44 replies to this topic

[Greg Locock]

FatBoy raised an excellent question in the FSAE thread - what is the proper way to set roll centre heights?

I should add that for many years I was completely cynical about RCH, largely because I had read a thesis in which some enterprising student had calculated both the geometric roll centre (GRC) and force based roll centre (FBRC) for a typical suspension as it went through a corner.

The two which are ostensibly the same thing, started off with different values and moved in different ways throughout the manouevre. They also responded in different ways to changes in hardpoint geometry.

Another reason for cynicism is that for a few years we were running an excellent beam axle (RCH 307 mm) and a fairly good IRS (RCH 120 mm) on the same model of car. On a smooth road there was very little difference in the handling.

Since the main program I use regards RCH as an output not an input then I don't actually need to know what it is, since I tune the car by looking at responses, not arbitrary drawing-board constructions.

However three years back I went through the (practical) exercise of altering the GRC on a car by +/- 20mm.

There were noticeable changes in the on centre feel, which is what we were after, so I now accept that GRC is telling me SOMETHING useful.

Of course when you set the roll budget up for the car you need to factor RCH in, but then that sounds dangerously like FatBoy's rule of thumb accusation - who says the split between GRC and sta bar contribution should be whatever it is? In practice we set RCH at a height dictated by experience (ie rule of thumb), set a roll gain from experience(rule of thumb), and make up the difference with sta bar. For a non aero race car that is not really an appropriate approach - your first requirement should be to use al the wheel travel up on a given track, not worry about arbitrary numbers like a roll gain.

I'd add that to confuse that side of things, I have calculated the speed with which roll forces move through the suspension and sta bar, and I would characterise GRC as fast roll forces, and sta bar as very slow. That may be a different perspective more in tune with Fatboy's approach.

One thing I think about occasionally is that GRC is effectively an 'anti', and antis are generally bad news.

[Joe Bosworth]

I have never convinced myself that any one GRC is better than others; however I have settled on +1 to 2 inches in front and a couple more inches higher in then rear as good starting points. I wouldn't fix an absolute value anf then sacrifice bump/roll camber change.

This was after building a few chassis with optional pickup points so that GRC could be easily alterred. If I were to build a solid rear axle car today I would still make way for altering the reart GRC.

Since what we really want is the most equitable front and rear roll couples front and rear and roll couple is a function of distance (CG-RC) Times spring rate resistance then GRC is every bit as critiical to tuning as are spring rate changes.

I have always worked on the basis of getting best mechanical grip balance on roll couples without roll bars or aero. Then add roll bars and once that is optimised tonly then add aero.

I don't think that I know enough to say that you can crank some numbers into a computer programme to get the best answer, a practice that everhyone seems to want in these days. No alternative to patient track work with a consistent driver!!! I used to do my own test driving so that I could feel the differences to balance against what the stop watch told us. Used to get some very good handling chassis.

I suspect that there is nothing here that you haven't already found,

Regards

[Goran Malmberg]

Intresting stuff and good points!

Fat Boy.
1 lack of actual engineering work and an over-reliance on rules of thumb.

2 You know the width of the racetrack. You know types of corners that you're liable to see. The compromises are that a wider car has less lateral load transfer, but a narrower car can take a better line.

3 Roll center heights are the same deal. They all have software that lets them play around with suspension points. None of them start from 'these are my tire characteristics and these are the corners we're going to see'. It's a lot tougher to do it that way, and the 'hot-rodder' guys that flock to FSAE just aren't prone to taking this approach.

Greg Locock.
4
There were noticeable changes in the on centre feel, which is what we were after, so I now accept that GRC is telling me SOMETHING useful.
5
Of course when you set the roll budget up for the car you need to factor RCH in, but then that sounds dangerously like FatBoy's rule of thumb accusation –
6
who says the split between GRC and sta bar contribution should be whatever it is? In practice we set RCH at a height dictated by experience (ie rule of thumb), set a roll gain from experience(rule of thumb), and make up the difference with sta bar.
7
I'd add that to confuse that side of things, I have calculated the speed with which roll forces move through the suspension and sta bar, and I would characterise GRC as fast roll forces, and sta bar as very slow. That may be a different perspective more in tune with Fatboy's approach.

