Roll Center vs. Jacking

[BoostCreeps]

Looking at some previous designs, some people have put their roll center well under ground. If my understanding of roll center placement is accurate, This would cause the car to understeer considerably right? Raising the roll center would reduce understeer, untill it was raised too high, then it would cause a jacking force correct? Well usually how do you know how high you can go with the roll center?

Could you run an upper control arm parallel to the ground, and a lower one at an angle going up towards the center of the chassis to raise the roll center? Or is this suspension geometry a bad idea?

This advice is for a non-fsae-competing car, but the physics still applies.

[BoostCreeps]

Looking at some previous designs, some people have put their roll center well under ground. If my understanding of roll center placement is accurate, This would cause the car to understeer considerably right? Raising the roll center would reduce understeer, untill it was raised too high, then it would cause a jacking force correct? Well usually how do you know how high you can go with the roll center?

Could you run an upper control arm parallel to the ground, and a lower one at an angle going up towards the center of the chassis to raise the roll center? Or is this suspension geometry a bad idea?

This advice is for a non-fsae-competing car, but the physics still applies.

[Matt N]

I know some real suspension experts will get in on this post, but as far as your geometry question goes ...

Look at the Jaguar F1 car. Horizontal upper control arm, upward sloped lower control arm. Exactly like you said.

Matt

[Schulberg J]

Be very careful looking at modern F1 cars for suspnesion geometry inspiration. Their priorities differ quite significantly from ours (think aero, aero, aero...)

[Tudor Miron]

Hi Clay,
First of all I'd like to warn you to not overestimate importance of roll center. Either of them geometric or datum center (force based roll center) is MUCH less important than your instant centers. Jacking and other "geometric" forces act through your instant centers.
I don't really see why car with ower than ground roll center (again this is not what you have to really look after as by itself roll center doesn't effect reaction forces at the contact patch.) will cause understeer. Unless you have exessive roll and too much camber loss. In fact lower than ground roll center (i.e. low instant centers) will in most cases and if done right give you a very stable front end. Going over the ground will cause jacking forces at laden wheel (actual amount will depend on your I.C. dynamic loacation) and if significant will play havoc with front grip. Geometric forces are usualy "shock" loadings. Also they are not normal to the ground plain - which does nothing good for your tires.

My two cents thought

Thanks
Ted

[Tudor Miron]

Schulberg is very right when talking about different priorities - F1 guys will trade their mother for stable aero platform and they have enough resources to simulate and test to deal with weired tire loadings. Also NASCAR guys use geometric forces extensively - but their cars are VERY heavy and without jacking and anti's they would have to use gigantic springs and bars.

Thanks
Ted

[BoostCreeps]

So what would be the best way to go about determining my control arm angles best for my setup? Ive read Carroll Smith's books about supension geometry and how differences in these angles can affect handeling characteristics, but im having trouble figuring out a good setup for me.

[Mark Bacchetti]

Make sure you think of roll centers in terms of the front or rear suspensions. The interaction between the front and rear roll centers plays a crucial roll in how your car handles.

I highly suggest reading Milliken's Race Car Vehicle Dynamics. In there you will find some equations for Total Lateral Load Weight Transfer Distribution. These equations will give you a lot of insight on how changing suspension parameters will affect the balance of the your vehicle.

-Mark
Cal Poly Pomona

[Z]

Clay,

This is a big subject! Not something that can be explained in a few sentences. But here's a start...

First, I agree with the above posts regarding F1, etc. Don't even think about them! (If you want to copy Jaguar you'd best first check how many wins they got... Even Williams had a terrible car, suspensionwise, this weekend. All its bouncing (on the smooth straights!) cost Webber a podium. Most F1 people think aero 24/7.)

I also agree that you have to think in terms of the whole car and all the relevant factors. You might want to read the "Zero droop behaviour" thread to get some idea of how many different factors can influence handling balance. Also, the "roll centre" is NOT the best way of thinking about independent suspensions (see my comments in zero droop thread).

