|
Bumping along smartly
|
01/04/2007
|
| |
The ideals of skyhooks and groundhooks are not achievable, but this hasn’t deterred engineers seeking to improve ride comfort and handling. Jeff Daniels looks at the challenges
The dream of every vehicle suspension engineer – indeed, of most vehicle occupants – is a car body that crosses any kind of terrain without responding to the surface beneath. Complete isolation from shock is the ideal, leading to the concept of the ‘skyhook’ which is the starting point for so many advanced suspension studies.
The skyhook is now a familiar idea, and with familiarity has come the acceptance that it represents an ideal that cannot be achieved, since it calls for infinite actuator input and zero response time. The challenge for suspension engineers has been to derive systems and control algorithms that approach the ideal as closely as possible, at moderate cost and with low parasitic power consumption. Thus far, most attempts have failed so far as series production vehicles are concerned.
A less familiar idea than that of the skyhook is the ‘groundhook’. Just as the ideal body motion is zero, regardless of terrain, so the ideal where the wheels are concerned is to keep them permanently in contact with the ground, thus ensuring optimum roadholding and traction. Attempting to approach the skyhook and groundhook ideals simultaneously is not for the faint-hearted, even in theory.
The real challenge is that the engineer has to make the most of what is available, to emulate the (clearly impossible) ideal in which there are actuators above the body and beneath the wheels. Since a conventional, ‘passive’ suspension consists of a spring and a damper, the options are to modify the characteristics of either or both, or to supplement or supplant them with a positively responding actuator.
Varying the rate (load per mm displacement) of a mechanical spring in response to a vehicle-generated parameter, such as speed of wheel movement, is difficult. However, many coil springs are designed to provide a variable rate as they compress, using coil geometry. As an alternative, auxiliary elastomer springs can stiffen the total spring rate towards extremes of travel. Both techniques are now commonplace. In practice, dynamic control of spring rate can only be achieved with hydropneumatic systems (as used by Citroen) or with air springs, increasingly used in the luxury car segment.
By contrast, dynamic variation of damper rate (force per unit wheel speed) in a conventional hydraulic damper is easily achieved. To alter damper rate, it is necessary only to vary the size of the orifice through which the hydraulic fluid is forced, and various methods of doing so have been developed.
The challenge here is to achieve continuous variation rather than switching between three or four distinct settings. This is the purpose of Delphi’s MagneRide principle, several years in development but now finding applications in several GM models and in Europe, starting with Audi and Ferrari. MagneRide uses a magneto-rheological fluid, and varies the damper rate by altering the strength of the magnetic field across the orifice by passing a variable control current through adjacent coils.
Variation of either spring or damper rate yields either ‘adaptive’ or ‘semi-active’ ride control, the distinction between the two being the speed of response. Where the response is slower than the natural frequency of the body, say less than 1Hz, the system is adaptive, meaning it is capable of setting basic parameters such as ride height and body levelling in pitch. Where the response is sufficiently fast, by implication comfortably in excess of the body frequency, at least 10Hz and probably more, the system is capable of applying dynamic correction – for example, varying damper rate to improve ride comfort – and it is classified as semi-active.
In its latest form, MagneRide has an operating frequency of around 100Hz; the fluid itself changes state in nanoseconds, fast enough to be regarded as instantaneous in relation to the data sampling and decision-making frequency of the system as a whole.
One problem encountered by MagneRide, but in principle applicable to any semi-active variable damping system, is that of force transmission. The effect of stiffening the rate of a damper in mid-travel can be considerable. Delphi engineers say that given the response rate of the MagneRide system, an additional force of 500N can be applied in a fraction of a second, to be transmitted into the damper mounting. This is an additional engineering constraint on the design of body structure for any vehicle equipped with variable damping.
