top view geometry

[spacer]

Hi all,

I have two simple questions regarding suspension geometry.

I'm currently in the process of designing an gravel rallycross chassis in SolidWorks to be built this summer, and am currently doing the front suspension. Car is FWD, front double wishbone (damper actuation method decided later on), designing for +80mm / -80mm suspenion travel.
I'm still a beginner that has read some of the recommended books (carroll smith etc.) and have now set up a quarter-car model in CAD to try and evaluate some kinematics in bump/drop/roll.

**First off, the books I've been reading lately all seem to mention suspension-design/evaluation methods which look at front view geometry and side view geometry in order to evaluate your kinematics.
However, I can't really seem to find (or get my head around) the effects of changing pickup points in top view.
Particular case being double wishbone setup, moving a inner wishbone pickup point laterally. Tons of cars run pickup points that don't have their inner pickup points located along the same longitudinal axis. Is this a packaging compromise (F1 gearbox mounted rear wishbones for example) that can be neutralized elsewhere in your geometry, or can this be used to some benefit?
At this point I believe it can be used to alter caster in bump/droop, but not sure of any secondary effects or to what benefit this might be.

**Second question, is my following conclusion correct:
Trying to get little camber change in bump/drop is always a compromise with trying to get little camber chang in roll (i.e. camber change rate 1:1 vs roll angle). The only parameter defining how big this compromise will be is track width.

[flynfrog]

other than the steering geometry the top view pickups have no effect on kinematics. They can be placed to reduce bending moments on suspension parts under breaking and acceleration.

[Greg Locock]

Top view Y axis placement of the inner hardpoints, staggering them fron to rear, on the wishbones is used to give particular castor effects, and does affect antidive (from memory). You can also get some geometric wheel recession in jounce, which is handy for reducing impact harshness. It isn't the only way to get those effects, and compared with the more usual approach I have never found that the supposed benefits really show up enough to make up for the basic complexity caused by setting up strange hinge lines for the wishbones.

Incidentally you might find that 80mm of rebound travel (especially) is lacking compared with your competition.

[Tim.Wright]

Apart from what Greg has mentioned I think there is often a packaging reason(s) for this setup.

Sometimes you might set your front view geometry but find you can't put one of the top wishbone hardpoints where you want. For example you might want them at y=300 but the engine or driver is in the way. One solution could be to put the troublesome one at y=400 to clear the obstruction and the other one at y=200 in an area away from the obstruction. Then your effective front view pivot point will lie somewhere between 200 and 300 depending on the longitudinal location of the wishbone pickups. I did something similar in my FSAE days on a rear axle to clear the engine (see below, mainly the lower wishbone). The front points of the wishbones were out wide to connect to the main roll hoop and the rear ones were closer to the centreline of the car.



Its not a perfect workaround because there are knock-on effects like Greg said but ok, in the end everything is a compromise to some degree.

You are right though, in that there is not a lot of publically available material on these effects. Pretty interesting discussion.

Tim

[flynfrog]

Top view Y axis placement of the inner hardpoints, staggering them fron to rear, on the wishbones is used to give particular castor effects, and does affect antidive (from memory). You can also get some geometric wheel recession in jounce, which is handy for reducing impact harshness. It isn't the only way to get those effects, and compared with the more usual approach I have never found that the supposed benefits really show up enough to make up for the basic complexity caused by setting up strange hinge lines for the wishbones.

Incidentally you might find that 80mm of rebound travel (especially) is lacking compared with your competition.

anti dive involves the angle between the two planes of the A arms. you can move the dots any where you want it wont chagne the anti dive

[Tim.Wright]

That geometric method of calculating antis is only realy valid in the case where the wishbones are hinged about an axis parallel to the longitudinal axes.

Also, the instant centre in your diagram there look like they are drawn from lines connecting the two chassis points of each wishbone. Instead, each one should be offset so its going through its respective outboard pickup point on the hub.

The reason for this comes back to the fundamental theory behind this method. You are trying to find the trajectory of two points on the wheel/tyre (not the chassis) and taking the perpendicular to those to find a common centre of rotation. This is then assumed to be the force/moment coupling point. The angle between the braking/acceleration forces and this coupling point tell you how much of your longitudinal forces are used to extend or compress the suspension against the action of the natural load transfer.

