How Not To Modify Your 205's Suspension...

[engine killer 5]

By the way,

 

Has anyone got any height measurement (from edge of the plastic fender to the ground) for a stock car (non-lowered)?

 

Thanks in advance.

[Cameron 16]

I've been doing some homework on this as part of a little (read massive) project I'm working on, and found some pretty interesting results about front roll centre height and how it affects body roll and load transfer. Annoyingly I still don't have internet access at my new place so I can't upload any of the info. I will do as soon as I can though.

[kyepan 291]

would be a really interesting read!

[engine killer 5]

I've been doing some homework on this as part of a little (read massive) project I'm working on, and found some pretty interesting results about front roll centre height and how it affects body roll and load transfer. Annoyingly I still don't have internet access at my new place so I can't upload any of the info. I will do as soon as I can though.

 

 

Great thanks!

 

Under normal circumstances the roll centre is below center of gravity, I also want to know if roll centre is the same point as center of gravity, what will happen?

[Cameron 16]

You will get no body roll during cornering, but a lot more weight transfer.

 

The roll centre is basically the (imaginary) point at which a the body pivots, and as such, the point where a lateral load can be applied and cause no body roll. Normally when the RC is below the CG height, the CG produces a torque (known as a roll moment) around the roll centre. This is because you have a lateral force (g-force, centrifugal force, whatever you want to call it) acting on the CG, and force x distance = torque. This torque translates through the vehicle track to an addition / subrtaction of load at either side of the vehicle - load transfer.

 

When the RC is at CG height, there is no distance for the CG lateral force to produce the roll moment (torque), so none is produced. BUT the force is instead applied directly to the roll centre, and since the roll centre is above ground level you will get a torque produced, and still get weight transfer.

 

What I've found is interesting because it shows that you get less weight transfer with a lower roll centre height, but at the expense of more body roll. Less weight transfer is always preferable as it means increased grip, but with a McPherson strut suspension system more body roll is definitely not preferable as it means positive camber change. A happy balance needs to be struck between the two.

[Rippthrough 98]

Which is usually why you have a lower front roll centre than rear.

 

HOWEVER (and it does need the capitals), geometric roll centre only tells half the story, even if you had the roll centre slightly above the centre of gravity, the car would still roll dynamically.

The best thing you should be aiming for at DIY level really (given what we can alter), rather than altering a few mm of it's height, is trying to keep the roll centre height consistant with the chassis movements throughout your new range of travel, which will generally give a nice predictable car which is easy to alter the attitude of with the throttle.

Having it too low usually results in a lot of migration laterally of the RC on strut setups too.

 

On the McPherson strut setup the ratio of movement of the RC compared to the chassis does change a fair bit depending up on the height of the car, so resetting it so the movement is similar to a stock setup is still a good idea, but don't worry too much about the height + anti-roll effects, as it doesn't control all of it.

Edited December 2, 2010 by Rippthrough

[Cameron 16]

My 205 currently has a RC height of -69mm. Not ideal!

[Cameron 16]

Ok so here is, as promised, the results of my little experiment with roll centre height. This is probably going to be a bit of a long post, and will probably need to be read in full to understand what's going on in the graphs. I'll try and explain as simply as possible.

 

To start off, I had to take a bunch of measurements off my 205, as without knowing the hard points (the susp pivots etc) there is very little you can do. I tried to be as accurate as possible but it's difficult to work around a complete vehicle in a single garage, in the cold. Anyway, I put these points into a program called Lotus Shark which is a suspension simulator, to find out what my 205's roll centre height was in it's current (very) lowered state. Turns out it's at approximately -70mm!

 

The great thing about this program is you can see exactly how the suspension behaves in cornering, and it will output some nice graphs showing camber change, toe change, roll centre height, etc etc! With a McPherson strut suspension you get very little camber compensation - meaning your camber change almost equals your roll angle. With this in mind it's probably a good idea to keep roll to a minimum, hence fitting ARB's as standard.

