In any given era, most cars are basically similar in both looks and engineering. Only a small handful are brave enough (or foolhardy enough) to swim against the tide and say, “What if instead of A, we tried B?” One example was the 1963–1977 Rover P6, a British near-luxury sedan brimming with novel features, including a very curious front suspension that took the typical double wishbone/coil spring layout and rotated it 90 degrees, with the springs set horizontally rather than vertically. If your reaction is, “Wait — what? Why?” then read on.
The 1963–1977 Rover 2000/2200/3500 — commonly known as “P6,” its factory project number — is one of those landmark cars whose relevance to the U.S. market was pretty close to nil. It was sold here for part of its run, but it wasn’t a commercial success, and its reliability problems made a bad impression on the few affluent American buyers who’d been intrigued enough by the Rover’s design and engineering to pay more than $4,200 for one.
The development and history of the P6 have been covered at greater length elsewhere, so I’m not going to try to recount it here except to note that the P6 had many advanced features, including four-wheel disc brakes (inboard in back), a De Dion rear suspension (which is worth its own post), unit construction with Citroën-like unstressed body panels bolted onto a sturdy welded “base unit,” and an unusual front suspension.
A dry tabulation of the major specifications of the Rover P6 would probably say that it had double wishbone front suspension with coil springs — which was true, and in no way novel in 1963. However, Rover did something quite peculiar: Instead of arranging each coil spring vertically, acting against a structural crossmember or the front fender apron, they turned the springs sideways, mounting them high and horizontally so they would act against the cowl/dashboard bulkhead (which the British call the “scuttle”).
Making sense of this layout from diagrams and cutaway illustrations is challenging: Aside from its horizontal springs, it appears at a glance to have a whole array of arms and links pointing in different directions to no obvious purpose. The American magazine Road Test (whose technical editorship was not strong) mistakenly thought it was some kind of weird derivative of the MacPherson strut, which it’s not. (The P6 DOES have an upper wishbone, although it’s arranged in an unusual way.)
Let’s start with the easy part. The lower “wishbone” is formed by two forged steel arms, connected by a greaseless ball joint. Their inner pivots have flexible rubber bushings that allow them to shift fore and aft to help absorb bumps. The outer end of the more forward of these two lower arms carries a socket for the lower ball joint of a tubular swivel pillar (sometimes called a “kingpost”), which carries the wheel spindle and serves as the steering knuckle. At the top of that pillar are the upper ball joint and the steering arm, which is an integral part of the forged ball joint housing. The steering tie rods move the arm, which causes the pillar to swivel on its ball joints and turns the wheel.
The role of the upper wishbone is played by a forged steel control arm, which pivots on a rubber-bushed fulcrum pin, bolted to the cowl with two forged steel brackets. The inside end of each control arm has a slot for the end of the hexagonal anti-roll bar, which is carried between the arms along the very back of the engine bay. (You can just make out the anti-roll bar under the left control arm pivot in the photo above.) If one upper control arm pivots more than the other, it twists the anti-roll bar.
The leading branch of the upper control arm extends forward to the top of the swiveling pillar, above the steering arm and upper ball joint. Roughly in the middle of that leading branch, there’s a lug that carries a steel pushrod. In essence, suspension loads cause the upper arm to rock so that the pushrod actuates the coil spring, which compresses against a spring seat on the cowl bulkhead. There’s a rubber bump stop that sits inside the rearmost coils.
Underneath the lug on the upper control arm, there’s an upper anchorage pin for the telescopic shock absorber (whose lower anchor is on the engine bay apron). The shock absorber works “backwards” compared to the shocks in a conventional wishbone or strut suspension, since it extends rather than compressing as the spring compresses.
All this will probably become easier to understand with this color-coded diagram. I unfortunately don’t know the provenance of this image (the site where I found it wasn’t sure where it came from other), but it makes the interrelationship of the components much clearer. I confess that trying to visualize the motion of the arms still gives me a terrible headache, but suffice it to say that this is still a double wishbone suspension, even though the wishbones are in different planes and the spring is lever-actuated.