Nevertheless, regarding "roll centre" height:

Generally (although not always!), an above ground RC will jack that end of the car up in corners, while a below ground RC will pull that end of the car down. Either of these (up or down) can be good or bad depending on a million other things. But usually TOO MUCH upwards jacking is a VERY BAD thing. The front and rear of the car moving in opposite directions (up/down) is also, usually, very BAD. A large change in jacking forces with changing chassis ride-height, pitch, and roll, is again very BAD.

So, in one sentence, start with your RC's close to ground level, and try to keep them there for all positions of chassis bounce, pitch, and roll.

Using long lower wishbones is a good place to start.

Z

[Brian Smith]

Lets talk about jacking forces.

If your Instant center is above ground then a lateral force will result in the chassis being raised up. If te IC is below ground then the chassis will be pushed down. But that dosent affect tire loading does it? Well it will raise or lower the cg during a turn and that effects loading, but is this the only way it changes loading. I know this raising of the cg is a really bad thing because its happening in a turn where you need a low Cg, but is this the only problem with jacking.

Also say you have a below ground IC front and an above ground rear. This will lower the front and raise the rear of the car. This changes the roll axis slope and screws with my precisely calculated elastic/geometric transfer percentages. So this says that the IC's should both be above or below ground. If they are then they need to be relatively the same heigth (no 6" differences http://fsae.com/groupee_common/emoticons/icon_wink.gif ).

So, if my IC's are both above ground, relatively the same heigth (more precisely the lines connecting the tire contact patch to the IC are relatvely the same slope, n-line slope) then my only worrie is that jacking will raise the cg, right?

[DJHache]

Also keep in mind that excessive jacking will scrub your tires laterally and affect your track width (thereby affecting wheel loads). Depending on your tires you may want some lateral scrubbing to put heat into your tires. Also if your roll center is high then you will have less adjustability in your suspension because more of the weight transfer will go through your links.

[Brian Smith]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> Also if your roll center is high then you will have less adjustability in your suspension because more of the weight transfer will go through your links. </div></BLOCKQUOTE>

You are saying that with a high IC, arb adjustments will have less affect on handling because most of the roll resistance is from the suspension links. right?

[DJHache]

Yeah, pretty much Brian. In the extreme, the RC is matched with the CG and the car doesn't roll (except for tires). At that point your ARB's are just ballast.

[Buckingham]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content"> Also if your roll center is high then you will have less adjustability in your suspension because more of the weight transfer will go through your links. </div></BLOCKQUOTE>

Depends what you want to adjust. If the springs have less effect on US/OS, then you have more room to play with in terms of wheel rate. You may be surprised when you actually run the numbers and find what your US/OS sensitivy to spring rate change actually is.

[Tudor Miron]

Brian... what makes you think that jacking forces does not effect tire loadings (apart of CGH change as you said)????????
Think - if some force acting through the supension links is raising or lowering the chassis than there is a reaction force at the contact patch. Magnitude of that geometric force depends on your IC DYNAMIC height and length. Worst think about this force is that it is NOT normal to the road and has "impact" nature. Think about an object laying on sheet of paper. If you pull that paper with Force X but relatively slowly than object retains it's position (relative to paper) but if you pull with same force X but fast enough than object will slide on paper. Jacking forces can change in magnitude and sign while cornering - this does nothing for stable predictable handling.
I don't say that jacking is absolute evil and as I mentioned above it is used widely (along with other anti's) in many series to help support the chassis. I may think that it could be used to benefit on narrow and twisty and short run to help heating front tires for example. How ever with light and not heavily aero laden chassis I would prefere to have those forces low and as constant as possible. This makes for much easier to setup and nice to it's tires car. Again - RC (which ever of them you refer to) is not the cause it's the result. Watching it's dynamic movement can give general indication of how noisy your IC's are but I would better pay more attention to CAUSES.

Thanks
Ted

[Z]

Re: "Let's talk about jacking."

Here's one way to look at it.

Imagine that you are skating (roller or ice-skates). You have got some speed up and are now coasting along with your legs straight and wide apart. Your straight legs represent the "n-lines" of the "front-view-swing-arms", your inner-thigh muscles are the pull-rod suspension, and the "roll-centre" is somewhere around your, err, wedding tackle.

