetrical and so on, and it was not possible to have RC to COG distance remains constant. I spent time DRAW, to show him it was possible.

You mention watts link, clearly it does not keep RC to COG distance constant.
You mention Force based roll center, I reply by joining the Mitchell document which deals with Force based roll center. I do know not if you read it

If your reasoning is: On Mac Pherson, roll center migration is massive , and Mc Pherson work pretty well for production car, so thinking about migration roll center is irrelevant. It is taking 2 facts without any link, to conclude something wich allow you to not ask you question.



RE: Roll center migration

Quote (Bluefoxy)

I created this thread regarding race car suspension, upper and lower arm, not MC Pherson. And ask if someone had ever try this kind of suspension with a roll center to distance remain constant and if they see any interest?

More or less, every open wheel race car since the 1950s.

Who do you expect to be "interested"? (Do you think people who have been working with vehicle dynamics their whole career don't already know how to do what you are suggesting?)

RE: Roll center migration

No interest because no one in their right mind uses kinematic roll center technology. Its an artifact of desk calculators and drawing board technology. Asymmetric suspensions rule out the use of this analysis. Mother Nature does not use kinematics for anything. [Just like she doesn't do a matrix inversion everytime she integrates accelerations to get movement]. "Yes or No ?" Take Claude's course, it's covered starting on page 666 in the spiral notebook.

RE: Roll center migration

(OP)
Brian,

'More or less, every open wheel race car since the 1950s.' It is your opinion,I do not think it is clever but I respect.

For sure and in any case it will interested people who had spent their whole carrier working on vehicule dynamics, and end up thinking: Mac Pherson works pretty well.

Cibachrome

'No interest because no one in their right mind uses kinematic roll center technology', that mean Eric Broadley, John Barnard, Robin Herd, Caroll Smith, Allan Staniforth loose their time with roll center?

Do one of you have work in race car design office?

RE: Roll center migration

Yep. What's your point? Race car engineering is relatively easy. Production Vehicle Dynamics is much harder. Comfort, cost, rules and requirements, foreseeable misuse, aftermarket parts, drivers and operators with limited skills and abilities, high mileage durability, maintenance, wearout, production volume, global specifications, multiple suppliers of the 'same' part, safety, packaging, customer xpectation, fuel economy. The list can go on.

Once you've had a few 'racecar' prototypes (spelled both forwards an backwards) on K&C machines, simulators and tire test machinery, you realize that all they are is super tires, super motors, super drivers, super expensive, titanium, kevlar and not a great long term career move for engineering staff. But pretty simple to analyze. Small angle approximation works for everything except crankshaft rotation.

BTW: show us a graph of your comparison of kinematic roll center vs. FB 'roll center'. Jus a leetle beet of kombliance make all bets off. Yes or No ?

RE: Roll center migration

(OP)
Race car engineering is relatively easy: In the sixties, it took million USD and 3 years to Ford to beat Ferrari at Lemans. Look how many manufacturer involve time and money to never win a F1 race or Lemans.
This affirmation have no sense.

In my carrer, I have been working as engineer for production manufacturer and race car manufacturer. Production and race car design are both not easy.

BTW: show us a graph of your comparison of kinematic roll center vs. FB 'roll center': FB you mean Force Base according definition of Mitchell?

'Jus a leetle beet of kombliance make all bets off.', sorry but english is not my native tongue, I tried google traduction but did not get the sense of it.

RE: Roll center migration

Quote (foxy)

Mac Pherson works pretty well.

As only an amateur at this stuff, I suspect that a MacStrut works as well as it does . . . more in spite of itself than because of any kinematic virtue.

The fact that you have to introduce a virtual UCA to establish MacStrut kinematics in the first place (an artifact with a fixed angular relation to the knuckle, no less) ought to be a clue.


Norm

RE: Roll center migration

(OP)
Norm,

I did not open this thread about Mc Pherson kynematics.

'The fact that you have to introduce a virtual UCA to establish MacStrut kinematics in the first place (an artifact with a fixed angular relation to the knuckle, no less) ought to be a clue.', English is not my native tongue, I did not understand the sense of what you mean, if you could put a drawing to explain, I will be thank full.