When you draw a line between the two wishbone joint locations, the line perpendicular to this is taken as the wheel trajectory (relative to the body). When your wishbones are not swinging on an axis parallel to the longitudinal axis of the car, then the trajectory is not a straight line between the two pickup points, but rather something of an arc.

The classical methods of deriving the anti characteristics then start to fall down at this point. Not such a big problem though. Front view and side view derivations are nice simplifications for a 3D case that with little effort will tell you 90% of how the suspension will behave. But don't kid yourself into thinking that they tell the whole story. You need to approach things from a force point of view to get the full picture.

[Greg Locock]

Yes I was using antidive in terms of 'things that stop the vehicle pitching' rather than the draftboard construction. Nice pic tho. So in all probability angling the inboard pickups will affect the RCH to some extent as well.

Having said that Tim's example is not an unusual approach for a rear suspension and I'm sure it can be made to work quite nicely. Placement of the toe link might cause some headaches, depending on what you want.

If you want to see an ugly example look at the Mustang Cobra IRS - first thing I do with them is junk the UCA and start all over again.

[spacer]

First of all, thank you all for commenting! I'd also like to thank marcush for sending some extended replies through PM but has disabled the option to send anything back

Top view Y axis placement of the inner hardpoints, staggering them fron to rear, on the wishbones is used to give particular castor effects, and does affect antidive (from memory). You can also get some geometric wheel recession in jounce, which is handy for reducing impact harshness. It isn't the only way to get those effects, and compared with the more usual approach I have never found that the supposed benefits really show up enough to make up for the basic complexity caused by setting up strange hinge lines for the wishbones.

Incidentally you might find that 80mm of rebound travel (especially) is lacking compared with your competition.

Apart from what Greg has mentioned I think there is often a packaging reason(s) for this setup.

Sometimes you might set your front view geometry but find you can't put one of the top wishbone hardpoints where you want. For example you might want them at y=300 but the engine or driver is in the way. One solution could be to put the troublesome one at y=400 to clear the obstruction and the other one at y=200 in an area away from the obstruction. Then your effective front view pivot point will lie somewhere between 200 and 300 depending on the longitudinal location of the wishbone pickups. I did something similar in my FSAE days on a rear axle to clear the engine (see below, mainly the lower wishbone). The front points of the wishbones were out wide to connect to the main roll hoop and the rear ones were closer to the centreline of the car.

https://lh3.googleusercontent.com/-yA4e ... ension.JPG

Its not a perfect workaround because there are knock-on effects like Greg said but ok, in the end everything is a compromise to some degree.

You are right though, in that there is not a lot of publically available material on these effects. Pretty interesting discussion.

Tim

Both these posts seem to confirm some of my suspicion (as well as marcush's comments); at this early design stage where I'm still busy figuring out raw camber curves, roll centre migration etc., it's better to avoid unnecessary diving into the top view y-axis placement as long as my packaging constraints don't mess up my desired kinematics too much yet?
At this point I'm also relatively free in placement of chassis hardpoints/bulkheads, so that might negate one of the possible reasons to dive into this.

Incidentally you might find that 80mm of rebound travel (especially) is lacking compared with your competition.

I'm triggered by your last sentence. The -80/+80 travel is actually based on the damper setup most of the field is currently running: 180mm travel dampers in macpherson struts often inclined to some degree, lowering damper motion ratio and resulting in some 160mm total wheel travel (to avoid some confusion, I'm merely adopting the currently used travel as a guideline, the design will be double wishbone no matter what).
So say the total damper travel is set due to cost/availability of our preferred damper cartridge, would you suggest sacrificing some bump travel in favour of rebound? Any rules of thumb regarding bump vs rebound travel ratio?

regards,
Tom

[Greg Locock]

That is very series dependent. Some circuit racers use rebound limit stops to get the car flat again ASAP, so that the underbody aero can start to work properly, whereas on a road car with RWD it can be handy to have 20mm more rebound than jounce travel at the rear, so you can get the power down on rough roads. I don't remember seeing an in depth survey on this, sadly.

Obviously the exact numbers you need depend on your spring rates and corner weights, and how you handle jounce bumpers and so on.

