Why Automakers Still Worship This 30-Year-Old BMW’s Suspension

To build the best product, you have to set technical goals. Those goals are based on a number of factors, including what the competition is doing and also what the “ideal” is. For many suspension engineers, the “ideal” has always been the BMW E39 5 Series. Here’s why it’s so good that it’s been studied under suspension engineers’ microscopes for over 30 years.

Let’s first talk about why suspension engineers would bother looking at another company’s cars in the first place. It all begins with defining the customer of your new vehicle program.

Defining The Customer

When a car company starts to develop a new car, one of the first things it does is define who is likely to buy that car. The company needs to know who the customer is so they can target their product to the right buyers.

Vast storyboards are created showing the lifestyle these customers live, their income, where they live, what activities they like to participate in. Are they single or do they have a family? How big is the family? Do they drive lots of kids and their friends to soccer practice? Do they take their car to track days? Are they active? Are they retired? Are they into natural foods or do they go to fast food places? Many of these things have nothing to do with cars specifically, but they give a broad picture of the type of person a company intends to attract with its product.

Here’s an example for the Subaru Crosstrek; its customers are described simply as “Youthful Explorers”:

Knowing your customers also means knowing what other cars they are likely to buy or consider. You have to ask yourself the question: “If our car were out today, what other cars would our customers cross-shop and what would influence their purchasing decisions?” From this discussion comes a list of potential competitive vehicles your customers may buy instead of yours.

The Competition

Naturally, the market segment your car will enter has a great influence on the customer and the competitive vehicles you choose. If you are planning to compete with the Toyota Corolla, your storyboards and vehicle list will be completely different than if you plan to compete with a Mercedes-Maybach.

As the planning process is taking place, the company will begin forming the various engineering teams that will be needed to design the new product. A management structure will be created to lead the project with a single person — a chief engineer — at the top who is responsible for delivering the final product. Every company has their own way of doing this, but the result is a group of engineers and managers in each of the major functional areas assigned to work on the new project: Body systems, Chassis systems, Interior, Exterior, Powertrain, Electrical, etc. as well as the various attribute areas: Vehicle Dynamics, Safety, Noise Vibration and Harshness (NVH), Ergonomics, Package, Regulatory Compliance, etc.

[Editor’s Note: At Chrysler, engineering was broken up into what was called BICEEPR: Body, Interior, Chassis, Electrical, Engine Systems, Powertrain, Restraints. -DT]

Each team will start out small — only a few people at most — and grow over time as the work progresses.

The work of creating the storyboards and defining the competitive vehicles is usually done by the sales and marketing teams in conjunction with the styling/design studio. There is little, if any, engineering input into this part of the process. However, that changes once a list of potential competitive vehicles has been identified. The program management team will go out and buy the top three to four competitive vehicles and the engineering teams will spend the next few months thoroughly analyzing them, measuring them, driving them, and figuring out why they work as well as they do. These vehicles will remain with the team for the duration of the program to be used as a reference as the design and eventually prototype testing develops.

Some OEMs will also take this time to determine on what level they would like to compete in the marketplace. This will depend on the type of manufacturer it is and how they have placed themselves in the market. If the OEM has built a reputation for engineering excellence, they may want to be the leader in vehicle dynamics, ride, or durability. If the OEM has instead built a reputation for value, then they may want to make sure they deliver a product at the lowest price while still offering acceptable ride/handling/durability attributes. Some OEMs may place more value on safety, demanding that their new product achieves five stars in all crash tests even if that requires a sacrifice to ride/handling/vehicle dynamics. It really depends on the OEM, how they have built their reputation, and what they perceive is important (and not important) to their customers.

Benchmarking

The process of understanding how good your competitive vehicles are is called Benchmarking. Benchmarking takes many forms depending on what you are trying to understand (ex: Munro and Associates tears cars down to find out how they’re built and how much it costs to build them), but the most basic form is to simply go out and drive the cars. This can be done on an individual basis or as a group.

I have spent many days in my career on benchmarking drives where we gather a number of competitive vehicles, go out to a remote location that has a variety of roads, and drive the cars. Someone would have gone out several days or weeks before and mapped out a route that would include some highway sections, twisty roads, rough pavement, and opportunities to do some hard braking and acceleration. Each driver would rotate through all the cars and rate them on many different attributes: ride, handling, interior noise, seat comfort, ergonomics, power, brake feel, steering feel, etc. There would be several dozen categories and sub-categories, and you would be expected to provide your opinion in the form of a rating on each one. (Side note here: We would always buy competitive vehicles in grey or silver, never a bright color like red. Red cars always rate better and can skew the results. Engineers try hard to be impartial, but the reality is that we still have our biases. We like red cars.)

