The Mark Ortiz Automotive
Reproduction for sale subject to restrictions. Please inquire for details.
Mark Ortiz Automotive is a chassis consulting service primarily serving oval track and road racers. This newsletter is a free service intended to benefit racers and enthusiasts by offering useful insights into chassis engineering and answers to questions. Readers may mail questions to: 155 Wankel Dr., Kannapolis, NC 28083-8200; submit questions by phone at 704-933-8876; or submit questions by e-mail to: firstname.lastname@example.org. Readers are invited to subscribe to this newsletter by e-mail. Just e-mail me and request to be added to the list.
SUSPENSION DESIGN PROCESS
How exactly do you design a suspension system? I mean procedurally, what are the steps? What comes before what?
There’s a good History Channel documentary about the WWII AVRO Lancaster bomber that a number of people have uploaded to YouTube (https://www.youtube.com/watch?v=fmcI3-dPBkY). The lead designer was Roy Chadwick. The documentary features comments by Chadwick’s daughter, Margaret Dove, who wrote a biography of her father. She recounts asking him, “How do you design an aeroplane, daddy?” Chadwick replied, “Well, I think about all the parts first, and then I sit down to do the design.”
That’s a bit simplistic of course, but the important takeaway is that a vehicle has to be designed as an integrated whole. Whether a single person is designing it, or a huge team, that integration has to happen somehow. Linear or incremental thinking doesn’t produce good outcomes. Designing one piece or subsystem at a time doesn’t produce good outcomes. Somebody, somehow, has to be able to think about everything at once, and envision and coordinate the whole thing.
And it isn’t merely all the parts that have to be thought about at once; it’s necessary to think about all the design objectives and design constraints at once, together with all the parts.
The suspension design process is one part of this overall process.
That’s if we’re designing a whole car. Sometimes we’re designing a suspension for an existing car, or just modifying an existing suspension system. But even then, the suspension is constrained by the rest of the car. The only difference is that we are limited as to what we can change in the rest of the car. When we are designing the whole car, it is possible to change other parts of the car, or negotiate with those designing the other systems, when we need real estate for our system, and we may even be able to reconsider or haggle about design objectives to some extent.
That said, it is possible to say that in most cases certain things should be determined before certain other things, but always with the caveat that everything leans on everything else, and it is very common to design something and then have to change it to accommodate something else.
Common advice to students doing Formula SAE/Formula Student is that the first step in designing the car is to read the rules; the second step is to read them again. The second part of that can be taken to mean that the rules are complex enough so you will need to study them at some length. However, it also means that you need to read them again once you’ve got things designed, and make sure your design is in compliance. Really, you need to do that repeatedly as you go.
There are rules for every kind of car. Compliance is vital for every car. It does no good to build a car that works well and then be prohibited from using it due to some technicality.
Regardless of the particulars of the project, the first step is always to define design objectives and constraints. If you are designing a car for yourself, you may not need to deliberate much about design objectives at all. You probably know what you like and what you’re going to do with the car. On the other hand, for mass-produced cars, establishing design objectives can involve months or years of market research and deliberation.
Early on, we need to think about aerodynamics. At a minimum, we need to know whether the car is going to have a lot of aerodynamic downforce, and whether this will be sensitive to under-car aero.
Once the constraints are established in general and legal terms, it’s time to define them spatially, on the screen or on paper. The known components need to be modeled or drawn, and laid in, along with any established dimensional limits. We will need to have an approximate idea what the total weight and the weight distribution are going to be, based on the overall layout, the type of construction, and the design payload.
If a decision on tires has not been made, that generally comes next. We don’t necessarily need a behavioral model of the tire. We just need to know what size we’re using. Having tire behavior data is nice, but what we absolutely must have is dimensional data. In many cases a car will run on a variety of tires throughout its production life or service life. We need to consider all the possibilities, and try to make the car work at least reasonably well with all tire packages that can foreseeably be fitted.
So the next step is to lay in a reasonably representative wheel and tire package. We must remember to allow roughly half an inch deflection where the tire meets the road.
We probably will be using existing brake components. We need to select brakes that provide appropriate front/rear bias, have the best possible fade resistance and power, and meet our cost and availability requirements. We lay those in, making sure everything clears the wheels and making sure the bleed screws are at the highest point in the calipers.