You turn one skate toe-in. What happens? Yep, that skate pushes inwards, putting a compressive load on that leg, which jacks your body upwards. You straighten that skate then turn it toe-out. This time the compressive load on your leg reduces and your body is jacked down.

So, LESSON 1.

A sloping n-line, combined with a lateral force at the wheelprint, causes an UP OR DOWN jacking force even when the car is travelling STRAIGHT AHEAD.

Does this really happen? Sure. With sloping n-lines any static toe angle will jack the car up or down, although usually only by a small amount. Likewise, toe change due to bump steer, or compliance steer due to accelerating or braking forces, can cause jacking. If one wheel gets caught in a rut and "tramlines" then it again causes jacking, this time possibly by quite a large amount.

Mostly we don't notice this jacking. On a straight smooth road it is almost impossible to notice any change in ride height due to steady jacking forces (except by data logging). On a rutted road we blame the bumps. However, it is easy to notice body roll angle during smooth cornering (compare windshield frame against horizon). This is why people talk of the "roll-centre" rather than the "jacking-centre" or "jacking-&-roll-centre", even though jacking is more common than roll.

And this is the crunch, because ALL OF THE STRAIGHT LINE JACKING EFFECTS ALSO OCCUR DURING CORNERING.

Furthermore, if a car has steep n-lines (high RC), and it has parallel steer of its front wheels, then on sharp corners its front wheels will be effectively toed-in by about 10 degrees, which can give large upwards jacking. If the car has large dynamic-toe-out (=pro-Ackermann) and the same n-lines, then it might jack down (depends on lateral load transfer). If the inner-wheel drops in a rut (say, the gutter) and is pulled towards the corner centre, then again there can be large downwards jacking. If the outer-wheel loses grip, because it is on oil, sand, or water, then the inner-wheel can jack the car down. Or vice versa to all of the above, depending on conditions...

Corollary to Lesson 1.

Any jacking of the car's body, up or down, must be accompanied by a change in vertical wheelprint force (Fz) during the jacking (from F=ma). In fact, the changing Fz force causes the jacking. The ratio of the vertical Fz force to the lateral Fy (road-to-tyre) force is equal to the slope of the n-line in radians (ie. Fz/Fy = rise/run).

So, not only is the body and its CG moving up and down, but the Fz forces are varying. The varying Fz forces inevitably have an influence on cornering handling balance. These influences on handling balance are largely due to the road surface conditions. The suspension designer would have to be very lucky for these influences to be of a beneficial nature, all of the time.

So, the easiest way to minimise these unpredictable influences is to keep the n-line slopes close to horizontal (ie. near ground level RC).


Z

PS. Longitudinal n-line slopes combined with longitudinal forces behave pretty much as above. These are usually called anti-dive, anti-squat, etc.

PPS. Jacking effects with beam-axles are much less trouble than above, but that's another story...

[Brian Smith]

allright z,

Assuming aboveground IC's and forgetting about changes in toe and having symmetrical IC's...

If we are in a corner and have weight transfer, then the inside tire will have less normal force than the outside. This means the inside tire will have less lateral force to create jacking. The inside creates downward jacking and the outside creates upward jacking. This results in a greater net vertical force than sitting still (about 40 lbs max for my "Excel" car). Since F=ma this force only sticks around for the time it takes to raise this simulated end of my car (force divided by springrates, inches).


But... on my excel car, the net normal force increases by 55 lbs on the outside and decreases by 15
lbs on the inside. This is a net normal force INCREASE. This would increase grip for a second and then return to normal (G's*cg*mass/track) weight transfer situation. I can't see any reason to complain about that.

Is this right?

Also, When trying to get a complete picture of normal loads through a corner, I have unsprung and geometric (near instant) and elastic (slower) weight transfers ?PLUS? jacking forces. Do the jacking forces appear instantly (as fast as the lateral force can occur) and then dissapear according to f=ma?

Thanks

[Z]

Brian,

Re: First paragraph - "If we are in a corner...". Spot on.


