Push vs Pull type rear suspensions... which is better?

[Ian P.]

A. Newey has said on a number of ocassions (that I know of) that car design at the F1 level is all about packaging.
Whether a rod is pushed or pulled (or both in all current F1 applications) it is a trivial matter to design the rod to withstand the forces. Keeping it small and light...that is the trick.
Buckling (not to take anything from the Swiss guy...) in compression is easy to predict and to design around. Tension forces even easier to handle.
The use of pushrods rather than pullrods on the front are entirely the result of the aero advantage of the high nose. Again, packaging.
At the rear, Newey has gone very low with the body work so the only option with decent leverage ratios was to put the linkage and dampers as low as possible, hence a pull rod suspension.
Check out the geometry of some of the other cars rear suspensions, the pushrods are laid over more than 45 deg. The actuation ratio is going to be over 2 to 1 between wheel travel and rod movement. Worse under full bump. The forces go up accordingly.
The pull rod geometry Newey has fiddled is better with a decreasing leverage ratio under bump.
The new fuel loads and any revisions to the aero regs. will dictate the designs for next year. Style doesn't play too big a role in F1. Thank gawd.
Gonna be interesting.

[xpensive]

With all the respect Ian P., I think you are passing some very sweeping judgments on what is easy or not within mechanical design engineering, as well as the different reasons for this or that solution.

[DaveW]

The "beam rocker" suspension layouts illustrated by Dave Killens' post were the norm before the adoption of ground effect aerodynamics in the late 1970's. They were found to be (much) too flexible to support properly the increased suspension loads that came along with ground effect. Some teams solved the problem with pull-rod layouts, whilst others chose push-rod. Pull-rod layouts were (are) the most efficient structurally, but lacked the stiffness of push-rods (partly because the latter had to be designed for buckling, rather than strength). Push-rod layouts also made springs & dampers accessible. For both reasons (stiffnesss & convenience) the push-rod layout became the preferred solution from around 1990 - apart from a very small number of exceptions, RBR being the latest.

I'm not sure about Ian P's "trivial" task of predicting buckling loads for slender beams, but he is correct about rear push-rod angles and the (implied) stiffness compromise caused by aerodynamics-inspired shallow "triangulation" angles. It is possible that the RBR rear pull-rod has a better triangulation angle when viewed from the front, but it is difficult to believe that the angle when viewed from above has no impact on rod loads and hence stiffness.

Overall, I would be surprised if the installation stiffness of the RBR rear suspension is not an issue. The fact that the RBR appears to work suggests that the ability to move the centre of gravity forward (i.e. the wheels back) is a fundamental requirement for performance in the F1 world post 2006.

[Richard]

Buckling (not to take anything from the Swiss guy...) in compression is easy to predict and to design around.

I'm not sure about Ian P's "trivial" task of predicting buckling loads for slender beams

The swiss chap is Leonhard Euler http://en.wikipedia.org/wiki/Leonhard_Euler

he did some useful sums that can be applied to the buckling capacity of struts
http://en.wikipedia.org/wiki/Buckling

[DaveW]

Good find, Richard. Your second reference contains the following sentence:

In 1757, mathematician Leonhard Euler derived a formula that gives the maximum (compressive - my addition) axial load that a long, slender, ideal column can carry...

A designer might be more interested in the minimum buckling load of a "real" column, & that will depend very much on manufacturing tolerances & imperfections when the column is slender.

For info., the Arrows front pull-rod could, with wheels in the air, be buckled by resting one hand on a wheel. I had the misfortune to prove it during a visit to Leafield. I'm reasonably sure, however, that it was strong enough to carry the design tensile loads, but the installation stiffness didn't rank amongst the best in the world.

[xpensive]

As I wrote before, as the neanderthalic kind of engineer that I am, spine tells me rods should be pulled and not pushed, just as coil-springs should be compressed and not elongated.
But what interests me in the pull-rod case, in practicality, when are you likely to see a compression load-case?

The pull-rod front suspension became in a way a Gordon Murray trade-mark, with coil-springs, not torsion-bars, arranged in a very intelligent way. Anyone with some inside on those designs?

At the same time, the before it's time Lotus 72 had torsion-bars, but when it was time for the equally ground-breaking Lotus 79, they were back with conventional coils and monstous top-rocker arms, why was that?