RE: Roll center migration

I realize that the MacStrut wasn't the original focus of this topic, but it's perhaps better to consider it anyway simply because a consistent approach to the usefulness of kinematic roll centers is either going to work equally well for the strut or weaknesses/limitations in the approach (small angle approximations more likely to become invalidated) might be more apparent because geometric roll center migrations are more severe with the strut suspension.

Pictures . . . the first is from Fred Puhn's "How To Make Your Car Handle (enthusiast-level book, 1976-ish)
https://www.mustang6g.com/forums/media/macstrut-ge...

The second is from a spreadsheet construction of my own.
https://www.mustang6g.com/forums/media/fvic-and-ti...


Norm

RE: Roll center migration

(OP)
Norm,

Thank you for the picture, I now understand what you met.

I will be happy to know your approach.

RE: Roll center migration

As y’all know, I don’t like the term “Roll Center”. I suggest:

The location at which the lateral force may be applied to create a roll moment about the center of gravity of the chassis equal to the chassis roll moment induced by the individual laterally loaded tires as transmitted through the suspension links to the center of gravity of the chassis.
It is a measure of chassis roll flexibility under lateral force, and it must be resisted by equal and opposite spring/ARB roll resistance.

RE: Roll center migration

yeah, I was looking at a website that somebody recommended for front view kinematics, and he'd put up a bug fix saying he couldn't figure out why it didn't roll around the roll centre. It gets even better when somebody connects the front and rear roll centres and calls it the roll axis.

Mind you to be fair Carroll Smith did include that in one of his books originally.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376: Eng-Tips.com Forum Policies http://eng-tips.com/market.cfm?

RE: Roll center migration

(OP)
Buggar,

The location at which the lateral force may be applied to create a roll moment about the center of gravity of the chassis equal to the chassis roll moment induced by the individual laterally loaded tires as transmitted through the suspension links to the center of gravity of the chassis.

It is more or less the definition of Mitchell FAP (file attached).

I do not share this definition.

RE: Roll center migration

(OP)
Greg,

You think FBRCH is an useful definition?

RE: Roll center migration

(OP)
And you still see any any interest of my suspension with nearly constant distance from RC-COG?

RE: Roll center migration

(OP)
Greg,

You are agree with Mitchell, when he wrote: 'Stability results when the FAP-CG moment arm remains constant as the vehicle rolls. The chassis “takes a set” rather than constantly seeking a new equilibrium. This can be expressed by minimizing the lateral movement of the KRC as the vehicle rolls. But this is an artifact: there are more direct ways to calculate this; namely with the change in FAP height resulting from ride. It should be one-to-one. (An easier way to visualize this is from the viewpoint of the chassis rather than the world. The FAP point should be constant as the wheels and tires move up and down.)

RE: Roll center migration

Bear in mind that the FAP is most relevant for the side that is more heavily loaded (the outside wheel) because of the way the FBRCH is calculated. And for that ... I bring back your favorite topic ... the MacPherson strut.

The instant center will be somewhere along the line that is perpendicular to the upper strut mount. In most normal applications (certainly both of the vehicles in my driveway), that strut is only a few degrees away from vertical. So the instant-center will be somewhere along a few-degrees-off-horizontal line through the upper strut mount. With practical and sensible geometry (lower arm somewhere near horizontal at nominal ride height and not unreasonably short), that is going to lead to a FAP that is near ground level. The instant center of the unladen side (the inside of the corner) will be doing all sorts of screwy stuff, but it hardly matters because there's hardly any load on that side. So, the FAP-CG moment arm is going to be somewhere near the height of the CG give or take a little bit.

Out back, if I look at the two vehicles in my driveway right now, in one case it has a twist-beam axle, which will have a roll center (whether force-based or kinematic) not much above ground level, and in another case it has a suspension design that you will surely love ... a dead-beam axle on leaf springs! The roll center (whether force-based or kinematic) is at leaf-spring height. So in both cases, that CG to FAP moment arm doesn't change much.

Whatever ails the handling of both vehicles (and they've both got ailments), isn't because the roll center moves around with suspension movement ...

RE: Roll center migration

(OP)
Brian,


As I far I understand, with few degree away from the vertical, It make an IC wich move on a line which nearly horizontal, which leads to a line Center of tire-IC, which is also quiet horizontal.
As the suspension compress (outside wheel) ,tire-IC went even more horizontal. So FAP of outside wheel does not alter a lot, and stay more or less near the ground.
As you do not care of the inside wheel, FAP of outside wheel become FBRC. And FBRC-CG remains constant.

