Roll Center Height and stability

[metal249]

I am building a double triangulated 4 link for an atv, and I'm having a hard time understanding how roll center height affects the the stability of a vehicle. I don't want to make this thing more unstable than it has to be, given its already high CG and short track. My current design has a roll center 4" under the CG, from what I understand, this will cause the chassis to lean away from the turn. Will this chassis lean make it more inclined to tip away from the turn? Would getting the roll center higher than the CG add stability? Thank you.

 

[BrianPetersen]

This is a solid axle suspension? High roll center is generally trouble with independent suspension, but if it is a beam axle, they'll tolerate a higher roll center than independent suspension will. You really do not want to have the roll center higher than the CG, though. An excessively high roll center, even with a beam axle, will cause a nasty side-stepping action whenever you hit a one-wheel bump, particularly if you do so at speed.

Rock-crawler vehicles frequently have suspension geometries that would be considered abysmal for any other type of vehicle ... but because they usually travel rather slowly, the bad side effects might not make themselves felt ... unless you put the vehicle in conditions outside its element, e.g. on a motorway.

One of the realizations that you need to make, is that body roll is a catastrophically disastrous thing to have and must be eliminated at all costs. Don't do that. If the geometry is such that there is a bit of body roll ... either just let it happen, or counteract it with spring rates and anti-roll bars so that it does not become excessive. The bad side effects of trying to eliminate body roll with suspension geometry aren't worth it; the cure is worse than the disease.

 

[metal249]

Thanks for the reply, this is a solid drive axle rear(triangulated), with an independent front. I came up with a roll center height of 10.5" for the A-Arm front suspension, with the current design of the triangulated rear having a roll height of 14". I believe my CG is around 18" off the ground. The rule of thumb in the 4x4 world is 8" above crankshaft center-line is CG, obviously with a small-block V-8 in mind. My motor is a GL1100 flat/opposed four, and being a motorcycle engine, has the transmission built into it, along the bottom. My measurement of 18" is right where the bottom of the crank case meets the top of the transmission, just under the crank. The front suspension needs to be widened 6-8" total, so I could possibly redesign it to function properly with the rear suspension. So, I should try to lower the CG of the bike, and level the roll center of both suspensions with the CG, or as close as possible? From what I understand, the closer CG and Roll center, less roll will happen. I'm completely fabricating the suspension, so I want to do it right, just not sure yet what "right" is yet. I attached a couple of pics of the project. Thanks, any insight into this will help me a great deal.

 

[metal249]

Extra note that might be important, the rear will use coil-overs.

 

[BrianPetersen]

There is no need for the front and rear roll center heights to match each other, or even to be remotely close.

Millions and millions of road cars were built with double-wishbone or MacPherson front suspensions with a low roll center and 4-link rear suspension with the upper links triangulated, thus placing the rear roll center above the diff.

Remember, solid axles will tolerate a high roll center much better than any independent suspension design will.

Your front roll center height seems excessive. It will lead to excessive jacking when going around corners - on the front end only, since solid-axle rear suspensions have minimal jacking effect even with a high roll center.

You are still fixated on control of body roll while disregarding jacking and side-scrub of the contact patches when going over bumps.

Again, as posted earlier, on an ATV that generally operates at low speed, some of these effects might not be so important as they are on a road car or a race car.

Still ... if it were up to me, I would drop the height of the attachment points of the front A-arms to get the roll center at nominal ride height down to something more plausible, preferably near ground level. If it needs more front roll stiffness, select spring rates accordingly or use an antiroll bar. If it needs plush front suspension with a lot of travel that leads to a lot of variation in geometry, use longer A-arms (do your widening by lengthening the arms). If you cannot drop the height of the chassis-end pivot points lower due to ground clearance then raise the height at the hub ends. Longer arms will help a lot ...

If you cannot do any of the above and insist on the high front roll center, it is an ATV and not a race car and certainly not a road car. Stay off the motorways with it and be mindful of that jacking effect when cornering! Lots of rock crawlers and off-road vehicles have suspension geometry that you could never get away with in a road car.