The range of relative motion of the control arms is enough to provide 8 inches of wheel travel. The British journal Automobile Engineer noted in their 1964 analysis that the static roll center height of the P6 front suspension is 7 inches and that the arrangement of the control links provides some anti-dive in braking by increasing the front castor angle in bump, but I’m afraid I’ll have to take their word for it. (There’s that headache again …)
I promised at the outset that I would try to explain WHY Rover adopted this curious though effective arrangement. There were several reasons for it. The first was that in this era, Rover was doing a lot of work on gas turbine engines, and in fact some aspects of the P6 design came from their experimental turbine cars. Although the P6 initially offered only an inline four-cylinder engine, Rover was still thinking that they might eventually want to install a gas turbine in this car, and the front suspension design offered lots of space for the bulky turbine power pack.
The turbine never made it to production, but Rover later installed its license-built version of the Buick 3.5-liter aluminum V-8 in the P6, whose largely unobstructed engine bay made that relatively straightforward.
The second reason for this arrangement was that it was structurally advantageous. In a unit-body car, the front suspension loads typically have to be distributed to the body structure through the front rails (either through a crossmember or through big suspension towers), which requires those members to be stronger and heavier, and tends to transmit more force through the floorpan into the interior of the car.
With the Rover arrangement, the spring loads are applied directly to the cowl bulkhead, the strongest and stiffest part of the whole “base unit” structure. This suspension imposes no loads on the unstressed front fenders, and all the engine bay apron structure and front crossmember have to do is support the weight of the engine, support the inner pivots of the lower control arms, and carry the lower anchor of each shock absorber. Nearly all of the vertical loads created by the wheels and suspension go straight back into the cowl.
Third, Rover argued that arranging the front suspension in this way was better for safety. The front and rear sections of the P6 were designed for controlled crush in a collision, with the engine arranged to “submarine” under the car rather than crash through the firewall. This makes it advantageous to have most of the suspension components up, back, and out of the way, which also reduces their vulnerability in minor collisions. (The cam-and-roller steering gear is mounted on the cowl for the same reason.)
Did this curious knee-and-lever arrangement actually work? Surprisingly, yes. I’ve never had the opportunity to drive a P6, but the general consensus is that it provided an excellent blend of ride and handling, especially by ’60s standards. The suspension was set up soft, with a modest average front wheel rate of 91 lb/inch, but if you accepted that the car was going to lean quite a bit in turns, the P6 would stick tenaciously while delivering a very supple ride, in the manner of the better French cars. As far as I’ve ever heard, the unusual front suspension layout presents no unusual wear or service issues, although of course the rubber bushings and tubular shocks do eventually wear out.
So, if this layout worked well and offered packaging advantages, why wasn’t it more widely adopted? The usual reasons: $$$ (or rather £££) — remember all those forged steel components I mentioned? MacPherson struts or Rambler/Falcon/Mustang-style high-mounted coil springs require bulky suspension towers, but they can get away fine with simple stamped steel control arms, with fewer parts and less complexity. Even the P6 successor, the Rover SD1, switched to conventional MacPherson struts in front.
The eternal question for most clever or promising technologies is not whether they work, or even whether they offer significant benefits, but whether those benefits are enough to justify departing from the tried-and-true formula. Too often, the answer is “no,” which might be better for shareholder value, but tends to squelch the engineering creativity that went into cars like the Rover P6. That has unfortunately made for a far more monotonous automotive scene.
Related Reading
Storage Field Classic: The Very Advanced (But Mostly Forgotten) Rover 2000TC (by Dave Saunders)
Vintage R&T Road Test: 1969 Rover 2000TC – “A Highly Individual Car” (by Paul N)
When the scuttle went rotten as did most british cars in the 1960s and 70s the spring would punch through and end up behind the dashboard and said car would be on its knees ah the benefits of not rustproofing and its other party trick was the thin metal plate on the engine block would rot out and then one day burst as you drove along All in all not the worst car of the 60s and 70s but up there in the topgroup when they had a few yrs on their back
The one inalienable right of all Britons is for their cars to rot in England’s green and humid land. If a metal component is exposed, it’ll disintegrate, if it’s stout and built-up, rot will set in internally. Kind of a lose-lose situation.