As we position the calipers, we want to be thinking about what the rest of the system is going to look like, and try to avoid putting the calipers where something else will need to be, such as the steering arms. There is some merit to placing the calipers above and behind the axle, as this minimizes wheel
bearing loads under braking. However, this is not crucial. Other considerations may lead us to put the calipers elsewhere. We may even decide to use inboard brakes. If we do, this affects not only the packaging at the wheels but also the way the brake torque is reacted. This in turn affects the suspension system’s jacking coefficients – its anti-dive/anti-lift properties – for any given geometry.
We need to think about what kind of steering the car is going to have, and where the rack or box is going to go. This will involve packaging the mechanism, as well as the steering shaft, within the
front end of the car. The location of the steering gear largely determines where the tie rods and the steering arms will have to go, and that will affect caliper positioning and upright design.
For the front suspension we will need to establish where we want the steering axis. I have written about this in detail in the fairly recent past, so I will not dwell on it here.
We may have freedom to select the general type of suspension, or this may be established during the product planning phase. The situation may be somewhere in between: suspension type may be tentatively selected; product planning and design can overlap; chassis engineering people may be included in product planning.
It is generally desirable to use carryover components from previous cars where possible. This saves money and design and development time. If we are using any such components, we add those.
If we are designing an independent front suspension system, my general recommendation is to start with desired force lines. These are the lines that run from the contact patch center of each wheel to the front and side view instant centers. We would like these to slope up toward the center of the car, at a slope of between approximately 1 in 50 to 1 in 6, or an angle of between 1 and 10 degrees. This means that as the suspension displaces, the contact patch center moves horizontally at between 1/50 and 1/6 times the rate that it moves vertically, and the system induces between .02 and .17 pounds of jacking force in the suspension for each pound of lateral or longitudinal force at the contact patch.
We then position desired front and side view instant centers on these lines. How far these are from the contact patch center determines the front and side view swing arm lengths, and the rate of camber and caster change with respect to suspension displacement. I generally recommend trying to put these no closer than the opposite wheel, and no farther away than twice that distance.
We now have two points that lie in both control arm planes. Three points determine a plane. In the case of a short and long arm (SLA) front suspension, the upper and lower ball joint centers of rotation then determine the upper and lower control arm planes. Usually, it is best to put the ball joints as far apart as possible. This reduces loads on the ball joints and control arms, although there is some penalty in upright weight. We will be limited by the need to have the wheel rims clear the
control arms at all combinations of suspension and steering displacement, and by the need to maintain adequate ground clearance. At all four wheels, we need to make sure nothing near the wheel hangs below the wheel rims. Even in a race car where the frame or tub hangs below the wheel rims, it is best to keep the uprights above the bottom of the wheel rims. All suspension components must clear the ground even when the tire goes flat.
Once we have the control arm planes tentatively established, we need to decide where to put the inner control arm pivot axes, which are defined by the bushings or spherical joints at the inboard ends of the control arms. These need to be in the control arm planes that we have established. We may find this is easier said than done. We need to make the control arms as long as possible, and we need to have a good upper to lower arm length ratio. As a general rule of thumb, the upper arm should be between the same length as the lower and 2/3 as long. We also need to have no interference between any of our parts and all the other things in the front end of the car. On top of that, we need to make sure that the tie rod length harmonizes with the control arm lengths, so we don’t get excessive bump steer. And on top of that, we need to have good load paths: the control arm mounting points need to feed the loads into the sprung structure in a manner that minimizes deflections, both local and global.
We may very well find that we need to go back and re-think the ball joint locations or compromise on the instant center locations. If we have to compromise on geometry, front-view geometry is more important than side-view.
We additionally need to think about what adjustments need to be provided, and how mechanics are going to work on the system.
A strut suspension is much like an SLA suspension, except that the upper control arm plane includes the center of rotation at the top of the strut and is perpendicular to the strut axis. A system with links in all sorts of different planes may not have identifiable control arm planes. It may only be possible to identify its kinematic properties by computer modeling. The rules for the relationship between jacking coefficients and lateral versus vertical movement at the contact patch will still apply. However, with any system the rules for side-view geometry change when brake or drive torque is not reacted through the linkage. The rules for beam axles are completely different.
Is it starting to become apparent why this cannot be reduced to a step-by-step process? Why incremental thinking leads to poor outcomes? Why Roy Chadwick thought about all the parts before he sat down to do the design?