Uh oh... one of those threads again (Roll centre/mac struts)

[RomanV]

One thing that has always bugged me, is the idea of roll centres, or roll centre adjusters.

A lot of people seem to think that RCAs are good, but cant seem to explain why or how they actually do anything.

My understanding, is that the car has a roll centre, and a centre of gravity.

The distance between the two is important, in the same way that using a power bar makes it easier to undo a stubborn bolt compared to a short ratchet...

The COG is the force that tries to tip the car, and the roll centre is the point that the car actually tips on.

If your roll centre moves far away from your COG, then the force required to make the car roll X amount is reduced.

And then from there it gets exponentially worse...
Because since its easier to tip the car over further, the car tips more, making the roll centre even worse, because the lower arm is on an even worse angle, making it easier to roll further, etc etc.

So doing some digging around, I found a diagram explaining
what actually determines roll centre position:



You draw a line perpendicular to the strut top, and then draw another line running through the angle of the lower arm. From this intersection point, you draw another line back down through to the point where the strut angle intersects with the ground.

Where this line intersects with the centreline of the vehicle, is the roll centre.

So the variables which actually determine roll centre position in in a car are:

Strut angle
lower arm angle
lower arm length

Tipping the angle of the struts in towards the centre of the car, raises the roll centre.

Having the lower arms angling 'down' towards the wheels raises the roll centre.

having longer lower arms raises the roll centre (But only because it has the affect of tipping the angle of the strut over)

Looking at the diagram, and doing a few sketches in autocad, what makes the biggest difference to roll centre position is definitely the angle of the bottom arms.

So it makes sense that having the traditional type of RCAs which space the lower arm back down, affect the roll centre geometry.

But it's interesting that it's not the only variable that affects it, and what this can mean in terms of how you're choosing parts for your car.

For the 80s RWD cars, the most desirable front struts to get have been the ones which give the most natural amount of negative camber. However looking at this, using the struts which give the most positive camber, and then using adjustable strut tops to lean your struts back over as much as possible will actually improve your front roll centre, so long as you can still get the desired amount of camber.

Also worth noting, is that this is all regarding static roll centre, obviously it changes as the cars suspension compresses etc.

Fitting longer lower arms reduces the angle that the lower arms move through per mm of suspension travel, so your dynamic roll centre moves less through out the range of travel, so your car will actually sit flatter through corners.

So it makes sense that roll centre position is important, and its interesting to note the ways that you can change it. Having a better roll centre means you need less in the way of swaybar stiffness to resist roll. Which means your suspension can work independantly more effectively over single side bumps, while staying just as flat through corners.

A lot of people get adjustable camber plates and dial in a shite load of camber, however I wonder if the actual improvement comes from the revised strut angle instead.

On a slightly related note, people are always keen on getting the shortest steering arms possible, to give the maximum amount of lock (presumably for drifting)

But the longer your steering arms are, the less bump steer you get... So to minimise bump steer, ideally you'd want the longest steering arms that give you the requried amount of lock, not other way around.

So, that's my rant for the day.

For those of you with a short attention span, here's a meerkat doing a rock to fakie.

[Bazda]

Its not always due to the angle of the bottom arm, its where the pivot point is on the bottom of the hub.
On RWD cars where the ball joint is upside down effectively putting a spacer in there will work. But on FWD cars most have the ball joint the other way round where the joint pokes into the bottom of the strut, So they think putting a spacer in the bottom arm will help their RCA, but this does nothing at all in this case.

[RomanV]

Have you got a pic to explain what you mean?

Are you talking about the type where there is a hub assembly casting with the steering arm built in which bolts to the strut, rather than having a stub axle coming off the strut itself, and seperate steering arm?

I've always wondered how RCAs would work in that case, short of redesigning the position of things on the hub casting.

Yeah I'm just talking about the type where the steering arm is sepearate, so the RCA spacer moves the lower arm and steering arm back down.

[h8wrxs]

For those of you with a short attention span, here's a meerkat doing a rock to fakie.

thats me

[RomanV]

Its not always due to the angle of the bottom arm, its where the pivot point is on the bottom of the hub.
On RWD cars where the ball joint is upside down effectively putting a spacer in there will work. But on FWD cars most have the ball joint the other way round where the joint pokes into the bottom of the strut, So they think putting a spacer in the bottom arm will help their RCA, but this does nothing at all in this case.

Aaahh! I get what you mean now.

Effectively the lower arm angle in the diagram is between the ball joint that the strut sits on, and the pivot point of the lower arm on the chassis.

Not the actual arm itself... makes sense.

[Malcolm]

On a slightly related note, people are always keen on getting the shortest steering arms possible, to give the maximum amount of lock (presumably for drifting)

But the longer your steering arms are, the less bump steer you get... So to minimise bump steer, ideally you'd want the longest steering arms that give you the requried amount of lock, not other way around.

