choosing a roll center height.

[englertracing]

Okay, building a tube chassis road racer.
300hp+ na pinto motor setback a foot behind he axle centerline.
solid rear axle
2000lb Target weight.
11" wide tires.
Will have quite a bit of downforce.

Im doing all the suspension design I have read milliken rcvd. Over and over and over again.
For the roll center height I know I want the rear higher than the front. I'm thinking that the difference in height (roll axis) should correspond with the front and rear cog heights. So the front x amount below cog and rear x amount below cog.
Example the front is 1" rch and 14" cog a and the rear is 7" rch 20" cog. It seems logical to me to place the roll axis along the "cog axis".
Question 1 am I off here matching cog axis and roll axis

Question 2 assuming question 1 is somewhat correct how do i pick the proper rch. I can go above ground and below ground on both ends.
Based on the 14" and 20" cog I made up.
I could choose 1" f rch and and 7" r rch
Or - 2" f rch and 4" r rch
Or a 4" f rch and 10" r rch

All input appreciated.

[RCJ]

Match the cog roll axis with the r/c axis is something that I would give a try.
The angles of the a-arms that you use to get your r/c are important.Get one of the programs that let you simulate dive and roll.If the r/c moves alot under dive and roll the handleing could be inconsistant.

[Jerminator96]

I'm in much the same boat you are in, only I'm building a mid engine single seater with independent suspension, but still from scratch and largely based on an understanding hewn from the work of Milliken and Olley.

I would recommend getting one of the free suspension design programs floating out on the web and trying a few (hundred) combinations to see what gives you the best outcome or the least RC movement. I do not recall a hard and fast rule in Milliken's book, but if you want to start at the beginning I will work from the back and maybe one of us will find something.

Just curious, what tire will you be using?

[englertracing]

Goodyear pavement stock car tires.

Yea I know I need a stable rch for sure I just wish I knew where to place it...

[Jerminator96]

Goodyear pavement stock car tires.

Yea I know I need a stable rch for sure I just wish I knew where to place it...

Milliken RCVD reference 111. Milliken mentions that the equations are infrequently used, but it might be a good place to start.

[RCJ]

I would pick 4'' front 10 ''rear.With the lower a arms level, using the angle and length of the uppers to get the camber curves,then work with spindle and pin heights to get the rc where you want it and be consistent in dive and roll.When designing a new car the main thing is to not totally screw up the r/c (4'' below the ground10'' to the left, moving to 10'' above ground 6'' to the right in dive and roll).After you get the car sorted out is when you can start to find if 2 '' is better than 4'' etc...

[MadBill]

Late to the party here (really late!), but FWIW, the concept of a F & R center of gravity, just like its big brother the 'centroid of area' is fallacious. There is only one for the sprung mass.

[governor]

do you plan to run conventional spring and shock combination or go with BBSS or coil bind setup, as these will have a much different need for roll center placement.

If you ever watch a Sprint Cup car on a road course you will be in awe at how fast they can get those 3000 lb plus cars around a road course and they all use some sort of tie down/bump stop setup.

May I also ask if you are using a drop spindle of some kind?

Gov

[Dan Timberlake]

C2 Corvette IRS has IC/Roll center 7.56" above ground.
http://www.web-cars.com/corvette/1963_S ... php?page=4

FWIW Some folks (including herb Adams) say as a result it has excessive jacking.

For a live rear axle the IC depends on the springing and linkages.
This looks to be about right.
http://photos.motoiq.com/MotoIQ/Tech/Th ... xle-X2.jpg

[Warpspeed]

Few things here to consider.

If roll centres set too low, there will be more body roll and stiffer bars will be required at both ends, that will effect wheel rates and traction on bumps when cornering, and especially on slippery surfaces.
Low roll centres and stiff bars (with the higher wheel rate during cornering this creates) will make it very fast in the dry but it will slide much too easily in the wet.
This is the one big reason production cars have moderately high roll centres.

Roll centres set far too high will cause jacking and produce minimal body roll which makes the antiroll bars much less sensitive, and basically makes the car untunable.

You need to figure a whole host of factors such as CG height, track width, maximum realistic cornering "G" and available suspension travel, and how grippy the road surface really is, to arrive at an APPROXIMATE roll centre height RANGE.
Not as critical as many people think, as there are a lot of tradeoffs.

Once the car is built with an adjustable roll centre height at one end (or maybe both ends) you can start to tune it.

First balance the car on a skid pad for a very slightly increasing understeering gradient with with increasing speed.
You can get this right with the roll centres located just about anywhere.
This is also the time to find optimum tire pressures and optimum camber.
This will give you stable steady state cornering and get both ends of the car working.