One thing I think about occasionally is that GRC is effectively an 'anti', and antis are generally bad news.

Me.
1
Rules of thumb easily become copying rather than understanding.
2
Often people ask for "what is best?” a question that automatically does not include circumstances.
3
Software is a tool.
4
GRC do have its own effect, and should be used when needed.
5
Well, if we don’t need the GRC effect why should there be a RCH?
6
There is one rule of thumb here, we need a roll to pith basic balance (non aero cars) that dictates the need of added roll stiffness, namely the TW to Wb relationship.
Depending on what I said under point 4, this added roll stiffness should be a balance consideration between the geometric and sprung part of the roll stiffness.
7
Well, as the geometric forces are mechanically fast the sprung forces is dependent on spring rate as well as the shocks valving, both over the Mr.
As I think the nr 5 ”dangerously like FatBoy's rule of thumb accusation” is no real problem since making those balance decision is what engineering is all about. As seen here we also have a third member regarding the members of “weight transfer speed”
the shock absorbers.
Goran

[Goran Malmberg]

It is commonly accepted (often without being questioned) that the rear Rc height is to be higher than that of the front. Then the question arises, how high front and how much higher rear? In order to answer that question we must understand what the angle of the force line does to the car, and in this case we will be in a position to know how much angle we want. Only knowing that "rear Rc has to be higher than the front” is not really enough. So, the Rc starting point could as well be zero ground level.
What we need is a blend of timed and proportional weight transfer during the cornering. With today’s multi way adjustable chock absorbers we are having a bigger window of adjustments concerning this issue. Geometric weight transfer is on very fast operator here, we could say mechanically instant like a force line iron bar.
Cit Greg
"One thing I think about occasionally is that GRC is effectively an 'anti', and antis are generally bad news."
Ok, geometric weight transfer is heavy and harsh medicine that also stays during the whole cornering, and is therefore used in some % of the CGH. And as Greg says, it has downsides, like all strong medicines.

By starting from zero and then and look at how we want the weight transfer to take place, we might manage with a good blend of using the springs, sway bars and shock absorbers for weight transfer speed adjustments. Maybe with some touch of small geometric percentage, if needed.

Goran.

[Greg Locock]

I was flicking through a large company's database of measured vehicle characteristics, and at least one German company makes a reasonable handling car with a very high front roll centre and a somewhat lower rear roll centre. Most of their cars are more conventional than that. Perhaps it was a parts-sharing compromise, but the fact is they were able to work around it.

In fact the whole "rear RCH should be 50mm higher than front RCH" thing is EXACTLY the rule of thumb that I think needs examining. I can put together an argument that says the rear RCH shoud be higher than the front, but where does this 50mm number come from?

[McGuire]

Originally posted by Greg Locock

In fact the whole "rear RCH should be 50mm higher than front RCH" thing is EXACTLY the rule of thumb that I think needs examining. I can put together an argument that says the rear RCH shoud be higher than the front, but where does this 50mm number come from?

Where all such rules-of-thumb come from, I think. Within a range of more or less conventional packages it will work, so it's a useful starting point. But the rule of thumb carries no internal information as to how or why it works, and it certainly doesn't preclude other solutions. So as I see it, rules of thumb are most useful to those who would otherwise be stumbling in the dark. They are less useful as guideposts for those who know what they are doing.

[Goran Malmberg]

In a way I think its Ok with some sort of general rule of thumbs that gives a guidlene. 50mm is easy to understand and is also easy to pass on to become common knowledge. However, I think these sort of statments at the same time becomes a wall for knowledge to go throught.
50 mm statement may also create another problem, if one happen to land at 45 mm then, it might be believed to be undrivable. Such recomedation is not including any range of proportions.
Goran

[Goran Malmberg]

Originally posted by Greg Locock
I was flicking through a large company's database of measured vehicle characteristics, and at least one German company makes a reasonable handling car with a very high front roll centre and a somewhat lower rear roll centre. Most of their cars are more conventional than that. Perhaps it was a parts-sharing compromise, but the fact is they were able to work around it.