Aways liked these, they were around during my childhood in Canada. Not common, but I saw enough of them to know what they were. The suspension is an ingenious design, but as you mention brought down by complexity and cost. If I ever drove one would I be able to stop thinking about the big coil spring pointed right at me on the other side of the scuttle?
When I was about 10 I went to an amateur off road race, there were a pair of P6’s there that different people were flogging around the track. The suspension aquitted itself well, leaping through the air. This continued until the engines blew in the Rovers. This was around 1977 so they were probably on their way to the scrap heap anyway.
Unfortunately, most USians know this car because Princess Grace (the erstwhile Grace Kelly) died in one…
The layout puts me in mind of a modern day Indy Car or Formula One car. In both vehicles the shocks are placed atop the central “tub” the driver sits in, usually about where his knees are. A pushrod attaches to one end of the spring, with the other end attaching to the lower A arm that holds the front suspension. The same set up exists at the rear of the car, with the springs attached to the top of the gearbox casing.
I’d never imagine a street automobile would use such a set up. Very interesting.
While it technically has two wishbones, it certainly doesn’t work like a typical two-wishbone front suspension. To my eyes, its geometry is essentially like a MacPherson type strut, with a complicated mechanism to transfer the movement from the top of the swiveling pillars.
Think of it this way: take a conventional upper wishbone, and rotate it 90° towards the rear, so the inner fulcrum axis runs transversely across the bulkhead, rather than longitudinally along the forward upper frame rail.
Then instead of having the spring compressed vertically between the lower wishbone and forward frame rail, it’s compressed horizontally between the upper wishbone and bulkhead. This is made possible by the upper wishbone having a kink in it with a pushrod extending from that elbow point, which pushes the spring. That same elbow point also pulls on the shock absorber as the suspension jounces, rather than pushing on it in the usual way.
It’s not a terribly complicated mechanism, just a somewhat unintuitive combination of 90° component rotations vs. conventional practice.
I quite understand how it works; I’ve been familiar with it for a long time. My point is the same as Robert Bray and XR7Matt pointed out below in the comments: in terms of camber change effect on the wheels, it’s essentially the same as a MacPherson strut. In fact, it essentially is a MacPherson strut (functionally) but with a complicated way of mounting the spring.
I had experience of these when they were already old cars, but they rode really nicely and stuck in the corners like ordure to the proverbial eiderdown, despite having relatively high roll angles.
The V8 versions were rather lovely and much prized, especially the manual 3500S.
The police in the UK used these well into the 1980s and were sad to see them go.
i always thought that the SD1, whilst looking modern was a bit of a backwards step in both sphistication, ride quality and especially quality…
Wow. Correct me if I am wrong, but it seems to me that with the upper control arm basically rotated 90 degrees the camber would stay more consistant than conventional SLA suspension. I see the increase in positive camber with suspension deflection, as the upper control arm moves upward the balljoint moves on an arc toward the rear tilting the swivel pillar (extended knuckle if you ask me) back. While the sway bar appears to be nothing more than a shaft tying the upper control arms together (an advantage of the arms being turned 90 degrees) the steering linkage must have been a real mess.
That’s exactly what I was thinking, the upper ball joint in a normal SLA /double wishbone follows the lower laterally, but due to their unequal length results in camber gain as the suspension compresses. That cannot work with this setup, it can change caster as the upper ball joint pivoted off the cowl will tilt the spindle longitudinally but I don’t see how that provides any real world benefit, let alone one that makes the compromise worthwhile.
It’s got all the shortcomings of a Macpherson strut without the simplicity of them.