In any conventional suspension, as the wheel moves in compression or extension, the spring and damper both create an opposing force – the spring in opposition to the amount of movement, and the damper to the speed of movement. If one draws a diagram of the force acting on the wheel versus wheel displacement or speed, adaptive and semi-active systems (like passive systems) can only operate in two of the diagram’s four quadrants, those in which the force and the deflection (or speed of movement) are in the opposite sense (the force arising from, and opposing, compression or extension).
In effect, a passive system traces a single path through the origin from one quadrant to the other, while a semi-active system can only change the slope of the path.
If you wish to enter the other quadrants – if, for example, you want to push down even harder on a descending wheel to make it follow a sharp downward ramp – you need to fit a positive-displacement actuator which propels the wheel downwards. In effect the actuator ‘injects’ additional energy to make the wheel move faster.
Any system with positively responding actuators is ‘active’. In practice, all the actuators fitted to demonstrated prototypes have been hydraulically driven. They enable the system to approach the skyhook ideal far more closely, and to perform other functions.
Early active-ride demonstrators from Lotus and Volvo, for example, could be made to adopt any roll angle (subject to the limits of suspension movement) when cornering, and to ‘buck and heave’ when at a standstill, by way of illustrating what high-powered actuators could achieve. The problems with all such systems, then and since, have been cost and power consumption. Estimates of the power needed for a ‘near-skyhook’ ride on a poor surface ranged up to 15kW depending on speed, while suitable actuators were only available from the aerospace sector, and priced accordingly.
Eventually it became clear that such a ‘full house’ system would probably never be offered, not only by virtue of its inherent drawbacks but also because conventional systems became more satisfactory, and because pragmatic engineers began to discover alternative approaches which conferred major benefits at a fraction of the cost and power consumption.
Citroën’s chassis engineers, their larger cars already provided with most of the means in the form of a hydraulically actuated hydropneumatic suspension, were first in the field. They had been providing some elements of adaptive control – self-levelling and adjustable ride height – since the announcement of the DS during the 1950s. During the 1990s they added some semi-active features, under the names Hydractive and Activa.
The first required no more than the addition of a few components to the existing system and the development of a clever control strategy. Activa, which added active control of roll angle, needed additional actuators. For their control strategy, Citroen adopted a predictive approach based on the sensing of steering, braking and acceleration inputs rather than direct sensing of wheel position. In the event, Hydractive survived and evolved but Activa fell victim to a policy of system simplification when the C5 replaced the Xantia.
This was one approach, but open only to a team that already enjoyed the benefit of hydropneumatic suspension. Short of a wholesale change in technology (and significant extra cost) it was not an option for anyone else.
Mercedes adopted a very different approach, reasoning that some frequency ranges are more crucial than others when it comes to ride comfort. Many parts of the human body resonate within the range 2.5-6Hz, with notable adverse effects on the neck and lumbar vertebrae and the trunk. It is a less demanding and ambitious target to minimize body vertical movement in this range only, allowing a well designed passive suspension to deal with the rest of the frequency spectrum.
This was one of the considerations behind Mercedes’ development of active body control (ABC), which combines an active module of limited movement and authority, integral with an otherwise conventional (but carefully calibrated) spring and damper strut. In this way it reduced the power consumption, and to some extent the cost, of the system.
Mercedes has continued to develop ABC and to apply it to a wider range of upmarket models. A ‘second generation’ ABC is fitted to the recently introduced CL-class, using 200 bar hydraulics and with a system frequency (determined by data sampling interval) of 100Hz. The system is calibrated to improve the ride at frequencies of around 5Hz and also now controls roll and pitch, with the additional function of being able to vary the ratio of front to rear roll stiffness to influence vehicle handling (more front roll stiffness creates more understeer and vice versa).
Thus far no other manufacturer has followed Mercedes down the active ride path. Its luxury-class rivals have preferred to move towards air-spring systems and semi-active damping.
A strong trend is the move towards such systems becoming ‘adaptive’ in the sense of changing the shape of the response curve according to quality of road surface or manner of driving, thus holding a firm setting for longer at high speed, or a comfort setting if the road surface is consistently poor.