 

So anyway, using this program, my measurements, and what I know about my car, I made a spreadsheet in Excel that could use all this data to calculate body roll angle and load transfer for different roll centre heights. I chose RCH as the only variable to keep things nice and clear. All this data went into this little table:

 

 

Using the following coordinate system:

 

 

The first thing I calculated was the body roll rate - measured in degrees per lateral g - plotted against roll centre height. This has to start by calculating the centre of gravity (CG) height above the roll centre (RC) axis (the line connecting front and rear RC's). This is done using the front and rear roll centre heights (RCH), the centre of gravity height (CGH), the centre of gravity x position and a bit of trigonometry. Why must this be done first? Because a roll centre is the point at which the body will rotate, so it follows that the body will rotate about the RC axis.

 

How much the body rotates (rolls) about this axis then depends on the roll stiffness - the resistance to rolling provided by the springs & anti-roll bars - and the CGH above the roll centre axis. This all gets calculated and produces the following graph:

 

 

So you can quite clearly see that the lower the roll centre is, the higher your body roll angle will be for the same cornering force. This is because the distance between the CG and the roll centre axis increases as the front RC height decreases. This then increases the torque created by the lateral force acting on the CG (think centrifugal force) having a longer moment arm (larger distance) since torque = force x distance. Think of the roll stiffness as a clock spring, if you had a lever connected to the centre of the spring to apply a torque, you will rotate the spring a certain amount; if you then increase the torque, you turn it more.

 

I also need to point out that the changing roll centre height only relates to the front roll centre, since the rear roll centre will always be at ground height due to the trailing arm design. That's why when the front RCH is 500mm i.e. at CG height there is still body roll, since the roll centre axis slopes steeply down towards the rear.

 

So this all looks pretty bad, right? You'd think from looking at that graph that you want the RC axis to be at the CG height, as then you would get zero body roll, but there's more to the story than this. This is where I have to add a little more depth, so I'm sorry if I confuse anybody!

 

The next thing to try and understand is how the changing roll centre height affects the weight transfer front and rear. I've heard a lot of people saying that weight transfer is a good thing, since the amount of grip provided by a tyre is proportional to the amount of weight (which I'm going to call vertical load from now on) on that tyre, so more vertical load = more grip! This just isn't the case in real life though, because with rubber the relationship between vertical load and grip isn't linear - the grip coefficient decreases as the vertical load increases. So if you have load being transferred from the inside to the outside tyre, you will see a reduction in the total grip level for that axle. With this in mind, you want to have all 4 wheels carrying an equal load in order to get the maximum possible grip, and you want to have both front and rear axles having an equal grip level to have a neutrally handling car.

 

So where does the roll centre height come into this?

 

As I said earlier, the cornering force acting on the CG causes a moment (torque) about the roll centre axis (RCA), equal to the force x the height above RCA. This is called the rolling overturning moment (ROM) since it acts to roll the body, and also to overturn the vehicle. But since the RC is the point at which the body rotates, the cornering force is also reacted here, thus producing a second moment called the non-rolling overturning moment, which acts around a point below the RC at ground level. So the NROM is therefore proportional to the RCH x the cornering force. The load transfer will then be equal to the sum of these two moments, divided by the track width. Note that this is calculated for each axle.

 

What's interesting is that if the RC is below ground level the NROM will be negative, since you have to calculate all moments in one direction (e.g. clockwise), and will act to cancel out part of the ROM. This means you will get less load transferred when the RC is below ground, which means a more even load distribution on the tyres, which means more grip!

 

I hope you managed to follow that! Here's a pretty graph to illustrate my point:

 

 

So the blue line is the front load transfer (in Newtons), which goes from about 1,000 at RCH = -100mm, to about 3,600 when RCH = 500mm.

The red line is the rear load transfer, which goes from 1,700 at -100mm to about 800 at 500mm.

The dashed line is the body roll angle, which goes from nearly 3 degrees at -100mm to about 1.4 degrees at 500mm.

The green line just shows the total front and rear load transfer.

 

What's interesting to note is the difference between the front and rear load transfer, and how this relates to the handling balance of the vehicle. When the RC is at it's lowest at -100mm there is slightly more load transferred at the rear, making for a naturally oversteering car. The front and rear load transfer amounts are equal at around 30mm RCH, which would create a neutral handling car, and from that point on the front load transfer overtakes the rear, tending more and more to understeer.