One thing I can say with absolutely certainty racking up 2k miles per month in mine this year. That’s suspension is worth much much more credit than you can give it on paper. It rides better than any car of that class I’ve ever been in, you cannot make it loose grip it sticks and sticks, the steering for such a complex setup of 6 ball joints total for the steering links and also the worm and peg steering it is very direct for that type of steering, and it soaks up bumps better than anything I’ve ever been in. It put the jaguar mk2 to shame. The only criticism I have is removing the front springs makes me genuinely nervous. As there is no way to get a spring compressor in there and the J bolts have a change of coming out if the back plate is rotten. However speaking of rot. They don’t rust very much compared with everything else of that time. Ford’s, vauxhalls, jaguars, triumphs they all dissolved. The rover lasted a good bit longer than all of them before succumbing to the same fate.
Found this P6 development diagram on AROnline showing how the steering linkage works. Basically, the track rod from the steering box acts on a bellcrank that converts transverse motion of the track rod into longitudinal motion of the tie rod end, pushing the steering arm fore/aft to rotate the swivel pillar/kingpost.
Here’s the front suspension camber curve. I highlighted the axes in red, since it’s kind of bitty.
The steering linkage isn’t really that complicated. Here’s now it looks in plan view. (This is for a RHD car, so the shaft at the right is the steering column.)
I bought an old one on a foolish whim in the early days of E Bay – in Wales it was so I sent one of my sons to drive it up to Scotland.
It made it fine and it did have a nice ride compared to most UK cars of that era but although it always started and ran and the brakes worked, it was very tired , the gear change linkage was worn out and I could never get the twin carbs tuned right.
Realising it had faults well beyond my skill level to rectify I punted it on E bay to another fool.
As a Certified Volvo Nut™, I’d like to note that Volvo, in this same era, had a similar line of thinking about wanting the ability to fit a wider engine (a proposed V-8 that turned into a 90-degree V-6). This was the transition from the 100-series to the 200-series.
The 100-series had a traditional double A-arm suspension. Volvo redesigned the unibody of the 100-series from the cowl forward in order to install bog standard MacPherson struts. They saw no real need to reinvent the wheel (suspension) in order to accomplish the goal of a wider engine bay. And I think I can safely say that the Volvo 240/260 was wildly more successful than the Rover P6.
Sometimes it pays to innovate and other times, it simply doesn’t.
A simpler and more ingenious horizontal absorption component can be found on the Citroen 2CV. This provides an advantage in equalizing the front and rear suspension when going over bumps. https://www.youtube.com/watch?v=j7pFxgDmZXQ
Ah, classic Rover! First article to read today.
Rover used to be one of those companies that seemed to be heavily into unorthodox engineering. We’ve considered their sloping-head IOE (F-head) engines before; I still go back to that atricle and marvel at the degree of imagination to dream that up. This is the first time I’ve seen the P6 front suspension treated in depth.
It says a lot for their engineers that they were able to think outside the box to this degree. It’s all quite logical and sensible once it’s explained, though I too get a headache trying to envisage the relative motions of all those parts.
Rover’s motto back when they were independent could well have been “Pretium maledictum; qualitatem senti.” (roughly “Damn the cost; feel the quality”). Unfortunately that attitude proved unsustainable in the long term.
Which would make pre-British Leyland Rover the UK equivalent of pre-Fiat Lancia, who also refused to build to a price
“…Automobile Engineer noted in their 1964 analysis that the static roll center height of the P6 front suspension is 7 inches…”
This really jumped out at me. The front suspension numbers that I’m most-familiar with are from the ’64–’72 GM “A-Body”, Chevelle, Tempest/GTO, F-85/Cutlass/Skylark.
The roll center is an imaginary concept, and on those “A-Bodies” it’s actually BELOW GROUND LEVEL up front.
This is important–and in GM’s case, a deliberate sabotage of vehicle “handling”. The distance between the roll center, and the center of gravity represent a lever of sorts, the longer the lever, the more the vehicle will tend to lean in corners. This gets counteracted by spring rate (compromising ride quality) and by anti-roll-bars (“sway bars”) which somewhat offset what could be “independent” front suspension–the sway bar ties the two together through a calibrated-stiffness linkage. The stiffer the sway-bar, the less independent the front suspension.