Longer arms don't necessarily mean less bump steer, it's all about the arcs that the arms are moving through (which I'm sure you realise), but as you move the steering arm higher the "ideal" length becomes shorter too, so if you steering arm was the same height as the lower control arm, you'd want it to be the same length, but if it's halfway between the lower arm and the top mount of the strut, you want it to be half the length of the lower arm.

For the 80s RWD cars, the most desirable front struts to get have been the ones which give the most natural amount of negative camber. However looking at this, using the struts which give the most positive camber, and then using adjustable strut tops to lean your struts back over as much as possible will actually improve your front roll centre, so long as you can still get the desired amount of camber.

This is where MacPherson strut is teh gays. Doing so will give you massive kingpin inclination, which gives you increasingly positive camber on both wheels as you steer (KPI = bad, caster = good).


The other thing that is worth discussing is whether a higher roll centre is even "better". Some of the trade offs when increasing roll centre height are:
- jacking forces, the line of force from tyre contact patch to the roll centre has a vertical component, and this component gets greater as you increase roll centre height. This has the effect of a net increase in ride height (i.e. average left to right) as the car rolls, increasing CoG height.

- less tunability, as you increase the roll centre height, the split between "elastic" weight transfer (that reacted by the springs, dampers and anti roll bars) and geometric weight transfer (the moment of the roll centre about the tyre contact patch) gets closer to the geometric side. As an example, if you have a roll centre at ground level, you'll have 100% elastic weight transfer, so doubling your spring rates will double your roll stiffness. If your roll centre height is halfway between the ground and your CoG height, then you have 50% elastic and 50% geometric, so doubling your spring rates will only increase roll stiffness by 50%. And FWIW, if your roll centre is at the height of your CoG, the car wont roll at all and changing your springs or ARBs will have no effect on the car's roll stiffness. When you consider the importance of getting the right front-to-rear distribution of roll stiffness to achieving the right understeer/oversteer balance, it's obvious that you need some degree of tunability to get it right

- time response (sorry, this is getting a little heavy). Going back to the stuff about elastic vs geometric weight transfer, geometric weight transfer is seen basically immediately at the contact patches, that is, as soon as you have a lateral acceleration, there will be a change in your left to right tyre loads. The elastic portion builds (relatively) slowly. If you just consider springs/arbs for a second (i.e. not dampers), the reaction force provided by a spring doesn't increase until the spring is compressed, in other words you don't see full weight transfer to the outside wheels until the vehicle has reached full body roll angle. This means a higher roll centre will make lateral weight transfer happen more quickly (typically), which is generally a good thing, but if it happens too quickly, it can have bad effects on the vehicle's stability, and the ability of the driver to maintain control (a more reactive car will tend to lose grip more suddenly, with less apparently warning). Another problem on time response is to do with the time response of the front axle vs the rear axle. It's not uncommon to have the roll centre higher at one end of the car than the other, to try and achieve/combat certain handling characteristics on turn entry and exit (i.e. having a lower front roll centre slows the front weight transfer down, making the roll stiffness biased further toward the rear at corner entry and therefore making it more oversteer biased on entry).

Hmm, might add more later, should be working...

[postfach]

Hold up - don't you want to lower your roll center?

[RomanV]

Epic reply Malcolm! Was hoping you'd pitch in.

I'll need a moment to think about most of what you've posted though, hahaha.

Longer arms don't necessarily mean less bump steer, it's all about the arcs that the arms are moving through (which I'm sure you realise), but as you move the steering arm higher the "ideal" length becomes shorter too, so if you steering arm was the same height as the lower control arm, you'd want it to be the same length, but if it's halfway between the lower arm and the top mount of the strut, you want it to be half the length of the lower arm.

Sorry I dont know the proper name for the... aahhh... end of the steering rack arm type things which connect to the steering arms on the bottom of the strut.

It's the difference in length between the lower arm and the steering rack end... thing... that makes bump steer right?

Because they travel on different arcs, the steering rack arm end thing effectively pulls itself in or out, relative to the main lower suspension arm.

If you have the option of two steering arm lengths, a longer steering arm will rotate the strut less, for every mm of side to side movement from the steering rack end thing.

Which is obviously why shorter arms give you more lock, but they also give you more bump steer (asusming you are not changing anything else)

Hold up - don't you want to lower your roll center?

Generally your roll centre lowers by a hell of a lot when you just lower your car and the suspension angles change. RCAs etc bring the roll centre back up again.

if you have a roll centre at ground level, you'll have 100% elastic weight transfer, so doubling your spring rates will double your roll stiffness. If your roll centre height is halfway between the ground and your CoG height, then you have 50% elastic and 50% geometric, so doubling your spring rates will only increase roll stiffness by 50%.

Yes, but since less roll stiffness is required, you can effectively run lower spring rates to acheive the same amount of roll resistance?

So you get the benefit of soft spring rates for going over bumps etc, but still good roll resistance during cornering without reducing the independance of the suspension with stiff swaybars etc...