Beyond this, the semi-active and active principles have been embraced via alternative routes. Two of the strands that came together were active roll control (ARC) and electronic stability enhancement (ESP). The ability to control roll stiffness became more of an issue with the increasing popularity of the SUV, in which two issues arose. One was the need to allow maximum wheel articulation in offroad use, while limiting it in faster driving on metalled surfaces. Allied to this was the concern about SUVs tipping over when cornering fast. A high-built SUV will always tip sooner than a low-built car in response to lateral acceleration (in fact, a low-built car is unlikely to tip at all unless it is provoked, possibly by running over a significant obstacle). However, if an SUV is ‘soft’ in roll its tendency to tip will be exacerbated, both by a rise in its centre of gravity and by the inertia of its body assuming a severe roll angle then hitting its bump-stops.
The answer was to fit the vehicle with anti-roll bars which could be disconnected, side-to-side, by some form of locking actuator, rather in the same manner as a lockable differential. The suspension would therefore be soft with the bar released, but much firmer with the lock engaged. It meant among other things that depending on the chosen calibration, the basic spring rate could be made softer, and the anti-roll bar stiffer, than would normally be the case: a limit to anti-roll bar stiffness is normally set by sideways head-level acceleration when the wheels on one side of the vehicle pass over a bump.
Such systems have been offered to OEMs since the late 1990s by most of the Tier One suppliers with an interest in suspension systems. TRW, for example, developed its SARC (semi-active roll control) system capable of accepting input from steering wheel angle, lateral acceleration or vehicle speed sensors. This was only a start, however, since the engineers concerned soon extended their efforts, replacing the simple anti-roll bar locks with variable actuators. This enables the systems to provide positive anti-roll operation, delaying the onset of significant roll until a threshold value of lateral acceleration has been exceeded.
So long as there are anti-roll bars at each end of the vehicle, the handling can be modified by differential operation of the two actuators. This is the principle behind TRW’s ADC (Active Dynamic Control) and other essentially similar systems. It is an approach which offers a possible alternative, or complement to, the widely fitted ESP stability augmentation systems which depend on brief, differential brake operation, further expanding the range of options open to chassis engineers.
While it remains to be seen where the principle of active ride goes next, it does appear that engineers are concentrating on addressing specific ride comfort issues by tackling particular aspects of vehicle body movement, rather than seeking an overall skyhook solution. Weight, complication and above all cost remain significant constraints.
Agility for Mercedes
Mercedes-Benz has developed two new suspension systems for its new C-Class range with the aim of combining a smoother ride with enhanced grip and body control, writes Ian Adcock.
Fitted as standard equipment across the range, Agility Control is based on an amplitude damping system. During normal driving conditions, with low shock absorber impulses, the damping forces are automatically reduced to improve ride comfort but without compromising handling. As shock absorber impulses increase during cornering or lane change manoeuvres maximum damping forces are set to stabilise the car.
The system is purely hydro-mechanical and uses a bypass channel in the shock absorber’s crank pin and a control piston moving within a separate oil chamber. When the shock absorber impulses are low the control piston forces oil through the bypass channel creating a significantly smaller damping force at the damper valve resulting in a softer shock absorber for improved ride quality. No complex sensors or electronics are required.
Subjecting the shock absorber to larger impulses results in the control piston moving to its final position, preventing oil from flowing through the bypass channel. This reduces body roll by up to 10% without any loss in ride comfort, says Mercedes-Benz.
The second development is the Advanced Agility system that uses a network of seven sensors around the vehicle – four wheel sensors, plus two in the front and one at the rear – to detect levelling information and body forces. In response to the data received, electronic valves are actuated on each damper to adjust each damping map according to prevailing road conditions. Mercedes-Benz quotes a reaction time of 10ms.
Drivers can select either comfort or sport settings with the hydraulic forces increased in the latter mode for enhanced stability and reduced understeer at speeds up to 120km/h.
|
| |
Author
Ian Adcock
|
| |
| Email this article |
| |
|
|
|
|