Suspension Design 4

[Tsiolkovsky]

Hello fellow engineers.

I am tasked to design a complete new suspension system for a concept race car. With no background in suspension design, I am currently in the self teaching/literature review phase (for 4 months and counting). I have done cad models but only as concept designs with little calculations backing them up.

This thread serves for me to ask, along the way, any questions I have regarding design of suspension. I am hoping there may be altruistic souls out there willing to shed light on some concepts that I do not yet grasp.

For today's question:

Why are there bushings that connect the control arm to the chassis mounting point? What would be the consequence if I directly bolt together the control arm to chassis?

I am assuming, firstly, the connection point would be un-damped so the ride comfort will feel more jolted.
However I don't see why this would be the case. After-all, all displacement, jolts and stresses are eventually transmitted to the spring damper-system that dissipates all shocks. Having said that,why not hammer in a close tolerance pin connecting the two. There would be no play between components whatsoever. Also, some of these sophisticated bushings do not deform nearly as much as old rubber bushings so the dampening effect is not there anymore. Ultimately, how justified is the dampening effect of bushings?

I also assume secondly, the other reason bushings are used is due to their classical definition i.e. A plain simple bearing that consists of a shaft and a journal (smooth hole). i.e. The control arm is constantly rotating with respect to the chassis mounting bracket and the bolt connecting the control arm to the chassis. However this rotation never does full circle and is at low speeds. Because of this, I cant see why a single pin that's greased can be used at the connection point instead of separate bushings.

Cheers

 

[Tsiolkovsky]

A double-wishbone suspension with the top arm shorter and / or the chassis-side A-arm pivots closer together (in the top-to-bottom direction) than the ball joints, will pull the top of the wheel inward as the suspension compresses. If that suspension compression is because of body roll, the body roll is carrying the top of the wheel outward. Superimpose the two, and they (more or less, depending on the geometry) offset each other.

Got it. However we must pay special attention to nomenclature. "Rolling moment" refers to the rotation of the sprung mass and the "non-rolling moment" refers to the rotation of the automobile as a whole: sprung and unsprung masses. Just to clarify: The effect of camber gain is offset by the the non-rolling moment of the vehicle.

I am assuming that's the reason SLA suspension is preferred over parallel-equal length wishbone configuration: The former has the property of camber gain/loss thus offsetting the non rolling moment thus keeping its original static camber angle through corners. The former, however, does not have the property of camber gain, thus the non-rolling moment cannot be offsetted by camber gain/loss. Correct?

 

[BrianPetersen]

Yes, but as usual, it's not that simple.

Front suspensions on mass-market production vehicles designed to be driven by ordinary people are deliberately designed to lose grip at the front first. A good many front MacPherson suspensions have very little camber change in compression. Often the rear suspensions on the same vehicles have a double-wishbone or multilink arrangement with shorter non-parallel upper links to get camber change in compression, so as to have lots of rear grip. Can't be having oversteer in a production vehicle these days.

Also, as mentioned by someone else, the caster and steering axis inclination angles have an influence on the camber when the steering is turned to the side.

 

[Warpspeed]

What I would be doing is first to define what type of racing this concept car is supposed to be competing in.

Then read the rule book though for the class rules very very carefully.

Then look very closely at the cars currently winning in that type of racing.

And beating the very best in the business at their own game, is not going to be all that easy.
But the winning cars define the current state of the art, and you could do far worse than use them as a case study, and a basis for YOUR design.

 

[kpt4321]

On one of the SAE design competitions (Baja), we used a semi-trailing rear suspension with pivots constructed of steel pins bushed with UHMWPE or Delrin. The pins, I believe, were hardened shoulder-bolt types, which we felt were more robust and easier to service than a pin with circlips. This is not a setup I would use on a production/street car, but for a race car it works very well. Low friction, easy to service, and replaceable bushings. Care must be taken that multiple pivots on the same member are coaxial, etc. On this same car, we used a similar setup on the front (unequal-length A-arm suspension), due to perceived higher level of robustness in the baja race environment relative to street/road course/auto-x, where rod ends are more common.

In other locations (front control arms on Formula SAE, for example), we have (many times) used rod end bushings (spherical bushings with a threaded rod extending from one end. They largely work well, but have problems when loads are applied in the wrong direction.

All this being said, if your current primary concern is what type of bushing to use, or what type of material to use in your bushing, then you should either be well into your design, or you are approaching the design problems in by far the wrong order. The first tasks should be to determine general types of suspension to use on the front and rear of the vehicle, and then determining approximately what behavior (camber gain, trail, total travel, natural frequency, spring/damper mounting) you want, and then looking at general geometry, and finally starting to look at actual structural designs.

Finally, as a previous poster already mentioned, it would behoove you to look at existing race cars in the class in which you wish to compete.

 

[isaacowen]

its also got a lot to do with the space that a mcpherson strut would take up in the boot of a sedan etc.
working on vw's all of them use a trailing arm style setup. id imagine its similar for other manufacturers?

 

[Tsiolkovsky]

Well its been almost 5 months since this topic started. And once again a new question arises got to do with camber gain. I can now see the qualitative properties of camber gain achieved through SLA suspension but I cannot seem to know how to design, or quantify it. How do I determine what values of camber gain I need per unit jounce? (angle/mm).