The ratings scales would depend on what is being evaluated. Most often, a scale from 1-10 is used where 1 is unacceptably bad and 10 is unbelievably good. The Society of Automotive Engineers developed such a scale as part of its J1060 “Subjective Rating Scale for Evaluation of Noise and Ride Comfort Characteristics Related to Motor Vehicle Tires” standard which I have used in all the OEMs I have worked for:

While this scale goes from 1-10, the reality is that we would never give anything a rating below five or above nine. Nothing is so good that it deserves a 10, and nothing is so bad that it deserves a rating below five. After all, the cars we are driving are all being sold so they are clearly not “unacceptable.” Plus, we weren’t benchmarking the worst cars in the class.

Such a scale doesn’t work in all cases though since vehicle attributes are so subjective. If you are rating brake pedal effort, for example, one driver may prefer a very light pedal while another prefers a heavy pedal. Using a 1-10 rating scale would not be very useful in that case. In those instances, we would use a scale from one to five where one might be very light and five might be very heavy. This way, personal preferences are captured and can be reflected in the data.

Usually, we would use a team of about ten engineers driving five cars, and the process would take a full day or even several days to complete. The results would then be tabulated and averaged, and the responsible team would write a report with the results.

Measurements

Another form of benchmarking is objective measurements. Here, the cars are run on a variety of roads and measured using various instrumentation. The NVH teams, for instance, will place a recording device and accelerometers inside and throughout the vehicle to record the noise and vibrations while driving over known rough surfaces, often at a test track.

The vehicle dynamics team will have a highly trained driver take the car on a very large asphalt surface called a VDA (Vehicle Dynamics Area or “Asphalt Lake”) through a variety of maneuvers like lane changes, step steer events (where you violently turn the wheel 90 degrees and hold it), max braking events, etc.

Some OEMs may even use driving robots to do these tests so that each car has exactly the same inputs every time.

A specially designed handling course will be used to drive the cars at their limit to see how easy the car is to control and how well “ADAS” electronic systems like stability control function.

During these tests, the responses of the cars are measured with accelerometers, GPS locators, height sensors, and other instrumentation to see how each car performed. The results are then tabulated, plotted, and compared against each other.

Some measurements may require highly specialized machines. Suspension engineers, for instance, will use a machine called a Kinematics and Compliance machine to push and pull on the suspension. This video shows one of these machines in action:

From these measurements, engineers can tell how the camber, toe, and caster change when the suspension is moved up and down, steered, or gets pushed and pulled like it would be during cornering or braking events. This data is invaluable to understanding why a particular car rides and handles the way it does.

Once all the data from subjective drives and objective tests is collected, engineers use it to decide how well they want their new product to function based on in what areas the company wants the product to be a leader (or just be competitive).

Futuring

Of course, deciding to be competitive or a leader compared to what is out there today will never be good enough. Benchmarking by definition is really just looking into the past. The vehicles you can buy today are the results of those OEMs using this same process four to five years ago. Those same OEMs are not standing still, and when your new car goes into production four to five years from now, those cars will have improved. So, if you are just competitive with what is out there today, you will be uncompetitive once your product enters production.

The trick is to look at the improvements that have happened over the past decades and extrapolate them into the future: a process called “futuring.” It sounds simple, but it is extremely hard, and not many people are good at it. The process also by definition looks at historical trends and cannot account for innovations that can completely change the marketplace. No one in 1979 knew that the world’s major producers of oil would reduce their production and cause a worldwide crisis that changed many customers’ buying decisions away from large gas hogs and towards smaller more fuel-efficient cars. No one in 1990 knew that Toyota was going to come out with its Lexus brand and rewrite the rules of what a luxury car could be and more importantly, what the whole customer experience could be. No one in 2011 knew that Tesla was going to come out with an EV that would make a typical ICE car feel like an old dinosaur.

No one can future against these sorts of events, but fortunately for those doing the benchmarking, they are rare. For the most part, the auto industry improves in small increments, and it is possible to look into a crystal ball and get some idea of where things will be four to five years from now. You can then use this information to decide how much better your car will need to be to remain competitive or still be a leader once it goes on sale. It goes without saying that knowing how good you want to be still requires translation into the hardware that will achieve those targets, and this is done through a process called a “target cascade.”

Target Cascade

Let’s look at an example of a target cascade. Suppose we were developing a car and the characteristics we were most interested in were Fuel Economy, Value, Performance, Safety, Durability, and Package. We’ve identified three competitive vehicles — A, B, and C — which we have bought and tested. After completing our benchmarking of competitive vehicles, we could summarize the results in the form of a spider chart (sometimes also called a radar chart) like this:

We can easily see how each competitive vehicle compares to the others for each of the characteristics we’ve chosen to focus on. In addition, we can add where we want our own car to be in relation to those vehicles. Of course, these are purely fictitious numbers. Expecting our new car to achieve a rating of nine for everything is probably unrealistic, but you get the idea.