I know of another sportscar fabricator using almost CGH Rc, I should se if I can come up with who.
I am not supriced that they can make a car work like this. But depending on what we want the car to do, it will come out short in many cases, no matter the status of compromice. But of course, i some areas it will work fine.
Whas it a racing car, hmmm?
Goran

[Lukin]

I've never worked from the blank sheet of paper side, only the tuning angle as it's something we play with a fair bit on our cars (1500 kg rear wheel drive front engine sedan). Our cars really only have, within reason, a 50 mm window at the front and 100 mm at the rear, so we are never making big changes.

Where they get the 50mm difference front to rear I don't know. If you ran that on our car it would be an interesting result!

I'm with Greg and FB; it does load the tyre much quicker so it makes it more responsive, but as Greg alluded to, using the suspension links to transmit the WT means you can't dampen them so it would lead to theoretically more contact patch load variation and a loss of grip.

From blank paper I would imagine the RC heights are a function of the tyres and chassis characteristics; how much response and how quickly are the tyres required to be loaded to get them to work. I imagine an open wheeler with aero has a much lower F/R RC than a big sports sedan because they are inherently more responsive while the sports sedan has a lot of roll inertia and the phase between input (steering) and response (lateral grip generation) is bigger.

From the tuning side it's the circuit grip and the corner characteristics that determine the direction you go. Sometimes you want a higher roll centre for initial turning response (and put up with less grip after turn in) and sometimes you want less intial response to give more consistent grip through the corner. Other times your spring/damper/ARB requirements determine your RC needs; If the car 'wants' a stiffer rear spring you often have to lower the RRC to maintain the lateral weight transfer distribution. However if the track is gripped up (relatively speaking) you can afford a little more RC while the opposite applies; if the track grip is poor you need to maintain a constant load at the contact patch so you might want lower RC's.

To me there is few rules of thumb.

[Bill Sherwood]

I KISS - > Keep them both about the same height above the ground, make them both move about the same amount & direction vertically - it's less important laterally but common sense applies - and try to keep the car as close to 50:50 weight distribution as possible.
(Though a little more on the rear is better)

That gets thrown out the window a bit with something like a nose-heavy FWD though. The car I developed in Malaysia for the MME 12 hour weighed about 50kg more than the Hondas that were in the same class, but the front of each car was pretty close to the same so the rears of the Hondas were a lot lighter than our Corolla.
I worked hard(er) at getting more weight out of the front and into the rear to try to give the car an inherently better basic balance.
... and that's my point - You can generalise a fair bit about roll centres and a lot of other things, but the real basics has to be right for the rest to work well. Something with a 65:35 weight distribution is always going to be a bit of a compromise. A good RWD with something like 48:52 will be vastly easier to set up.

[Fat Boy]

I gotta reply to this one, but don't have time right now. Give me a little bit....like about enough time to write a book. I promise to get back to it.

[RDV]

Ditto, but a bit busy at moment...car pits in two laps!!
but just note=

Goran Malmberg-It is commonly accepted (often without being questioned) that the rear Rc height is to be higher than that of the front.

...in a very heavy, front centered six in line engined car, a high front RC was just the thing....will elaborate later...

[mariner]

When looking at suspension geometry it is always worth asking the question " what about a Kart?".

They have , theorectially, no suspension so no RC and so no RC.

So why do they corner so well and how are they tuned?

I don't really know the answers to that but since they clearly do corner well and have lots of chassis tuning tweaks it is a way of testing all the car based geometry theories. Simialarly the latest F1 cars appaer to throw al suspension geometry out eh window in order to get the front links to slope up out of the airflow.

It is clear on TV that ehre is some suspension movement front and rear but eh F/R roll couiple may well be weird.