 

[metal249]

I can definitely lower ride height and was planning on adding 3-4" a side to the arms. If I lower the attachment points on the chassis end, could I further lower roll center by changing the vertical distance between the uppers and lowers? What would be an acceptable range for front roll height? Any other general changes that would reduce/eliminate jacking? What would be the affects of using a swing-arm instead of a triangulated link in the rear? This would help remove body roll and would be simple to build, not quite sure the downsides of it other than no rotation of the axle. It drove decent on asphalt with the swing arm rear, but it had very high spring rates all around, and was very lightweight, which seems to be why it drove well.

 

[BrianPetersen]

If you change the vertical distance between the arms, make a drawing to find the roll center. The effects might not be simple.

Regarding swing-arms, I presume that you are talking about motorcycle rear suspension type swingarm except with the entire axle moving up and down with a wheel on each end rather than a single wheel in the middle as it is on a motorcycle. That type of suspension arrangement has essentially infinite roll stiffness because the two wheels are rigidly linked across the vehicle. It won't do so well in high-speed one-wheel bump - or in uneven chatter-bumps that are different left to right. As noted before ... In a low speed ATV used on soft-ish ground, you can get away with stuff that you could never get away with in a road car or racecar. Also ... Any lousy suspension geometry can be made to sort-of work if you don't let it move (high spring rates). Extremely high rear roll stiffness is a recipe for oversteer at higher speeds. Good if you are flinging the vehicle around a dirt track. Bad if you are trying to avoid an obstacle on a rain-soaked motorway without ending up sideways, backwards, upside down, or into the guard rail.

Tractor-trailers with air suspension use essentially the swing-arm geometry that you propose, with the only difference in left-right movement being that allowed by compliance of the bushings and deflection of the components. Very high roll stiffness, necessary to act against the trailer tipping over. But that's why tractor-trailers don't have good ride quality even though they use air suspension with very soft vertical spring rates. One-wheel bump causes a "head toss" side-to-side movement.

It will really help if you tell us what the objectives are with this vehicle.

 

[metal249]

The goal is a long distance, freeway capable(75mph/120km) atv. A soft ride is not required, and the machine will see very careful off road use occasionally. It will need to pull a small trailer(600lbs total), safely. I've approximated the weight at 916lbs, full fuel, 175lb rider.

From what I understand, the ideal solid axle suspension for street use is a parallel 4 link or a 3 link with a pan-hard bar or watts linkage. I have no place on this machine to add a pan-hard of any significant length, and a watts linkage only gives a few inches of travel, so i'm told. I could run a triangulated 4 link or a swing-arm easily.

 

[BrianPetersen]

OK, so you need much more careful attention paid to stability and steering characteristics than a "normal" ATV would.

I agree that a linkage-articulated axle of the types mentioned would be generally most suitable at the rear. Rule out the traditional motorcycle-swingarm-type rear suspension and don't give it a second further thought. It has excessive roll stiffness.

Study the suspension designs used on normal road cars - pick one that works well, and get an understanding of why it works. Don't forget to consider the effects of scrub radius, steering axis inclination in front view, caster in side view, anti-dive introduced by having the chassis-side pivots on the front control arms non-parallel in either side view or top view or both (some cars have the upper control arm pivot non-parallel with the vehicle centerline in order to induce more anti-dive as the suspension compresses without having so much effect on varying the caster in the normal range of suspension movement!).

This is not a simple task by any means.

 

[wiloberlies]

Tire characteristics and yaw damping....

 

[BrianPetersen]

... bump steer / roll steer / compliance steer on both front and rear ...

 

[metal249]

Let me start simple, could I use a sway bar to control roll as in the picture? They are normally 4130 steel as far as I know. Given the light weight of the application, I could possibly use mild DOM tubing?

 

[metal249]

I plotted a new front suspension layout based on Mazda Miata front spindles, and 195/50r15 tires. When I lower the ride height to 7", and lengthen the arms 4", I end up with a roll center of 2.75" and a CG of 8".