GM WANTED lots of body lean in corners, so the idiot driving the car could not fail to notice the uncomfortable lean, and the squealing tires. They they put some band-aids on the so-called “performance” cars with higher-rate springs and stiffer anti-roll bars. The aftermarket goes even farther–much stiffer springs, much stiffer anti-roll bars.
But the real solution is to raise the roll center, and/or lower the center of gravity, so that there’s less leverage trying to tip the body in turns. This is done by clever placement of the suspension pivot points, (roll-center) and by reducing weight especially “high” on the vehicle (Center of gravity) For example, removing iron V-8 intake manifold and cylinder heads in favor of aluminum manifold and heads.
Short story: The English car, with a roll-center 7 inches above ground level, will allow softer springs, and less sway-bar to remain composed in turns. This increases passenger comfort compared to the American method of promoting wild corner-lean, showing a basic distrust of American drivers’ ability to judge safe cornering speeds.
An interesting point of comparison in this respect is that the roll center height of a 1963 Corvette Sting Ray was 7.56 inches.
One thing I’ve never got with this set-up: the damper seems to pulled on compression of the suspension, as the bottom of the damper is fixed and the upper is attached to the knuckle thingy as it swings away (or up, if you like).This is the opposite of how these things work in everything else. Are they unique to the Rovers, I wonder?
Dampers (shock absorbers) inherently are bi-directional, although they can be tuned to have relatively more or less damping on rebound than on…bound?
I’d put it as different damping rates on compression vs. extension. Normally we might say jounce vs. rebound respectively, but the P6 shocks extend on jounce and compress on rebound.
So the P6 had a good combination of ride and handling, but how did it compare to contemporary cars with more conventional suspensions?
It would be interesting to build a 3D CAD model of the suspension and see how it worked in jounce and rebound. That might help with Aaron’s headache. in 3D CAD it’s possible to create a visible axis between any two points, which would also be interesting.
I suspect that if the P6 had had good build quality and a good parts and service network in the US, the suspension wouldn’t have been a strike against it.
One of the commenters on Aaron’s AUWM history of the P6 related that at the time the Rover SD1 was launched, an interviewer asked Spen King why the design of the SD1 wasn’t as forward thinking as that of the P6. King said, “We were young and starry-eyed back in the day.”
So the P6 had a good combination of ride and handling, but how did it compare to contemporary cars with more conventional suspensions?
The P6 was always highly regarded for its excellent ride and good handling. From our vintage review:
“The result was a plush ride, remarkably so for steel springs. But that was not at the expense of very good control and stability, and the Rover had very high cornering power for a sedan. There was a certain initial “squishiness”, but as more steering lock was added, the car’s response increased, to the point of hanging its tail out gently without provocation. It can be driven at that attitude for brisk cornering, but it’s not necessary to maintain rapid rates of driving.” (my commentary based on the text).
An Alfetta would be the best apples-to-apples comparison with the P6, at least as far as the de Dion rear suspension was concerned.
My dad bought a TC 2000 new in 1968. British racing green. He was a car guy. I remember he researched a number of cars. I remember seeing a number of Road and Track editions around. R&T seemed to favor foreign cars. His choices came down to the Jag XJ and the Rover. His friend, Fred bought the Jag, dad bought the Rover. Even as a teen, and before I could drive, I knew this was a different car. It was our first foreign car. When I got to drive it, I drove it a lot. It had a nice combination of a well damped ride and a cornering ability that belied its soft ride. I “tracked” this car with my friend’s Austin Healey Sprite. Fun times.
The Lucas electric bugaboos were pretty well known then. The only repeat problem I can remember was the starter Bendix would malfunction leading to the common compression or bump start. The Rover was my dad’s primary car and he kept it until he bought a new 1986 Honda Accord. He and I were very fond of that unique car. Turns out there was only one other TC 2000 in the city. My dad became good friends with the other Rover owner.