[Goran Malmberg]

If building a new car we are not tied up by as much limitation as that of a rebuild. This gives a larger freedom of suiting the construction to circumstances. Rc (as anything on the car) I not a stand alone issue. As Lukin say, tires and grip are a big factor, as well as where the chassis has its CG(H). In my book I therefore start of with tires, wheel position and aero. After that the position of masses for as low as possible CGH and suitable polar inertia. We are now getting a brief picture of loads and weight transfer, and from there the suspension geometry design will begin along with positioning the Rc.

As Bill states,” Keep them both about the same height above the ground, make them both move about the same amount & direction vertically - it's less important laterally but common sense applies” is not a bad idea for a neutral build philosophy that might call for less harsh ”anti figures”.

If some more anti should be looking for, a slight adjustability could easily be built in to the A-arm chassis mounts.

Goran

[Bill Sherwood]

Originally posted by mariner
Simialarly the latest F1 cars appaer to throw al suspension geometry out eh window in order to get the front links to slope up out of the airflow.

I suspect they do that to get the roll centre above the centre of gravity, and so make the car roll into the corners, etc.

[NRoshier]

Originally posted by Goran Malmberg
If building a new car we are not tied up by as much limitation as that of a rebuild. This gives a larger freedom of suiting the construction to circumstances. Rc (as anything on the car) I not a stand alone issue. As Lukin say, tires and grip are a big factor, as well as where the chassis has its CG(H). In my book I therefore start of with tires, wheel position and aero. After that the position of masses for as low as possible CGH and suitable polar inertia. We are now getting a brief picture of loads and weight transfer, and from there the suspension geometry design will begin along with positioning the Rc.

As Bill states,” Keep them both about the same height above the ground, make them both move about the same amount & direction vertically - it's less important laterally but common sense applies” is not a bad idea for a neutral build philosophy that might call for less harsh ”anti figures”.

If some more anti should be looking for, a slight adjustability could easily be built in to the A-arm chassis mounts.

Goran

I have to disagree with some aspects of this WRT overall rules/purpose you are trying to design to. For local motorsport here aero would be a minor consideration WRT some of the other aspects - depending on type of motorsport. Sure an F3 car aero would be that high, but for most other classes aero would be after most of the others.

[Goran Malmberg]

F1 are very focused on downforce and CGH. The two factors makes for a very low relative weight transfer.
I think these factors overrule normal geometry thinking. They also have completely different tires.
Goran

[jeremy durward]

Originally posted by Bill Sherwood


I suspect they do that to get the roll centre above the centre of gravity, and so make the car roll into the corners, etc.

my understanding is that would create a weight jacking effect... i could be wrong, please tell me if i am but everything i've read and experienced tells me that is the case

[Bill Sherwood]

Originally posted by jeremy durward


my understanding is that would create a weight jacking effect... i could be wrong, please tell me if i am but everything i've read and experienced tells me that is the case

Yes it would, though they have such a low centre of gravity it'd be liveable.
(I suspect)
In any case, it can't be denied that with the suspension they're running they have a very high pair of roll centres.

[Goran Malmberg]

Anyone having an opinion about this Q:s.
Let us say that we do have a 1” front and 3” rear Rc, all in order for a good turn in balance.
1
Then we do have this anti effect during the rest of the cornering, would this be desirable? Or how would it affect the average cornering balance.
2
We where talking Rc height (intersection of the forcelines). Let us assume this intersection height remain constant but moves sideways. As the jacking is dependent on load*forceline angle, jacking will become different as the forceline angle alter during it’s sideway intersection movement. Keep in mind that the outer pair of wheels may be heavily loaded.
3
Wouldn’t it be a good idea to speak “cornering weight transfer speed” and then look at Anti, Shock settings, Springs and Swaybars as tool to get the right mix.
Goran

[cheapracer]

I think with F1 the aero supersedes all other considerations and not what Gregs looking for.

I'm with Bill, flat roll line (I dont care about the roll centers, its the roll line thats important) which under braking will have the rear higher anyway to get some weight/feel on the fronts on turn in.