We can now focus on each area individually. If we looked at Performance specifically, we might break it down into 0-60 time, Braking, Handling, Steering, and Ride like this:

We could then drill down even further. Looking at Steering, we could break it down into On-Center Feel, Effort, Linearity, Vehicle Response, and Returnability like this:

Going even further, we could look at On-Center Feel and break it down into Steering Gear Ratio, Pull-through Friction, Intermediate Shaft Torsional Stiffness, and Gain. We are now getting to the point where we are talking about characteristics to which we could potentially assign an objective, measurable number — one that we could provide to a supplier or engineering team to use in the design of a component. And that’s what we ideally want to get to: numbers that can be used to guide the design of the components of the car.

Once the design is completed, these numbers, and the cascade that got us there, create a direct way to evaluate how well we did. When computer models and later prototype vehicles are built, their performance can be compared with the targets, and we can decide if we think we’ve achieved our top-level goals.

The Best Car

Having been through this process many times during my 32 years in the auto industry, I have driven competitive vehicles of all types: small sedans, large sedans, high-end sports cars, SUV’s, pickup trucks, top-end luxury cars, and on and on. From cheap Corolla competitors to Rolls-Royces, I’ve had the opportunity to drive it all. But the one car that has consistently stood out to me time and time again was the E39 BMW 5-series.

I know for a fact I am not alone in feeling this way. During my career, many engineers, especially vehicle dynamics engineers, owned these cars. In many cases, the competitive vehicle sets for the new car programs I worked on included the E39 or its predecessor. In fact, the incredibly advanced, all-electric Lucid Air, which launched for the 2022 model year, used the nearly 30-year-old car as a benchmark, per Motor Trend. From the car site:

As a Lotus alum, Lucid Motors CEO and CTO Peter Rawlinson wanted his Lucid Air electric luxury sedan to at least invoke the feeling of nimbleness and linearity of responses to control inputs without latency or phase lags that the iconic Lotus cars are known for.

[…]

Obvious benchmark vehicles for the Lucid Air’s dynamic handling included Porsche’s Panamera and Taycan, while more attainable benchmark cars Lucid purchased for study were a Chevy SS and an E39 BMW 5 Series. Each of these vehicles features modest levels of ultimate grip but feel very playful as they approach them.

When the programs I worked on tested the E39 5 Series, the vehicle consistently rated at the top of almost all functional categories, and really became the benchmark we all tried to achieve. To this day, I have not driven a car that has such a perfect combination of ride, handling, comfort and utility as the E39. The car was exceptional in twisty backroads but was also relaxing to drive on the highway. The faster you drove, the easier the car was to control.

Making a suspension great is a combination of the right tires, springs, damper settings, and bushing tuning. It’s a combination of many small details, There is no magic bullet that makes a suspension great. Of course, you need a good suspension geometry first but good tuning on the tires, springs, and dampers can make up for a lot. As an example, for decades Porsche used a very simple MacPherson strut in the front of the 911, as did BMW with the 3 and 5-series. Simple but very well executed. At that time, BMW were masters at finding the right combination of these factors and the designers of the E39 managed to find that perfect balance.

Most other cars that were fun in the twisties were darty or nervous on the highway. Not the E39. On top of that, you could put five people and their luggage in it. I never did entirely figure out how BMW managed to achieve all of that, but I’m glad they did. But if there’s one obvious piece, one little bit of the secret sauce that makes the E39 just feel right it’s something called an “Integral Link” and it is a unique design different from any other suspension.

It uses a small vertical link in the rear suspension between the lower arm and the knuckle to control wind-up during braking. The whole design is very compact. If you watch the video above you’ll see how it works and where it is. The downside is that it doesn’t allow for very much rear steer which is why you see some companies moving away from

I feel that no one, not even BMW, has bested this car in the years since. All the electronic gizmos and bells and whistles available today cannot make up for the fact that the E39 was so fundamentally good that it didn’t need those devices to be a great car.

It didn’t need infinitely adjustable shocks, or dynamic stability control, or cameras that looked at the road ahead, or lane keeping assist, or extreme low-profile tires, or any other toy to be an absolute joy to drive. BMW nailed the fundamentals of what makes a car great with the E39, and the result was a car that was at the top of every engineer’s wish list. I have spoken to and worked with many engineers from other OEMs, and my feelings are universally shared.

So, if you plan on starting your own car company or are working for an OEM starting a new car project, you can go through a massive benchmarking activity that will cost you millions. Or… you can go out and buy an E39 BMW for a few grand and be done with it.