 

[GregLocock]

" My current design has a roll center 4" under the CG, from what I understand, this will cause the chassis to lean away from the turn. "

no.



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[BrianPetersen]

You could use an antiroll bar in the rear as shown ... but it must be made of spring steel and properly heat treated. Mild steel tubing will not have anywhere close to the required fatigue strength.

And ... You are showing it in the rear suspension. The rear suspension is going to be the one with the high roll center. You are going to need the antiroll bar up front, and most likely not in the rear.

 

[metal249]

"And ... You are showing it in the rear suspension. The rear suspension is going to be the one with the high roll center. You are going to need the antiroll bar up front, and most likely not in the rear."

This solves a lot of the confusion I was having, it makes sense now. Since I'm running a solid rear, which is more tolerant of a high roll center, it can be set up with a roll center close to the CG, and wouldn't need an anti-roll device.

The double wishbone front will need a low roll center, to prevent jacking, and a sway bar to fight body roll.

That being said, and hopefully correct, this is my plan.
First and foremost, the whole bike will be lowered by 50%, at the very minimum. After playing with a suspension calculator, I realize the CG has to be lower to get anything near my goals.

My rear will be the double triangulated 4 link, with coilovers. The suspension links are DOM tubing, with threaded rod ends, adjustable hiem joints on one end, and rubber bushings on the other. I'll setup the suspension to spec with my calculator model, and the threaded hiem joints will allow me too adjust the lengths of the links to some degree. I will use adjustable mounting brackets for the chassis end, which will let me change the vertical position of the links. With the calculator as a benchmark for the initial build, and lots of test driving and adjusting, I believe this will work.

I have been trying to find anything comparable, and available in the States, in which to base the front geometry. Mazda Miata's seem to be the closest, next to anything exotic. Locost car builders use Miata's quite a bit, and a lot of info is available on its geometry. I need a smaller track width than a miata(55.5"), more around 42". Since the Locost guys use miata's, I can find a few A-Arm's, for very good prices, designed to fit miata spindles, but with "book" length wishbones/A arms. "Book" is whats specified in Ron Champion's book, "Build your own sports car for £250", UCA length 11.25" and LCA length is 16.25". If I shorten the lengths of the arms, or narrow their attachment on the chassis, it would change the geometry. Would there be any point in basing it off existing geometry if I just have to narrow it and completely change it?

An alternative to this would be lengthening the current lower control arm to the desired width, and using a suspension calculator to figure out the position of the upper. The upper could also have adjustable chassis attachment points, and adjustable lengths that allow for caster and camber adjustment.

I'll source an anti-roll bar from a UTV, probably a Polaris RZR 800. It's a little heavier than my projected weight, and the anti-roll bar will probably too stiff, but there just isn't anything smaller with an anti-roll bar. I've noticed that cars with anti-roll front and rear, the rear is generally softer. Perhaps the rear sway bar of the slightly heavier RZR would be softer and more suitable for the front of my ATV?

 

[Tmoose]

If I recall correctly in years gone by some threads have included comments by folks far more knowleadgeable than me that the perfect road suspension would need no anti roll bars, but that as a practical matter ARBs are a useful way to balance front/rear roll stiffness and thus understeer/oversteer.

If a car has a moderately stiff chassis I'm not sure that the end with excessive roll center height requires or even deserves the first ARB.
Early model Corvairs with swing axles at the rear have an IC 13.6 inches off the ground and demonstrate significant jacking and oversteer when pushed, especially on trailing throttle. The front is where the ARB goes first.

 

[metal249]

I'm going to start widening my lower control arms this weekend, I've lowered the bike by 5". I'm not able to lower it any further, as it sits, there is some kind of bind. I think the ball joint is at its angle limit. When I widen the LCAs, i'll straighten the ball joint/LCA mount and it should fix the problem. As it sits now, I can move the suspension in its full range and detect very little, if any bump steer. After I get the LCA's to my desired track width(42"), I'll build a simple, fully adjustable set of upper control arms, based on DOM tubing, radius rods, and a rod end ball joint. This should allow plenty of caster/camber adjustment.