 

The standard roll centre height is somewhere in the region of 150mm, where there is an 800N difference between front and rear load transfer, causing a slightly understeering car. Basically what all manufacturers want as it's nice and safe!

 

The problem with all this comes when you look back at the body roll angle and remember what I mentioned earlier about how camber change is proportional to body roll. You may well get less load transfer with a lower RC height, but you also get more body roll; which could take you into a situation of nasty positive camber and mean you actually end up losing grip! You therefore need to strike a balance between the two - a compromise.

 

The problem with all this is that it hasn't accounted for the change in CG height with lowering, since I only wanted to have one variable to keep the explanations as simple as possible. Hopefully if you've understood everything so far you'll be in a much better place to understand things when I take it to the next level, and find out how lowering the car affects things. I'll be using the suspension simulator to calculate camber change, and comparing that to the load transfer as you go from standard ride height to lowered ride height. I'll try and get this done early next year.

[Rippthrough 98]

And then when you've had all that fun you'll want the lateral movement under roll too

 

Just keeping you busy over Christmas

[kyepan 291]

very interesting cameron, well done.

 

So in summary - a roll center at 30mm off the deck = balanced front rear, what front ride height does this correspond to?

 

secondly, standard roll center height is 150mm vs -100 for your lowered car, there is only a roll difference of 0.5deg, will this affect camber or actual roll that much?

 

are the relative corner weights of the car going to factor in because they affect the vertical load on each car... or is that taken into account in the initial calculation (weight distribution)

 

and one would imagine that that relative vertical loads change front rear when for example, trail braking into a corner.. vs neutral throttle mid corner.

 

Please upload the camber stuff when you get an opportunity, this is fascinating, and i feel like some kind of rule of thumb optimum set up is on the cards here.

 

Cheers

J

[Cameron 16]

And then when you've had all that fun you'll want the lateral movement under roll too

 

Just keeping you busy over Christmas

 

Do you though? I know the roll centre can migrate laterally but as long as the height doesn't change it shouldn't make too much difference. Don't all the forces still resolve to more or less the same moments if the RC height stays the same? Please correct me if I'm wrong! I was just told that roll centre migration isn't such a concern as long as the height remains as constant as possible.

 

very interesting cameron, well done.

 

So in summary - a roll center at 30mm off the deck = balanced front rear, what front ride height does this correspond to?

 

secondly, standard roll center height is 150mm vs -100 for your lowered car, there is only a roll difference of 0.5deg, will this affect camber or actual roll that much?

 

are the relative corner weights of the car going to factor in because they affect the vertical load on each car... or is that taken into account in the initial calculation (weight distribution)

 

and one would imagine that that relative vertical loads change front rear when for example, trail braking into a corner.. vs neutral throttle mid corner.

 

Please upload the camber stuff when you get an opportunity, this is fascinating, and i feel like some kind of rule of thumb optimum set up is on the cards here.

 

Cheers

J

 

To be honest, I can't answer your first question as I don't know how far my 205 has been lowered from standard. My estimation of the front RC height at 150mm is only based on me adjusting the ride height in the software until the wishbone angles looked more or less like standard, the error could be in the region of 20-50mm! My 205 currently has a RCH of -67mm at about -80mm in ride height.

 

Yes, the corner weights will be a factor, but they don't need to be accounted for when calculating load transfer. The figure for load transfer for each axle is halved then added and subtracted from the static loads on the outside and inside wheels respectively, to give the value for load on that tyre during cornering.

 

Yes trail braking will affect things but the process for calculating that would be EXTREMELY complicated! Something that is best left for quite advanced simulation! Everything I'll be looking at here will be steady-state, so representative of how the car will behave mid-corner.

 

I'd love to be able to work towards a recommendation, it would also be very interesting to be able to validate everything with some on-track testing!

Edited December 21, 2010 by Cameron

[Rippthrough 98]

It does alter the ROM slightly with lateral migration, nowhere near as much as the RCH though.

[Cameron 16]

I think for the purposes of anything up to the higher levels of motorsport it's probably safe to ignore it, as long as the RCH remains constant through the migration.