If I well understand, according to you, a suspension which has an FBRC-CG constant is a suspension that is working well?

RE: Roll center migration

Quote (foxy)

You are agree with Mitchell, when he wrote: 'Stability results when the FAP-CG moment arm remains constant as the vehicle rolls. The chassis “takes a set” rather than constantly seeking a new equilibrium. This can be expressed by minimizing the lateral movement of the KRC as the vehicle rolls. But this is an artifact: there are more direct ways to calculate this; namely with the change in FAP height resulting from ride. It should be one-to-one. (An easier way to visualize this is from the viewpoint of the chassis rather than the world. The FAP point should be constant as the wheels and tires move up and down.)
This "taking a set" . . . do you think that is strictly a displacement effect (geometric steady-state)? Or could it be an approximation of force equilibrium using a time history approach to forces associated with roll centers and the various elastic suspension components? Seems it might matter, since "taking a set" is something that the driver perceives.


Norm

RE: Roll center migration

Quote (Bluefoxy)

If I well understand, according to you, a suspension which has an FBRC-CG constant is a suspension that is working well?

But there are a thousand other factors that are a higher priority.

Examples of suspension designs that have FBRC-CG dimension more-or-less constant in roll include: Rigid beam axle and leaf springs (both FBRC and kinematic RC at leaf spring height), pure trailing arms (both FBRC and kinematic RC at ground), swing axles (both FBRC and kinematic RC at swing axle pivot point height - which is fixed to the chassis i.e. the CG!), Ford Twin-I-Beam (instant centers at chassis-end I-Beam pivot points), double trailing arms traditional VW Beetle front suspension (FBRC and kinematic RC at ground).

Only car I know of with swing-axle-geometry front suspension: https://www.shannons.com.au/library/images/news/Q6...

Hark the Herald axles swing! https://www.canleyclassics.com/images/infodatabase...

So here you have a collection of bad suspension designs that have FBRC-CG at least more-or-less constant in roll, if not necessarily in two-wheel bump or rebound.

And yet, MacPherson which is no better or worse than these in terms of that specific measure, works OK.

And if you wish to limit the discussion to upper-and-lower-wishbone designs ... A Formula 1 car has upper-and-lower-wishbone suspension layout. So does a 1955 Chevy. The F1 car is not particularly designed to keep the FBRC at constant height ... well, let me correct that. It does keep the FBRC at constant height (near ground level with almost-nil camber gain with suspension movement), by using brutally stiff springing and damping so that the suspension hardly moves, in which case, it doesn't matter!

Look at it this way. A FBRC that is slightly above ground level and doesn't move much in bump travel isn't something that you primarily design for, it is something that "falls out" when you impose all of the other things that you want the suspension to do in order to work the way you want it to work - and if it happens to move more than you'd like it to but your other constraints are satisfied, it doesn't matter.

RE: Roll center migration

(OP)
Brian,

So I do not understand why you do not like the concept of twin arm suspension with constant distance RC-COG.
Because as far RC-COG remain constant, both FAP are very near (you even do not have to negligate the inside wheel), it result a FBRC -CG distance constant:


RE: Roll center migration

(OP)
Brian,

Regarding Mac Pherson,

I take the time to DRAW, because DRAW help me to understand what happen.
Regarding the position of FAP, it stay constant until the lower arm get the same angle as the perpendicular to the upper strut mount, if it come more angle it went totally opposit.

Both suspension you mentioned with FBRC-CG constant ha a massive jacking effect.

RE: Roll center migration

It's not that I "don't like" that design.

It's that I'd rather focus my efforts on other things that achieve proper function, and let the roll center be wherever it ends up.

RE: Roll center migration

(OP)
At least, recognize FBRC-CG stay constant near the ground!

RE: Roll center migration

You're acting like you're trying to use that as a selling feature. I'm trying to explain why you won't have any success with that.

RE: Roll center migration

(OP)
I m not trying to sell anything.