I can't really suggest a way forward on this, maybe use the standard shocks for now, and 80/80 like everybody else, but perhaps you could arrange that your design could go to 100/80 or 80/100 without binding up, if and when you can afford non standard shocks. I strongly suggest fitting some sort of travel detection to the shocks so you can see how much travel you are actually using - crudely you should be using as much as you can on each wheel.

[Jersey Tom]

However, I can't really seem to find (or get my head around) the effects of changing pickup points in top view.

So at this stage in general, are you trying to move pickup points around and understand what the downstream effect is? May want to consider taking a "top down" rather than "bottom up" approach, or thinking some from that direction. IMO pickup points can be turned into a no-brainer... if you spec out some high level suspension parameters (like VSAL, or camber curve, or various anti's) then often the hard points "design themselves" in that there is only one unique solution which best meets the performance characteristics you want.

**Second question, is my following conclusion correct:
Trying to get little camber change in bump/drop is always a compromise with trying to get little camber chang in roll (i.e. camber change rate 1:1 vs roll angle). The only parameter defining how big this compromise will be is track width.

For a SLA suspension, yes I believe so. Compromise is one word for it, choice would be another.

at this early design stage where I'm still busy figuring out raw camber curves, roll centre migration etc.

I'm curious - is the design intent to achieve some specific RC migration? If so, why? Are you more interested in the movement of a virtual point, or say.. the mechanical balance of the car? (More to the point of top-down vs. bottom-up). And don't forget, for as much as you can play around with geometry to get some RC movement or whatever, your actual jacking forces will be impacted by static camber, toe, Ackermann, tire pressure, tire construction, etc etc. Big picture...

[Greg Locock]

Ah, I missed all that.

What I do is decide on my target curves for each parameter of interest vs jounce travel (say toe, camber, RCH in practice, and castor for live axles) and then apply 'reasonable' tolerance bands to those curves. Then I run my favorite hardpoint analyser (depending on the project this might be ADAMS, wishbone.bas (it's free, and excellent), or something I have put together in excel or octave) with constraints on the hardpoint locations. After 3 or 10000 runs (depending) I end up with some candidate geometry that can be checked thoroughly for all the other parameters. Bear in mind that usually my outboard hardpoints are usually fixed in advance, as is rack height pretty much.


This of course assumes that I know what I want in the first place. Roll steer and RCH are set by my linear range understeer budget, camber gain is usually set by straight line traction and tire wear considerations, and castor for live rear axles is set by allowable and preferred UJ angles. Ackerman and geometrical antidive and antipitch and scrub radius and so on and so forth all need a bit of a think.

[Tim.Wright]

...Roll steer and RCH are set by my linear range understeer budget...

That caught my eye. Why do you use RCH instead of roll stiffness to influence understeer? Is this because the springs are chosen based on ride requirements so the only possibility left for tuning US is with RCH and stabi?

Tim

[Jersey Tom]

This of course assumes that I know what I want in the first place.

Indeed. I suppose that's ultimately my main point.. time is well spent figuring out what it is you want at a high level, and once you do all the hard points figure themselves out. Noodling on the hard points directly.. not the ideal approach!

[Greg Locock]

...Roll steer and RCH are set by my linear range understeer budget...

That caught my eye. Why do you use RCH instead of roll stiffness to influence understeer? Is this because the springs are chosen based on ride requirements so the only possibility left for tuning US is with RCH and stabi?

No Not at all. First I start with springs, to meet ride frequencies. then sta bar to try and meet roll targets, then RCH to meet roll targets, then roll steer and the other things mentioned in this http://en.wikipedia.org/wiki/Bundorf_analysis to meet the understeer target. However each parameter also has other constraints - for example you can't use too much roll steer or the car will be darty on rough roads. You can't use much compliance steer or the steering will feel rubbery, and so on and so forth. geometric RCH is a bit of a weird one - if a car has an IRS with a GRCH of 130mm, instead of a live axle at 305 mm, one might expect to notice some significant differences. There are measurable differences in steering response, but on a smooth road in the linear range I doubt many people could pick them apart subjectively.

However, RCH is usually constrained by camber gain targets so the whole thing becomes iterative. Then add in that you never really know what CGZ is going to turn out at, and the whole thing becomes a bit of an exercise