Again with Bill, get the CG weight down and inbetween the weelz if possible then the RL is less important.

I like the idea of a live rear end with a beam front too, why I was dissapointed when I asked Bill about the type (forgot the name) he had that he never used that gets the roll center down, darn you!


Oh by the way for those of you who could do with a simple desktop RC proggy, get this demo..
http://www.performan...ends.com/rc.htm

Now the demo of course locks out your ability to input your own figures bar 2 into the frontend's 2 available demo files ("dbl-aarm'" double A arm and "Mustang" McPherson strut) but if you open these 2 files that are in the "RCfiles" folder with notepad you will notice that they are totally editable to your own figures and saveable. Slow but worthwhile.

[cheapracer]

Originally posted by Goran Malmberg
Anyone having an opinion about this Q:s.
Let us say that we do have a 1” front and 3” rear Rc, all in order for a good turn in balance.

Dramatic difference on the roll line for a 80" WB compared to a 120" WB isn't there!

Even getting your RC determining links as far away from each other as possible is worthwhile.

[Greg Locock]

The ugliest spreadsheet I ever wrote is also the one that attracts the most hits

http://www.geocities...cock/slarck.zip

That's a roll centre calculator

[imaginesix]

You mean the ship's hull resistance program isn't most popular? Go figure.

[cheapracer]

LMAO, I went straight to the "battleships"! Ahh some of us never grow up

[phantom II]

This post of bill's in response to Mariner needs closer examination. F1 cars have unbelievable roll couples. Very high rc in the front and very low at the back. The front camber goes thru a loss in braking and turns.
On the National Geographic channel, they had a series of shows called technology of the old west. They showed a stage coach cornering abilities and how they reduced frequent roll overs. The cab was suspended as low as possible in the frame allowing the CG to be below the roll center. The cab would lean into the corner like a motor cycle.
If you consider some of the comments on this thread, particularly the high front RC on old 6 cylinder cars one may conclude that there is more to the rc height than what is written here. The couple of the CG and RC seems quite important. I try to get the roll center and cg as low as I can. The effective roll couple is determined by the location of the CG. Usually a front engine car RC couple points uphill and a rear engine car points downhill. The smaller wheels on the front of the stage coach cause a lower RC. The CG is toward the rear wheels and the cab rotates clockwise below the RC on a downhill axis in a right hand turn.
On a car, the load sequence from transfer can be optimized in the suspension design if you know the relationship of these couples and how you may generate these forces during your packaging process.

Originally posted by Bill Sherwood


I suspect they do that to get the roll centre above the centre of gravity, and so make the car roll into the corners, etc.

Originally posted by mariner
Simialarly the latest F1 cars appaer to throw al suspension geometry out eh window in order to get the front links to slope up out of the airflow.

[cheapracer]

Yes Ph11, the relationship is important, for example BMW race cars when lowered the RC goes away from the CG so they like to get the RC back up so they don't need to use so heavy anti roll bar and weight the outside tyres. There is a question of camber gain as well in their packaging.

If a rear engined car (or specifically a rearward mass biased car) has a forward sloping roll line and a CG above the roll line then you have a situation where the CG wants to take an easy path downwards (or falling over the roll line if you like) under braking/turn in. Not ideal if your CG is high but maybe good for balance if the CG is just slightly above the roll line (depending on where your RC's migrate to of course).

By the way, those stage coach tests were done with or without a heap of baggage sitting on the roof?

[Goran Malmberg]

Originally posted by cheapracer
they like to get the RC back up so they don't need to use so heavy anti roll bar and weight the outside tyres.

Isnt weight to the outside tires a function of (cgh*g*w)/Tw only?
Goran

[Joe Bosworth]

I strongly suggest that F1 is no example of best use of RC. Rather they are willing to sacfrifice anything for underbody clean air flow adding to the aero effect.

They probabbly can sacrifice all kinds of mechanical grip for increased aero grip. Clearly their recent use of the keel nose was to get a decent front geometry but they were willing to forego the keels to gain performance elsewhere.