Re: Second paragraph - "... would increase grip for a second ... I can't see any reason to complain about that."

Well, no complaints, not for that (split) second. BUT! The increased Fz force has pushed the body upwards giving it a bit of upwards momentum. The spring forces (let's assume they're vertical and directly over the wheelprint) decrease as the body goes up until "return to normal (G's*cg"mass/track)...". However, the upwards momentum of the body continues and total net Fz force then decreases for another split second. Its a bit like jumping upwards - greater footprint pressure as you jump, followed by lessor pressure as you take off, followed by greater pressure when you land again ... until everything is damped out.

So, any variable lateral Fy forces acting on steeply inclined n-lines create varying Fz forces that are a bit like those over bumps. Usually you don't want to drive over the bumpy part of the track, but sloping n-lines can turn a smooth track into a bumpy one (although Fz = Fy*n-line-rise/run, which is usually small, so the "bumps" aren't that bad).


Re: Last paragraph "Also, when...".

Yes, "the jacking forces appear... as fast as the lateral force can occur", but no, they don't disappear. They end up equal to (Fz=Fy*rise/run) and this increased n-line force offsets the decreased spring force due to the spring now being a bit longer (higher ride height). At steady state (n-line-Fz-force + spring-force = Net-Fz-wheelprint-force-due-to-lateral-load-transfer). And, don't forget that as the ride height (at each wheel) changes you also have to include damper forces.

So, at each wheel the net Fz-wheelprint force includes:
1. Spring force - function of wheel-to-body position (slow acting).
2. Damper force - function of wheel-to-body velocity (faster acting).
3. N-line Fz force - function of lateral-Fy-wheelprint-force (fastest acting). (This is sometimes said to be a function of lateral body-to-road acceleration, but that's not really right for individual wheelprints).
4. Force due to lateral acceleration of the wheel-assembly (ie. "unsprung mass", but I don't like this term because it is not really "unsprung"). (This contribution to wheelprint-Fz-force depends on front-view-instant-centre position. This is a long story, but briefly it is treated wrongly in most textbooks, where it appears in the equations as if the IC is at infinity. Since the wheel-assembly/body mass ratio is fairly low (0.1 to 0.2) then these "mistakes" don't make too much difference. IC at infinity is same as adding wheel-assembly mass to body.)

Err, I think that's all....


Z

[Z]

Here's another reason why NOT to think about "roll-centres".

Picture the suspension in the usual "end-view". Assume the right-wheel (in the end-view) has an n-line sloping at 1 degree upwards from horizontal towards the car centreline (to our left). The left-wheel's n-line slopes down towards the car centreline at 1.000,001 degrees (ie. at one-and-one-millionth degrees).

The roll-centre, as it is usually defined, is at the intersection of these two n-lines. By my calculations (on a small scrap of paper) the RC is 1,200 kilometres to the left of the car, and at a height of 21 kilometres above ground level (that's more than twice as high as jumbo jets fly)!!!!

Is this good or bad? Some experts, who don't like wandering RCs, would be dumbstruck by this RC position.

What is the effect on the car of this RC position?

Assume a leftwards lateral force Fy at each wheelprint. These cause forces along the n-lines (and also small Fz forces of the previous posts). We vectorialy add the n-line forces at the RC, and then decompose this resultant RC force into horizontal and vertical components. The horizontal "cornering" component of the resultant RC force, acting at the RC altitude of 21kms, creates a HUGE anti-clockwise moment about the car's CG. The vertical "jacking" component of the RC force, acting 1,200kms to the left of the car, creates an even BIGGER clockwise moment about the car's CG. WOW!!!

The sum of these two massive moments is a clockwise moment approximately equal to 2*Fy*CG-height. Since CG-height is about 0.5 metres this moment is a small and quite normal roll-moment acting on the car.

There is nothing wrong with this suspension. In fact, the n-line slopes of only about 1 degree from horizontal would indicate benign handling, all other things being equal.

The "roll-centre over the rainbow" is just a confusing distraction. It is better to think in terms of n-line slopes.


Z