The next step is to start adjusting the roll axis inclination for best TRANSIENT response.

Load will transfer faster at the end of the car as the roll axis height is increased at that end.
A low roll centre will feed the lateral load transfer produced from body roll mainly through the springs and bar.
Body roll is not instantaneous, and neither will be lateral load transfer at that end, response will be slower.

A high roll centre will produce less body roll, because most of the lateral load transfer goes direct through the suspension links, it responds faster to sudden changes in direction.

What we want is a stable vehicle with fastest possible response.
We do this by building in some steady state understeer (thicker front antiroll bar).
And giving it just a bit of transient oversteer (higher rear roll centre).

If it is done right, it will be fast and stable around a chicane, where you rapidly turn first one way then the other.
All this assumes there are no bump or roll steer problems to complicate things.

Every time you move one of the roll centre heights, you ideally need to then rebalance the car on a skid pad before testing for a change in transient response.

Vehicles with extreme or unusual weight distribution can require unconventional roll centre inclinations, the front roll centre is not always lower than the rear, but it usually is.

The fastest way to get from one steady state condition to another is with just a bit of overshoot.
That is where the transient oversteer comes in.
It will not cause instability, unless there is far too much.
What it does is speed up the response.

[mbrooks]

I would suggest designing as much adjustability into the thing as possible cause sooner or later you are either going to make a compromise or throw your hands up in despair. Listening to a proficient designer can make your head swim and there are so many factors to consider, driving style being one. Here's a little note someone and myself passed back and forth a few days ago....

With 1 degree of roll the RC migrates to -1.063 Horizontal (towards inside) and remains at -0.669 vertical. I pay more attention to the moment arm (CG to RC which is approximately 19 inches front and 17 rear) and in particular to the resulting Roll Axis.

To be perfectly honest all the below discussion is interesting but the reality is the tuning of the dampers is what the driver feels and what matters as far as lap times are concerned. (As long as you have optimum camber angle once steady state is reached) The trick is controlling the rate of loading.

The next thing that is not discussed is the tires. Each carcass is different and requires different static camber settings and camber curves (and these both throw your roll center all around) as well as pressures and load timing. (once again dampers)

So limiting ones thoughts to roll center migration as it relates to chassis dynamics as discussed below is a small part of the exercise.

......just enough to get one started!

[Warpspeed]

Roll centre migration is a very controversial topic.

But one point about it that most people seem to agree on, the roll centres at each end should behave in a fairly similar fashion as the vehicle is driven hard around a race track.

For example, some suspension location systems fix the roll centre height a certain distance above ground.
Other suspension location systems have a roll centre height that rises and falls as the CG rises and falls.
Both will work fine, but mixing them on the same vehicle may not be such a good idea on a very bumpy or undulating track.

As pointed out in my previous post, any roll centre locations can produce a balanced vehicle in steady state, but it may have some fairly dramatic transient handling issues and feel unstable or have very poor response to rapid steering input.

A quick thought experiment will tell you that if you can get it up onto two wheels while cornering, the roll centre is now located at the outer two contact patches, regardless of where some "theory" or computer program tells you it is.

Like politics and religion, the protagonists of forced based roll centres and geometrical based roll centres will probably go on arguing till the end of the world.
Roll centres located very near ground level, can move vast horizontal distances with even the slightest body roll, and the vehicles all seem to handle just fine.

The whole roll centre migration issue is clouded in mystery and confusion, which is why the authors of many otherwise excellent suspension design books often choose to politely ignore the whole subject altogether.

Maybe it is just an abstract concept, like the centre of gravity of a donut.
It's the bit that isn't there that makes you fat.

[Olefud]

For example, some suspension location systems fix the roll centre height a certain distance above ground.
Other suspension location systems have a roll centre height that rises and falls as the CG rises and falls.
Both will work fine, but mixing them on the same vehicle may not be such a good idea on a very bumpy or undulating track.

With a solid rear, i.e. staybar/Panhard rod, you'll probable have a mixed setup. I don't disagree with what's been said, but keep in mind that you'll have at best a compromise reached as much by art as by technology. For instance, with the L/S arm front suspension, the A-arms determine not only roll center but also camber gain, king pin axis (scrub radius) and bump steer with antidive only simi independent. As has been suggested, try different designs on paper (screen) with the expectation that your choice will need real world testing and tuning.

[Warpspeed]

More than one way to skin a cat.
It's a bit clearer with a Watts linkage, you can do it both ways.

[MadBill]

It's worth noting that an SLA front suspension can be arranged to maintain a RC fixed Vs. either the ground or the CG, as required to match an axle or chassis-mounted Watts pivot.