I find a way to design a twin arm with Rc not moving, which many author claim it is a good thing. honestly I have no clue if it is good not or not, so my initial question on this thread.
I never give an attention to FBRC concept, because as far as you deal with force the law in static sum of force = masse x acceleration, and then you calculate every force on every compenent and you are able to conclud something or not. FBRC only deals with lateral.
In my point of wiew, for a steady state corner, a suspension is a mechanical system wich mass is summit to vertical and lateral acceleration, on this mecahnical system, only spring are deformable, once you have calculate every force on each component, and end up with the force on the spring and thus their deflection, you could have an acceptable wiew of what happen in roll.
For having done this, the suspension does not roll around roll center definitely as long as you get jacking effect, and as long as the track alter between static and roll.

What I dislike in the Mitchell paper, it start from static to a roll position, then include force, then include FAP. the question is how his value of roll is done (I imagine symetric, which is rarely the case due to jacking effect).

You wrote 'Bear in mind that the FAP is most relevant for the side that is more heavily loaded (the outside wheel) because of the way the FBRCH is calculated', I was thinking it was important for you, so I notice RC-COG constant means FBRC-Cg constant.

If you mention the nowadays F1 suspension, you have to keep in mind, that wishbone are place this way only for aero reason, It is not me who say that, it is written in Adrian Newey book.

'I'm trying to explain why you won't have any success with that.', I will be pleased to know.

RE: Roll center migration

Much of the attention paid to roll center position and height is due to the excitation and transmission path. Roll is not induced buy steer (OK a bit by caster) but by lateral acceleration and lateral force. We refer to it by the mathematical term Convolution. It is a process by which steer induces lateral forces / accelerations followed by the roll response as a secondary characteristic. The timing of all this is worth studying because placement and movements of the unsprung mass(s) force points defines the initiation of events. In closed loop control (driver-in-the-loop), the lateral position and heading angle implied to the driver modulates their steering and throttle/braking action. Movement of the nose is part of this signal conditioning feedback. That's why a 'hood ornament' makes a good driver training and engineering feedback aid. It implies a heading correction to be anticipated. If sprung mass front end is orbiting the desired path instead of staying tangent to it, then excess control and poor steering judgements result. It's all about perception. That's why even seat position has a great influence on a driver being able to 'get up on the wheel'. Routine ISO testing doesn't reveal this because the steering is open loop (no feedback correction).

It DOES show up on the ISO Weave test (sine steer), though. With simultaneous path, frequency, and g-level constraints, some vehicles are very difficult to drive. The indications of this are in the variation computed during multiple steering cycles. We measure it, integrate it and produce a score to rate he vehicle. You should not be surprised to see a list of best to worst vehicles and their measured K&C data for roll, friction and tire relaxation properties. AND why this variation does NOT show up with a mechanical, robotic driver. Check it out !

https://www.youtube.com/watch?v=S4O7b0-epBo

RE: Roll center migration

Graphics attached.
I'm attempting to load a pdf file and it won't load! It is my treatise on the "roll center".

RE: Roll center migration

(OP)
Sorry but nothing is attached

RE: Roll center migration

Quote (ciba)

Much of the attention paid to roll center position and height is due to the excitation and transmission path. Roll is not induced buy steer (OK a bit by caster) but by lateral acceleration and lateral force. We refer to it by the mathematical term Convolution. It is a process by which steer induces lateral forces / accelerations followed by the roll response as a secondary characteristic. The timing of all this is worth studying because placement and movements of the unsprung mass(s) force points defines the initiation of events.

In other words, this really needs to be looked at using a time history approach rather than 'before' and 'after' snapshots.


Norm

RE: Roll center migration

(OP)
'In other words, this really needs to be looked at using a time history approach rather than 'before' and 'after' snapshots'
You are rifght but my calculation are static not dynamic.

What mention Cibachrome is tested to see the dynamic of car. First those test are for production car, all the suspension are silent bloc and they use MC Pherson full of friction, tire etc. For sure you will see time delay, hysteresis and so on. It is not the case on race car where suspension are rigid joint and so, only the tire are rubber. I understand the point of wiew of Cibachrome, roll center is irrelant when you test, because some other point are more much important.

I asked yesterday to Claude Rouelle: Do design a upper and lower wishbone suspension, with invariant distance of roll center to COG , and with minimum lateral displacement of roll center, on roll and suspension travel, is relevant?

He replied: Of course. What is important is not only the initial position of the roll centers but the variation of their vertical and lateral coordinates.