Here is my reasoning, correct me if I'm wrong:

1) It depends on the tire properties. An easier flexing carcass means the tread will conform more easily to the ground even with the smallest of lateral tire loads. This in turn means that if the tire carcass is "soft" (don't know the correct term for this), the camber gain must be more pronounced. If it wasn't more pronounced, the tire tread will "over" distort with a given lateral load through cornering.
2) It depends on the non-rolling moment of the automobile (which in turn is dependent on the position of CG WRT RC). If the non-rolling moment is high (the automobile-as-a-whole rolls easily), then you would want a higher camber gain to "offset" this disturbance in camber angle.

If the above is correct (there may indeed be even more factors), then how would you quantify and calculate it?
Am I correct in saying the answer is: You dont? The reason being is that its an experimentally deduced number. You set it to an arbitrary value, run the car on corners and then bring it in to measure the tire temperature along the tire tread. If its hotter in the inside, you have too much camber gain. If its hotter on the outside, you have the camber gain set to too low a value. This makes sense to me, I think I even read it in some literature (the experimental nature of determining camber gain). Adjusting the camber gain itself is of coarse only a function of SLA geometry so you could incrementally lengthen/shorten the bottom wishbone and then do another test run)

Ultimately, how true is the above?

 

[BrianPetersen]

From what I can tell, "you don't" is the correct answer.

We don't even know what sort of "race car" we are talking about here. World Rally cars have very compliant suspensions and with that and the nature of the rally stages, comes a lot of suspension travel during operation, and with *that*, comes not wanting too much camber change with suspension travel. Subaru cars use MacPherson front and rear, and the rears have long lower links - a recipe for not much camber change.

Open wheel cars designed for smooth pavement race tracks don't need much suspension travel and generally have to be designed with ground-effect aerodynamics in mind. With that, comes a need for not letting the ride height vary too much no matter what happens with the car. With that, comes very stiff spring and antiroll rates. With *that*, comes very little camber change regardless of how the suspension is designed, because it hardly moves. Formula 1 cars have very visible upper and lower A-arms, and they're visibly almost horizontal, suggesting minimal camber change with what little suspension movement that there is.

Ordinary road cars that are expected to have an acceptably smooth ride and still have decent cornering properties and have varying height due to cargo loading ... are a tricky application, but even then, there is a wide variety of suspension designs, particularly for the rear, some with no camber change (trailing arms), some with wheels perpendicular to the ground at all times (beam axle), some with a lot of camber change on bump compression (late model Honda Civic upper and lower arms), some with not much (MacPherson strut).

 

[Warpspeed]

If you try to design with only camber gain in mind, you will very likely end up with a fairly unfortunate roll centre location, and wheel track changes at the contact patches under suspension movement.

The whole thing is a rather complex process of evolution.
Choosing the correct compromise for the actual racing conditions is what wins races.

 

[NormPeterson]

2) It depends on the non-rolling moment of the automobile (which in turn is dependent on the position of CG WRT RC). If the non-rolling moment is high (the automobile-as-a-whole rolls easily),

I think this statement ??? is causing some confusion, as relatively high non-rolling moments imply rather high geometric roll centers and relatively little roll. Try thinking in terms of lateral load carried through the roll centers instead, and resolve that into changes in wheel loading. FWIW, there is a little non-geometric camber change here due to tire vertical stiffness effects, maybe 1°/g.

then you would want a higher camber gain to "offset" this disturbance in camber angle.

Don't overlook the static camber setting as a means of adjusting where the camber under operation ends up. Hopefully we aren't looking at extremely large camber changes.


Norm

 

[Tsiolkovsky]

Everyone on this thread has been 100% correct in that methodology of suspension design is critical. Literature quotes that the best design process is as follows (I summarised in point form):


Front hub height
Rear hub height
Wheelbase
Track
Front RC
Rear RC
Type of Front Suspension
Type of Rear Suspension
Relative position of outboard pivots (upright/knuckle control arm pickup)
wheel travel
suspension unit travel
ROUGH Geometry of control arms (length, angle, position, pickup outboard & inboard)
Undertray Line

Pause: Feed this information, and the following information to chassis designer
return any input from chassis designer
Locations of all internal components
Bulkheads Main/subsidiary structural members
General traverse members


more detailed susp Design Begins:


Plot Wheel Limits of Travel
Plot roll centers during bump/droop
Plot camber change during bump/droop
Move wishbone pivots to constantly control this
Fix maximum desirable roll angle
Hence wheel inclination at maximum roll
Virtual swing arm length (rule of thumb/baseline: 1.5-3 X track)
Static camber angle agjustments

Through RC & tire contact & outboard pickups, control arm front view angles are found
Inboard pickups
again, Roll centre height change on bump-droop
Keep roll centre movement to minimum without keeping camber gain too low
Detailed factors calculations: Determination of Roll Angles and Wheel Loading (by calculating the roll moment)
Inboard pickups (to be fed directly into chassis)

Mountings for susp units
before wishbone pickups, determine Type of mountings/bearings
Steering linkages (anti ackerman and toe in/out under bump/droop)
Castor/KPI angles
Anti features and other non standard features

Repeat and constantly increment all steps in order and feed all these factors into each other until good compromise values are found