Regards

[mariner]

One practical point in all of this is the actual length of the linkages involved, the longer the transverse linkage the less angular displacement per cm/inch of suspension movement and so the less RC movement.

Single seater designers can go for very long wishbone lengths if they so desire. If they are nearly half the track width ( as per F1 front) then the RC movements are actually very small. Then you get good RC control and probably less jacking efect.

Conversely road car designers ( like Greg) may be restricted on linkage length and they have to provide for much more suspension rise and fall. This is likely to give a much harder challenge on RC movement.

This can be true of front suspensions as bottom links conflict with sumps etc. and top link length is tough with transverse engines ( hence the popularity of strut set ups). At the rear boot/trunk space again limits lengths. The five link virtual set up ( used first by Mercedes I think) helps as it allows he tranverse length the be a virtual one beyond the actual linakge length, as well as resolving bushing loads into purely fore and aft.

So maybe for raod cars packaging trumps all else?

[murpia]

This is a motorsport view, rather than road car...

I think that the front view geometric roll centre height will never completely die as a concept in professional motorsport while the current generation of senior designers and engineering managers are still active. Also there are a lot of textbooks and 'non mathematical' texts out there that appeal to the amateur and the interested student that put forward this view of roll centre.

For those of us that have moved on to multibody simulation or other force-based techniques it's clear that the lack of symmetry in a real car under real conditions does invalidate the geometric approach.

I have had the experience of testing a car with multiple geometry options. For each option we did indeed calculate a geometric roll centre height (to please management) but more importantly calculated the vertical force distributions per g at the contact patches (we call it roll couple distribution or mechanical balance). We matched these distributions for each geometry by adjusting the side spring rates. What we found was that the geometries made little difference once the roll couple distribution was equalised. We were adjusting the rear geometric roll centre height either side of a reference. Always above ground. We tried front = rear, front = 80% rear, front = 60% rear (approx), by adjusting the rear.

This is for a car with significant aero and very low compliance in the suspension. The car has quite a lot of antis of all kinds - these do not seem to be much of an issue with an aero car.

So to answer the original postings about how to choose a roll centre, my current opinion is not to choose roll centres but to design in a roll couple distribution that suits your tyres (and aero, and weight distribution). So you have more design freedom to achieve it, but you do need to know ahead of time what will suit your tyres and car. Allow yourself the adjustment range of say 10% either way.

Regards, Ian

[mariner]

re roll couple distribution Colin Chapman 1975 " all that mattaers is how wide it is, how long it is and where you put the weight"

Still true on aero cars in 2008 maybe ( accepting that he and today's engineers play with spring/damper and A/r rates) - he wasn't rejecting those in the quote.

[Ben]

Originally posted by murpia
This is a motorsport view, rather than road car...
I have had the experience of testing a car with multiple geometry options. For each option we did indeed calculate a geometric roll centre height (to please management) but more importantly calculated the vertical force distributions per g at the contact patches (we call it roll couple distribution or mechanical balance). We matched these distributions for each geometry by adjusting the side spring rates. What we found was that the geometries made little difference once the roll couple distribution was equalised. We were adjusting the rear geometric roll centre height either side of a reference. Always above ground. We tried front = rear, front = 80% rear, front = 60% rear (approx), by adjusting the rear.
Regards, Ian

Isn't that just the same as calculating lateral load transfer distribution as per Milliken with different roll centre heights? Achieving the same load distribution with different levels of anti-roll / roll centre height isn't the issue so much as what other effects those different geometries have other than on the "steady state" loads.

What I'm getting at is the rate of change of load transfer as opposed to the total amount.

Random question as a result: what are the relative effects and trade-offs between running a higher roll centre and softer springs vs. a pre-loaded ARB, stiffer springs and a lower RC if both have the same load transfer per g in the middle of the corner (assuming you've exceeded the preload)?