My question was: I find a way to design a twin arm with Rc not moving, which many author claim it is a good thing. honestly I have no clue if it is good not or not, what do you think?
I should add: if you have no experience in race car or you are thinking 'Mac pherson works well', please do not reply. If you reply, please use clear technical explanation (drawing calculation, measure) (not photos of old suspension, and not picture of car on 2 wheels). I have a mechanical engineer diploma, I m 47 year old, I m working for 20 years in motorsport, and I m still learning, but it is not a reason to waste me my time, by reply 'out of subject'

RE: Roll center migration

You may be asking too much from static-only analyses. I'm coming up on 8 years retired, and I remember that certain types of analysis in my career were best handled by time history methods. Even if the analyses themselves between what I did and what you're trying to do now are different, the approaches can be more similar across disciplines than you perhaps realize. I see a time history approach being of value here, unless you're making everything so rigid and the "roll moments" so low that the kinematic roll center hardly moves off its static location at all and all the forces are either zero or maximum.

At least keep the matter of time in mind. LLT through the roll centers (using that approach) builds more rapidly than LLT through the elastic suspension elements (roll takes a finite amount of time, assuming some sanity in choosing component stiffnesses). Even within the 'elastic' portion, damper forces (being velocity sensitive) don't peak at the same time as spring or bar forces (displacement sensitive). Meaning that over time even the amount of understeer isn't a constant.

You might also want to get further clarification on "the variation of their vertical and lateral coordinates" - is it only the coordinates themselves that matter, or is their variation over time also important? I'm a firm believer that the driver responds to subliminal-level effects, and while everything in the suspension is still moving everything is changing at least slightly.


I've already stated my thoughts regarding the MacStrut, and FWIW I've more or less been my own vehicle dynamics guy (off and on since the early 1970's; Structural was what I got paid to do). It's for sure been better than taking the word of everybody who'd sell me something that their solution was going to be my best solution.


Norm

RE: Roll center migration

(OP)
Buggar,

Nice job. but:

1. If you isolate a wheel with upright, sum of the force apply to W+U must be zero. This mean line force upper link, line force lower link, and line force from the tire contact must be concurent on a single point, otherwise your system is not static.
More clearly you have a degree of freedom (z).

2. How do you determine your jacking effect?.

RE: Roll center migration

(OP)
Brian,

I assume you mention this kind of graph



So if I well understand built up of geometric force (partially depend to Rc-CG) come fisrt with no movement of the suspension (no roll), then built up of elastic force (mainly dependent of spring, etc) with movement of roll.

As soon as no movement of the suspension is happens relative to geometric force, the movement of RC during roll (elastic force) is no relevant.

Do I well undertsand?

RE: Roll center migration

(OP)
"It is not the case on race car where suspension are rigid joint" While it would be possible to build a racing car that has a structure as stiff as a ball joint, it would not be competitive.

I meant rose joint not silent block like production car

RE: Roll center migration

The evidence for a dynamic influence on 'roll center migration' ought to consider the entire roll axis location trajectory because it does show up in frequency response tests and simulations of frequency response tests. First, here are some time traces from a simulation of a 'real' racecar on real (tested) tires and real vehicle parameters. A realistic step steer input based on driver steer velocity capability is used as input. The signal traces are normalized by their steady state values.

RE: Roll center migration

So, once you can get your arms around the physical mechanics with migrating roll centers (actually a migrating roll axis). You can play sandbox with the sim(s). The fundamentals are simple: Roll frequency is roll stiffness and roll inertia modulated by damping.. Stiffness is not the issue. Roll inertia is. Usually roll frequency is fixed: it's an inverted pendulum: No matter what the speed is.

But, yaw velocity and sideslip (hence lateral acceleration) IS speed dependent. It is possible (and GM did it) to configure a vehicle such that the roll frequency overlays the yawrate or sideslip frequency. When this happens, the 'Q' of the system can get very high and can self actuate unwanted responses. There are some other cool direct results from this analysis but you will have to guess what they are. No need to debate over the fence, just do the math.

I can see how a roll axis departure from its static position drives up the roll inertia via the change in distance to the sprung mass cg. I suggest you expend your energies recreating this analysis to make and prove your points.

