T He Stabilished Car
Imagine a very softly sprung car which doensn't roll in corners, doesn't pitch or drive in braking and acceleration and always keeps its wheels at right angels to the road. Automotive Products (AP) and Rover have produced one; it was demonstrated at the Safety Congress at Crowthorne earlier this year and nobody but Motor seemd to take much notice. We were fascinated both by the car and its implications and since that time we have become the first outsiders to drive the original Rover 3500 test car both on the road and test track. A second similar suspension has now been fitted to an MGB which has gone to the States as part of the British Exhibit at the Washington Transpo safety conference.In a nutshell, it comprises a very fast (in fact almost instantaneous) hydraulic levelling system acting on a hydr-pneumatic suspension like Citroen use. On the face of it this is not practiable for reasons which suspension egineers will see at once. Herein lies its ingeniuty so we will return to this point later.
Automotive Products started work in 1964 by puttin a couter-banking roll system on the front axle of a small British car. Encouraged by the results, they tried a more elaborate system on both ends of a Lotus Corina and measured gains in cornering power of over 20 per cent. The Rover Company was particulary interested in this work and in 1966 suggested a joint experementon a Citroen DS, which already had the right sort of basic suspension. This led them to design a complete installation of their own which was fitted to a Rover 3500. By this time (1969) Rover had largely taken over the whole project but, partly because of the British Leyland merger and various changes of staff, work stopped for a couple of years until AP stepped in again, bought the 3500 and restarted development work.How it Works
Let's start by saying why it can's apparantly work. All full levelling systems operate by measuring the distance between the body and the wheels (back and front) and correcting this to its designed static value. The correction may be done hydraulically, pnematically or electromechanically - it doesn't matter. This is fine for compensating for load variation but, of course, the system can't distinguish between a decrease in body height due to four passengers getting in and a temporary decrease due to a wheel passing over a bump. Or vice versa for porholes. So it will tend to oppose all normal suspension movement and it it were quick enough to keep the wheel-to-body distance always constant, the car would have no suspension at all by definition.
Present cars with levelling systems (eg Citroen, Rolls-Royce, Mercede) avoid this dilemma in one of two ways. Either they impose a delay between the signal and the response (about 10-12 seconds in a Citroen as you can see by pressing one end down) or they make levelling such a slow and gradual process that there is no significant response at all to very short duration changes.
But the whole essence of the AP system is that it can respond so quickly that it will cancel roll, pitch and dive as they begin to happen. The dilemma is resolved with a most ingenious valve. Figure 1 shows one front suspension unit diagrammatically; movements of the suspension arm operate a piston wich compresses the gas of the gas spring through hydraulic fluid which is separated from the gas by a flexible diaphragm. The balanced three-line valve (extreme left), normally in the central position, can move left or right to add to or subtract from the quantity of fluid betwee the piston and the gas spring which, in turn, raises or lowers the body.
The valve mechanically operated by the link from the suspension arm - but not directly. Interposed is the horizontal pendulum supported by a coil spring and damped by its own small damper. When a wheel passes over a bump it produces, through the main spring and damper, an upward force which accelerates the body of the car upwards.
The pivot of the offset pendulum moves upwards with the body and this would leave the mass of the pendulum behind, thus operating the hydraulic valve, were it not for the input from the suspension arm acting on the base of the pendulum spring. This is arranged to accelerate the pendulum mass upwards at the same rate as the car body accelrates the pivot leaving no net pendulum or valve movement relative to the body. In other words, the pendulum suspension system is a perfect model of the car suspension system both in frequency and damping so that it filters out all transient signals from the unsprung parts whilst reacting against all the steady changes which would be caused by load variation, roll and brake drive.
The general arrangement is shown in Figure 2. The two valves at the front control the front struts in roll, pitch and height (the single valve at the rear controls pitch and height only), the two rear struts being interconnected and balanced. However, there are additional hydraulic sylinders in the rear struts which are diagonally cross-connected to the front ones so that the front valves also control the roll couple applied at the back and, in this way, the distribution of roll couple correction front to rear can be pre-determined and also synchronised; on the Rover about two-thirds is applied to the front, one-third to the rear.Pros and Cons
This advantages of constant height, attitude and suspension rates irrespective of load are obvious but in general these are also realised with less expensive conventional levelling arrangements. The real bonus is that you can make the suspension as soft as you like for maximum comfort without worrying about roll, brake dive and all the syspension compromises that go with these movements. From a design point of view this is a tremendous liberating factor because once you eliminate roll and keep body height constant, the wheels will also remain vertical (or whatever camber angle they were initially set at) and the suspension layout and geometry can be chosen for convenience or ideal movements characteristics instead of demanding, as now, a deliberate compromise between entirely conflicting requirements.