Ben

[phantom II]

Yeah, but the safe holding the notes and the gold were kept under the seat. If you want to see these things in action, check them out at the Calgary Stampede in Alberta. It is modeled after Buffalo Bill’s Wild west Show. You will see Stagecoaches and chuck wagons in 4 wheel drifts on a huge dirt oval with guns a blazing. You gotta see this show before you die or before the tree huggers shut it down.
Closer inspection of these vehicles shows modern technology. They are made of fiberglass and steel with modern wheel bearings. Even the straps that suspend the cab are other than leather.
Another thing, I always wondered why cars and race cars had positive camber in the front for so may years.
Even my soap box racer in the 50s has negative camber. OK, so the axles bent. It sure cornered better.
How come Ettore Buggati didnt learn this after all those years of building and racing his little cars? Shirley the wear on those punu little tires should have told him something?

Originally posted by cheapracer

By the way, those stage coach tests were done with or without a heap of baggage sitting on the roof?

[murpia]

[QUOTE]Originally posted by Ben
Isn't that just the same as calculating lateral load transfer distribution as per Milliken with different roll centre heights? Achieving the same load distribution with different levels of anti-roll / roll centre height isn't the issue so much as what other effects those different geometries have other than on the "steady state" loads.[/QUOTE]
Probably the results are broadly similar. However no assumptions are made about the existence or validity of a roll centre height, it's purely done on calculated forces at the contact patch. Also Milliken assume a fixed motion ratio (unlikely) and an unmoving roll centre height (also unlikely). Likewise the Claude Rouelle / Optimum G 'Magic Number' spreadsheet.

From memory, the definition is something simple like:

Mechanical balance = (left front contact patch vertical force - right front contact patch vertical force) /
(left front contact patch vertical force - right front contact patch vertical force + left rear contact patch vertical force - right rear contact patch vertical force)

The calculation is static, no damper or inertia forces involved. I'm sure it could be duplicated in ADAMS etc. very easily. One day I may duplicate it using one of the open source games physics engines.

With an aero car the ride height change with speed means we would quote the number for 100kph ride heights.

What I'm getting at is the rate of change of load transfer as opposed to the total amount.
[/QUOTE]
What we found was, neither the driver nor the stopwatch were sensitive to this effect, just to the total load transfer. But of course that was our car not yours so once again 'rules of thumb' aren't so useful...

Maybe the tyre relaxation lengths coupled with surface temperature build up effects dominated the rate of change of load transfer effects? Maybe the way the driver traded off steering input and brake pressure to transfer vertical load to the outside front tyre to get the turn-in response he needs was dominant?

The test was used as justification to package rear suspension for the best structural and aero performance (within reason). A similar test on rear camber gain concluded something similar - as long as the dynamic camber of the outside rear tyre was matched, the driver could not feel the effect of more or less camber gain.

So, if an FSAE team came into my judging bay with a similar analysis plus some test data to back it up, I wouldn't dream of asking them what their geometric roll centre heights were...

Regards, Ian

[Goran Malmberg]

Originally posted by Ben


Achieving the same load distribution with different levels of anti-roll / roll centre height isn't the issue so much as what other effects those different geometries have other than on the "steady state" loads.

What I'm getting at is the rate of change of load transfer as opposed to the total amount.

Random question as a result: what are the relative effects and trade-offs between running a higher roll centre and softer springs vs. a pre-loaded ARB, stiffer springs and a lower RC if both have the same load transfer per g in the middle of the corner (assuming you've exceeded the preload)?

Ben

We are having a total (average) load transfer dependent on CGH and Tw. Within this frame, the suspension geometry, wheel spring rate and shock dampening pattern will make this load transfer to take place in different shapes. And, is beeing used to give different front to rear axle Wt distribution. Concernng swaybars, they just add to normal spring wheelrate in roll. (Althrought they under some circumstances will act different). For downforce cars we may need the oposit to swaybar, and use more vertical spring rate.
So, there must be some "circumstances" while driving on the track that calls for a special weight transfer pattern.

Goran

[cheapracer]

Originally posted by Goran Malmberg


Concernng swaybars, they just add to normal spring wheelrate in roll. (Althrought they under some circumstances will act different). Goran

Well they add (outside wheel) and subtract (inside wheel) at the same time, they aren't a fixed spring.