RE: Roll center migration

Can't figure out what you want and can't waste the time

RE: Roll center migration

BlueFoxy. Minor point. In your diagram. Rear tyre delay period should begin when front lateral forces begin.

The "roll axis" clearly has some significance - eg when calculating the roll moment (as seen in cibachrome's last link).

je suis charlie

RE: Roll center migration

(OP)
cibachrome

Thank you for your reply.

But I must admit I did understand a few bit of what you explain.
1. English are not my native tongue, and most of time I have to use google traduction to understand, so I might lost the sense of your meaning.
2. I m not a sim specialist and I do not have access to, I m not able to the maths at this level. But I understand the importance of dynamics, clearly I have picture 1 and picture 2, but I do not know what happen between. I think your are far better than me to analyse what happens between.

My question is still: I find a way to design a twin arm with RC-CG not moving, which many author claim it is a good thing. honestly I have no clue if it is good not or not, what do you think?

If as a specialist of vehicule dynamics,if you think: it is no relevant, just tell me, and explain me as you will explain to an idiot. Example: it has no influence, because when you do dynamics, stiffness and damping are more important.

RE: Roll center migration

I have been designing quite a few single seater suspensions; It is difficult to reduce the vertical migration with ride height movements and in any case I never saw this bringing or take big advantages.
Instead it is crucial to have a laterally stable RC: drivers will report an unpredicatble car in case of migrating rc. The effect is less important at the rear.
Obviously if the car has a very limited roll movement the problem is damped down.

RE: Roll center migration

(OP)
Hello UBYVISI,

Thank for your experience feedback.

What kind of racing car did you experience this?

Regarding rear caster, like for a formula ford style suspension, it is a way to set the bumpsteer. Increase caster go for toe in in bump, decrease toe out in bump.

Regards

N.Maurel

RE: Roll center migration

Hello Bluefoxy,

cars I designed were F Ford, F Renault 2.0 and F Renault V6.
At that stage our capability was purely cinematic so no frequency was analyzed but what I stated remains valid.
I can confirm it since later I have been working and still I work with F Renault 3.5 and various F3 cars all of them confirmed those concepts.

Thank you for your point about rear caster, this is something I know and I fully agree; but I think that there is something more and somewhere in this forum I found a post about, I think our frien Cibachrome was the author.

Ciao

UV

RE: Roll center migration

I suspect that lateral RC migration is really an indication of something as opposed to an end in and of itself. If the RC isn't moving very far, that means the inclinations of the force lines from the contact patches to the FVICs can't be varying much. Which sounds to me that the anti-roll and jacking force effects would also be stable.


Norm

RE: Roll center migration

(OP)
Ubyvisi,

I know well Formula Ford, I worked from 2001 to 2006 at Mygale design office.

Regarding rear caster, from side wiew if you isolate wheel + upright, you put vertical and longitudinal load on the wheel contact.For push rod, Upper uright point will carry longitudinal load, and lower upright point carry vertical and longitudinal load. If you alter the caster you modify the repartition of the longitudinal effort between upper and lower point of the upright.

You can end up to a point of having zero effort on the upper upright point for braking or accelerating but not both.

But as far as far the longitudinal effort is small on the upper point regarding the inner point, this should not be a big thing.

If you have only vertical load on the wheel, and you add caster, you will get longitudinal effort on upper and lower upright point.

If you want, I can put some drawing to explain.

Regards

RE: Roll center migration

(OP)
Hello Norm,

'I suspect that lateral RC migration is really an indication of something as opposed to an end in and of itself. If the RC isn't moving very far, that means the inclinations of the force lines from the contact patches to the FVICs can't be varying much. Which sounds to me that the anti-roll and jacking force effects would also be stable.'

It might be an explaination.

RE: Roll center migration

Ciao Bluefoxy,

then we were almost competitors.....
Your notations about efforts are well clear.
The doubt I have is due to the fact that the rear susp (it is the usual double whishbone with pushrod) has the toe link welded on the top whishbone.
Assumed what we already said about bumpsteer, car seems to react to rear caster adjustment; more caster (positive as in front, nose up) gives more OS and the other way around.
As you say maybe this forces are deforming this whishbone so as there is some bumpsteer that cannot be seen measuring on a bench.
My kinematic analysis gives no bumpsteer......