[Malcolm]

Sorry I dont know the proper name for the... aahhh... end of the steering rack arm type things which connect to the steering arms on the bottom of the strut.

It's the difference in length between the lower arm and the steering rack end... thing... that makes bump steer right?

Because they travel on different arcs, the steering rack arm end thing effectively pulls itself in or out, relative to the main lower suspension arm.

If you have the option of two steering arm lengths, a longer steering arm will rotate the strut less, for every mm of side to side movement from the steering rack end thing.

Which is obviously why shorter arms give you more lock, but they also give you more bump steer (asusming you are not changing anything else)

Yep, yep, you're right, you were using the right terminology and I was assuming you weren't, haha! Steering arm is the part that attaches to the upright and creates a moment to cause it to rotate about the kingpin, as you say a shorter one gives more lock and will magnify the effects of bump steer. The tie rod is the arm that runs roughly parallel to the lower control arm and connects the steering arm to the steering rack.

Yes, but since less roll stiffness is required, you can effectively run lower spring rates to acheive the same amount of roll resistance?

So you get the benefit of soft spring rates for going over bumps etc, but still good roll resistance during cornering without reducing the independance of the suspension with stiff swaybars etc...

Yep, definitely, this is the general direction I started thinking doing suspension design for the FSAE car - if you have a traditional suspension set up with a coil spring on each corner doing most of the work in pitch, heave, roll and warp (the 4 chassis modes, heave is all 4 wheels moving in unison up/down and warp is diagonally opposite wheel pairs moving together), then you end up with your stiffness in one mode affecting all the others, so you use the height of your pitch and roll centres to achieve you desired stiffness in these (rotational) modes, while using spring rates suitable to achieve you goal stiffnesses for the other modes. There's a bit more to it that than, because obviously you have the effects of ARBs and dampers to consider (and there's always the possibility of using anti-pitch springs, "z bars", and roll dampers), but it's often better to keep a simple view of these things rather than getting bogged down by overcomplicating and trying to get things "perfect"

[Bazda]

maybe you mean center of gravity postfach...

I just went through this with my car and made secret custom bits to change all the geometry so yes i can run a much lower spring rate as well. Didnt really like running 14kg springs :S.

[wde_bdy]

Most of the RWD mcpherson strut RCA's though are mis-named and have nothing to do with roll centre, they are simply there to return the suspension back to a factory range of motion and reduce bumpsteer. Factory setups are designed to operate within a range that has minimal bumpsteer while keeping handling predictable, move out of that range though and bumpsteer is amplified. RCA's put you back in the factory range for cheap without redesigning your whole suspension.

Callum

[RomanV]

Here's a sketch of 'normal' lower arm angle, and the construction lines that show where the roll centre ends up: (yellow dot is roll centre)



And here's when the lower arms are back up near horizontal, thanks to the car being lowered:



When the lower arms (and the ball joint) are spaced back down, it does put the roll centre back further up again... But only if as Bazda mentioned, that the ball joint gets spaced down too. So you could be right there Callum.

Juts expanding on a bit of a ramble above, about the narrow and wide body 80s cars.

The narrow body cars generally had smaller/lighter engines like 4AGEs and 3AUs etc, so presumably a lower centre of gravity, with a smaller force acting apon the roll centre due to the smaller engine mass.

So the strut towers/tops are more upright, and close to the outside edges of the engine bay. With a smaller kingpin inclination angle on the strut, and (presumably) less castor, which means less steering effort when very few of these cars had power steering.

The wide body cars generally had bigger heavier engines like the 5MGE, 1GGTE, etc etc... So higher centre of gravity, larger mass in the engine bay requiring greater factory roll resistance... A higher factory roll centre to accomodate the higher centre of gravity and higher total force acting apon the roll centre.

These cars have got the struts further tipped over towards the inside of the car, with the top mounting points further inboard... And larger kingpin inclination angle, to give the right amount of camber on the stock setup.

Also presumably they would run more factory castor, increased steering effort is irrelevant because they pretty much all have power steering anyway.


My issue, is trying to raise the roll centre back up on a narrow body car, where there's only so much space for the spacer block type RCAs in 15" wheels, and maintaining ground clearance to the lower arms.

Mimicking the wide body strut angle with adjustable camber plates on the top of the strut, and using wide body struts, would give me additional roll resistance as per the factory widebody cars.
Without giving me stupid amounts of negative camber.

Running more castor to cancel out increased kingpin inclination angle wouldnt be an issue...

Increased steering effort on account of more castor is irrelevant, because my car has never left the garage

[loudstealthGT-Four]

I too, am still stuck on the meerkat doing a rock to fakie.

[al_feinted]

Wow, they are a couple of smart mere cats!

Once you drive the car you will be saying "Damn, my carina handles awesome" then all this roll centre stuff will be pushed to the back of your brain never to cross your mind again