Edited December 21, 2010 by Cameron

[Rippthrough 98]

Probably, given the 205 doesn't have that much travel, but you looked like you were having fun

Edited December 21, 2010 by Rippthrough

[Cameron 16]

Cheers!

[oli-pug 6 2 Cars]

Impressive stuff Cameron... makes me wish i'd knuckled down properly now and gone to Uni!

[Henry 1.9GTi 36]

nice. I have made a quick load transfer spread sheet for anyone interested. Unfortunately the derived results based upon front and rear track differences arnt accurate as it doesn't take into account roll center analysis. And also the change in the fraction of load transfer front to rear does not take into account wheel lift off. Is this something lotus shark can model cameron? As 205s lift a rear wheel as do most fwd tin top racers and it seems something worth modelling. Also during wheel lift off I assume the roll centre becomes the contact patch of the wheel still on the road. A good way a controlling its height!

 

http://www.sendspace.com/file/min2k2

 

I have some tyre data with slip angles against lateral force for varying normal loads at 3 different pressures. I was going to try and balance my 205 for slip angle, assuming identical front to rear tyres, by changing the weight transfer and tyre pressures but havnt got round to it yet. Also the 205 is going on the 4-poster late Jan, should be interesting

 

keep up the work cameron really good read.

Edited December 24, 2010 by Henry 1.9GTi

[Cameron 16]

Not that I know of, Shark is really just a suspension design program. It'll show you kinematics for bump and roll, and tell you all sorts of handy values (like anti-dive / squat, RC heights etc etc) but doesn't do load transfer. It disregards roll stiffness, so you need to calculate what your body roll angle will be then feed it back into the program.

 

It's quite easy to work out when lift-off occurs, basically when the load transfer is greater than the static load on the tyre, the wheel will lift. I'm pretty sure the roll centre position will stay the same when the wheel lifts off. You get very large RC migration (side-to side movement) during roll with a trailing arm suspension but it will always stay at ground level.

 

What tyres is your data for, and how did you get hold of it?

[Henry 1.9GTi 36]

its a generic 195/60/15 road tyre a lecturer gave me the data, not sure what brand it is although I think its quite recent. He said despite being different to the actual tyres on the car it will put you in the right ball park to balance the vehicle as the front and rear tyres are the same.

 

Can I ask a favour and get a copy of your hardpoint measurements? Would love to chuck it all into ADAMs and have a play.

[Cameron 16]

Yeah I can give you the hardpoints, just need to get them off my PC so will probably be tomorrow evening.

[Henry 1.9GTi 36]

cheers Cameron.

 

Some interesting results from this:

 

 

Absolutely masses of data which I have yet to go through or understand. But there was an issue with the front; ~45kg constant friction in the front suspension. Causing the body to be dragged around by the suspension until the force was overcome. Disconnected the front ARB but made no difference. Also wound the dampers to full soft and made no difference. Gonna get them off and put them on the damper dyno to see whats going on. On the plus side the rear was very well controlled and showed good characteristics. Better corner weighting and getting rid of this stupidly high amount of friction in the front and the car should be better to drive as well as having more availble grip. Computer model of the rig testing to come so I can get it right there and then get it onto the car right first time.

 

Dodgey conrer weights were causing the car to roll when subject to a heave only input. Will try and get some graphs out of the matlab data and do a write up when I have an idea of what I'm looking at

[Henry 1.9GTi 36]

Heres a quick indication of the friction. Rear suspension on the left, front and the right. Strut displacement vs force as you can see a very large force in the front before any movement occurs.

 

[Rippthrough 98]

What front struts are on the car?

 

edit: N'mind, seen the sig, I can't see it being normal static friction in those as the seals are so small and the gas pressure is low, unless one's knackered, or oil's not getting to the upper shaft bush.

Edited January 26, 2011 by Rippthrough

[Henry 1.9GTi 36]

Don't want to slag them off just yet as it may be something else contributing. Will get the dampers on the dyno first, although this wont account for any bending moments induced by the suspension geometry. The rears were freshly fitted Tarmac spec Bilsteins and the fronts are in my sig

[Cameron 16]

Are you using Polybushes?