UV

RE: Roll center migration

You will need to use software which can analyze member with elastic elements (as in wheel bearings, wheel, wishbone attachment stiffness. Then it will all make sense, If something is flexing, something could break from fatigue at any tome. Usually a 'toe' link mechanism is on swivels or ball joints when it is used as a ride/roll steer setting device. Shouldn't it be linking the upright to the chassis at a point off the control arm/wishbone plane ?

RE: Roll center migration

(OP)
Dear Ubyvisi,

On this kind of suspension, where toe link is weld on the upper wishbone.



From my experience , we put a little bit caster at the rear, may be 1-2°, to get a bit of toe in in bump (+0.2mm for 1 inch bump).

With this small angle, the resultant vertical effort on the upper wishbone is small, so twist (from side wiew) of the upper wishbone might be small.

But if you end with 10° caster to achieve no bumpsteer, then you will get a vertical resultant effort wich could be important and twist more the upper wishbone under load, end up with undesirable toe variation under load.


RE: Roll center migration

Hello everyone,

reply to Cibachrome: for this thread and the otherone too: first of all thanks a lot. I deal with racecar since a lot of time but when I was a student I did aerodynamics so, even if I studied by myself, several subjects remained a bit "foggy". In fact, I thought that, as you say introducing a moment would also induce something inside the tyre. About your picture, I fully agree: a spearated toe link is the best way to control (eventually also in an active way) the toe movement. Sometimes, tough, I have to work with cars like the one in Bluefoxy pic, and I have no right to modify them by regulations.

reply to Bluefoxy: yes the layout is exactly like that (picture). Everything seems to match, when the caster is introduced the car oversteers because something get deformed (toe link?), in fact this effect lowers a lot when (I checked the set-ups) the trailing arm goes to smaller values. Thanks a lot to you as well.

UV

RE: Roll center migration

(OP)
Hello Norm,

'I suspect that lateral RC migration is really an indication of something as opposed to an end in and of itself. If the RC isn't moving very far, that means the inclinations of the force lines from the contact patches to the FVICs can't be varying much. Which sounds to me that the anti-roll and jacking force effects would also be stable.'

By thinking for 2 days and I have a look on the force based roll center concept, I think you are right.

If the roll center stay in place, between static and roll, the inclinations of the forces line stay the same, this mean during roll same feeling for the driver.

If the roll center move to the outside tire contact patch (let say 2mm upper the tire contact), I think it is the worse case, because under small jounce, the force line on this outside tire can get reverse direction. As it is the loaded tire, it become a very unstable system. This kind of situation can happen if lower and upper arms have quiet the same lenght.

If the roll center move to the inside tire contact patch, The force line of the outside tire will not be angle, the force line inclinaison will not be the same as static, but angle should not alter under small displacement.

So nothing new by designing a twin arm suspension with an invariant position RC/Chassis, but a tips to get a stable suspension if Rc is not to far from the ground.

Red Flag This Post

Please let us know here why this post is inappropriate. Reasons such as off-topic, duplicates, flames, illegal, vulgar, or students posting their homework.

Red Flag Submitted

Thank you for helping keep Eng-Tips Forums free from inappropriate posts.
The Eng-Tips staff will check this out and take appropriate action.

Reply To This Thread

Posting in the Eng-Tips forums is a member-only feature.

Click Here to join Eng-Tips and talk with other members! Already a Member? Login


Resources

Solutions Brief - Protecting and Rescuing On-Ground Personnel
Keeping our warfighters safe and delivering them a competitive advantage is a key goal of departments of defense around the world. It’s a goal shared by embedded computing manufacturers like Abaco: we never forget who we serve.This case study describes how a major international contractor integrated an Abaco single board computer at the heart of its CAS/CSAR solution. Download Now
Datasheet - Top Enhancements Creo 7.0
PTC's Creo 7.0 has breakthrough innovations in the areas of generative design, real-time simulation, multibody design, additive manufacturing, and more! With Creo 7.0, you will be able to design the most innovative products faster than ever before, keeping you on the cutting edge of product design and ahead of your competition. Download Now

Close Box

Join Eng-Tips® Today!

Join your peers on the Internet's largest technical engineering professional community.
It's easy to join and it's free.

Here's Why Members Love Eng-Tips Forums:

Register now while it's still free!

Already a member? Close this window and log in.

Join Us             Close