And I think of them differently - if you have no suspension (or unison suspension such as an ATV) then the RC is at the contact patch of the outside tyre so it figures to me that swaybar's effect is to migrate the true(?) RC towards the outside tyre depending on strength.

[cheapracer]

Originally posted by Goran Malmberg

Isnt weight to the outside tires a function of (cgh*g*w)/Tw only?
Goran

Tuff question with all the variables and my lack of math (well just lazy and no time actualy).

The distance between the RC and CG is a bearing factor on this, low RC and hi CG will put a lot of weight on the outside wheels and vise versa. Leverage action.

Ph11, positive camber with lots of castor for the old rigs will get the tyres flat while turning. I have a feeling they wanted the shoulder to 'dig in' on initial turn in.

[Goran Malmberg]

Originally posted by cheapracer

Well they add (outside wheel) and subtract (inside wheel) at the same time, they aren't a fixed spring.

As for what I made an reservation (dynamic situations) I look at it this way. If you have the car on a parking lot and use some lever arm to roll the chassis, (this is just to get an clean roll motion to describe what i mean), one side of the swaybar will go up and the other down. This means that the middle of the bar wil remain in rest and the bar could simpley be looked at as two half bars, one each side where its inner location could be welded solid to the chassis. In other word two torsion bars. The difference between the two torsion bars and the regular springs will then become that the torsion bars is unloded at rest while the regular springs is loaded by the weight of the car.

I mention this just as another way of looking at its function rather than the "inner wheel lifting theory" that is common description, and which I think easely counfuse understanding.
If I am not making in even worse .
Regards
Goran

[murpia]

Originally posted by Goran Malmberg
Isn't weight to the outside tires a function of (cgh*g*w)/Tw only?
Goran

Yes. A free body diagram of the whole car easily confirms this.

But...

cgh may not be constant once a lateral force is applied.
Lateral cg movement will have an effect.
Tw may change with suspension movement.

Note that Tw should be at the cg position, if F & R tracks are unequal.

Regards, Ian

[cheapracer]

Originally posted by Goran Malmberg

. This means that the middle of the bar wil remain in rest and the bar could simpley be looked at as two half bars, one each side where its inner location could be welded solid to the chassis. .
Regards
Goran

But it's not and they aren't.

[Goran Malmberg]

Originally posted by cheapracer


But it's not and they aren't.

If the car makes a clean roll, whats sort of difference would it make if the bar is welded solid to the chassis or not?
Goran

[Goran Malmberg]

Sorry, I forgot to write, welded solid in the middle.
Goran

[cheapracer]

Originally posted by Goran Malmberg

If the car makes a clean roll, whats sort of difference would it make if the bar is welded solid to the chassis or not?
Goran

well then you would clearly have 2 independent torsion bar springs working as a normal spring independent of each other. When you hit a bump on one side you can see that a swar bar doesn't just twist from the middle can't you? Same as around the RC, if the RC isn't perfectly in the middle sway bar then the swar bar won't just twist from the center. put 1 end in a vice and twist it and see what happens.

[Goran Malmberg]

Originally posted by cheapracer


well then you would clearly have 2 independent torsion bar springs working as a normal spring independent of each other. When you hit a bump on one side you can see that a swar bar doesn't just twist from the middle can't you? Same as around the RC, if the RC isn't perfectly in the middle sway bar then the swar bar won't just twist from the center. put 1 end in a vice and twist it and see what happens.

That is what I reffer to as the exception from a clean roll situation, right? I guess the main perpouse of the sway bar is not transferring bump movements from one side to the other wheel in straight ahead driving (or in roll) , but rather to act as an extra springrate in roll. There are a number of this other than plain "rollspring" effects from swaybars, do we really asked for them? And how do we predict them? I am trying to analyse the clean ROLL situation and show that it under such circumstances the swaybar is nothing but added spring stiffness in roll.
Regards
Goran