In fact, although we won't go into it here, much of the existing paraphernalia of suspension theory and handling development would become redundant - roll centre heights, roll axis slopes, wheel chamber changes, anti-roll bars and roll steering effects generally would all be consigned to the past. Who, we wonder, will be the first to investigate these implications in motor racing? An Indianapolis car could well be the ideal application.
Keeping the tyres at an optimum angle to the ground is one of the most important features. It should improve tyre wear, facilitate the trend towards wider low-profile tyres with flatter treads and, of course, improve braking, cornering power and control during avoidance type manouvres. On these things are founded its qualifications for appearance at safety exhibitions. A lot of timed tests which have been done on a standard chicane using the dame driver have shown an increase from 48.5 mph to 54.5 mph as a result of fitting the AP system - the latter being very fast indeed for a fairly large saloon.
But, of course, there is one very big snag - cost. The suspension parts themselves should cost no more than those of several cars in mass production now or planned for the near future, but the variable delivery hydraulic pump itself could be costly. To act quickly enough it has to produce a very high flow rate at about 2000-2500 lb/sq.in. pressure and absorbs some 15 hp at full output although naturally much less most of the time. This pump could cost about £ 45,- according to present estimates. Perhaps volume production would reduce this to a price which wouldn't look so high when split betwen the other features it could energise - fully-powered anti-lock brakes, powerassisted steering and a servo clutch.Driving impressions
In fairness to AP, We should say straightaway that their Rover 3500 is no super sales and demonstration car --it is a hard-used engineers' test rig and the anti-roll suspension is a ccomplicated device stillin its earlier stages of development from which some of the newer and better parts have been withdrawn to equip the MG in Washington. So its behaviour is perhaps not so mauch a blinding revelation as a conscious process of peering through the metaphorical grime to the gold beneath; this won't surprise anyone used to working on test vehicles.
Unlike most self-levelling cars, which assume their proper attitude so slowly that you can't always see the process, this one springs up with a great leap when you start the engine. If you then jump on the front of the car to defect the springs, it doesn't apparently move at all - you might as well jump on a Sherman tank. On the other hand, if you move one of the control valvesunder the bonnet by hand you can make the whole car leap about like a flea with St. Vitus dance.The first impression when we drove away was the sound of the hydraulic pump; it drones away like a low-pitched siren rising and falling in note with engine speed and dominating all other mechnical noises. This isn't considered a serious problem, as it is mainly a question of isolation and a quiet installation has already been made.
So far little attention has been given to ride and this particular car isn't as comfortable as a standard 3500. It is thought to be considerably overdamped at the front and this is certainly the impression it gives. It really doesn't show any sign of brake dive or acceleration squat and the abscence or roll makes the passenger feel much more stable and comfortable, as we found on the test track where a standard 3500 was available for direct comparison. Watching it go round in circles at a speed near the breakaway point the only roll you can see is due to tyre deflection which at this sort of lateral 'g' can contribure 1° or more. As a matter of fact it is very easy to modify the valves to compensate for this or even to over-compensate so that the car rolls in the opposite direction like a bicycle or a speedboat, but the AP engineers think this would be a mistake. Psychologically it might lead to overconfidence on the part of the driver, technically it re-introduces all the roll steer effects and their associated time lags which have just been eliminated.
The only way We could produce some roll was by doing a very rapid wiggle-woggle (slalom) through closely spaced bollards. In this case it seems likely that the hydraulic demand exceeded the supply from the pump in a worn system with considerable leakage. Probably the most interesting difference to the driver when manouvring quickly is the enormously more rapid response to the steering. Indeed, at first this is almost embarrassing and We watched another AP driver do a hasty correction at the first left hand bend he came to to avoid hitting the kerb.
Coupled with it is a feeling of initial oversteer - an impression that if you steer the car quickly and firmly into a corner, it will spin in the initial stage. It doesn't seem to do so but the whole feel of the car remains wuite different throughout the corner. In the standard car it was possible (on an airfield) to get the car into an understeering condition where you had considerable lock wound on and were exerting so much force on the wheel that you felt that you were lugging it round by brute force. On the AP car the steering in similar conditions remained quite light so that you could always put on a bit more lock quite easily - and the car would respond to it.
Because of its the Dion rear suspension, the AP modifications have probably reduced the understeer considerably. How much of the difference in feel this accounts for it is difficult to say but we believe that tyre self-aligning torque is much less for an upright wheel than for one with several degrees of positive chamber.
Special Thanks to: Charles Bulmer