[DaveW]

But what interests me in the pull-rod case, in practicality, when are you likely to see a compression load-case?

I guess in the garage with its wheels in the air... But, for me, the real issue is installation stiffness. I have yet to see a decently stiff pull-rod suspension.

At the same time, the before it's time Lotus 72 had torsion-bars, but when it was time for the equally ground-breaking Lotus 79, they were back with conventional coils and monstous top-rocker arms, why was that?

Aerodynamicists wanted to get air into the tunnels, I suppose. The installation stiffness of the arrangement resembled knicker elastic (it had undamped hub modes that fired off with no damper movement to speak of).

[xpensive]

But as I can recall, during the entire venturi era, 78 - 82, eveverybody followed suit with the big fat rocker arms except for Brabham with the pull-rod BT49, correct?

Didn't John Barnards 1981 carbon fibre MP4-1 have rockers both front and rear, no?

Push-rods were never an issue with the venturi cars whatsoever, why?

[mep]

With all the respect Ian P., I think you are passing some very sweeping judgments on what is easy or not within mechanical design engineering, as well as the different reasons for this or that solution.

No, he does not. It is exactly the point.
The angles play a very important role for the forces who occur in all the suspension links.
The a-arms including.
For the springs, you want them to see a linear force (with constant ratio) while moving the wheel.
With the very low bodywork, this might get difficult to realize.
Or at least you get to high forces in the strut.

It is not that important whether the rod is pushed or pulled.
You can make a pulled rod lighter but the difference in weight might by minimal.
It is much more important to have the forces attacking in the right way.

[xpensive]

Thanks for sharing.

[mep]

Thanks for sharing.

hm, this comment sound a bit ironicaly.
I guess you want a better explanaition.

[xpensive]

In all honesty no, your earlier posting was quite enough to convince me of what I was not interested in hearing.

[DaveW]

xpensive. You are absolutely correct. Beam rocker suspensions did exist throughout the F1 venturi period. Apologies for the misleading generalisation.

However, my comments on the installation stiffness of different layouts remain valid, I think. Some estimates of front axle installation stiffness are:

Beam rocker (a venturi, unskirted Lotus) : 530 N/mm
Pull-rod (2000 F1) : 840 N/mm
Modern pull-rod (IRL) : 1.35 KN/mm
Older push-rod (IRL) : 2.2 KN/mm
Modern push-rod (F1) : usually > 2 KN/mm

The front tyre dynamic stiffness for the beam rocker vehicle was 390 N/mm but, to be fair, it was probably not representative of a "period" tyre. Hub mode damping ratio was 15 percent of critical with a natural frequency of 27 Hz. at both axles. Important? A more modern vehicle with a similar characteristic at the rear axle was prone to occasional "snap oversteer". The problem disappeared after the rear structure had been stiffened. The saving grace, perhaps, of the beam rocker Lotus was that the two axles behaved in a similar way.

So, what is an acceptable installation stiffness? A rule of thumb we used when designing active suspension vehicles was to target an installation stiffness of 3 x tyre stiffness. By that measure, the beam rocker was something of a disaster. The IRL pull-rod was OK on road course tyres, but marginal on speedway tyres. Push-rod layouts should, however, present no set-up problems. I believe it was this (as well as component accessibility) that has resulted in the general adoption of push-rod layouts in modern aero race vehicles.

[xpensive]

Dave,
Thanks for the engineerish data, not everyday you get that kind of info over here.
But are you certain of those stiffness values, say 2 kN/mm, what components exactly do you include in that, if its just the rod itelf, it's not that impressive is it?

Most of all, the beam-rocker must have been a bitch when you had all available forces wrapped into one, no?

But why did they hang on to it for so long, I mean, there was hardly any aerodynamical advantages with it, or was it?

[DaveW]

what components exactly do you include in that, if its just the rod itelf, it's not that impressive is it?

Obtained by reconciling damper displacement with suspension load & inertial displacement of the wheel rim (above the contact patch) relative to a sprung mass hard point in the plane of the contact patch. Measurements are recorded during a swept sine "heave" input to the vehicle through its tyres & are averaged left/right to minimize the effect of any vehicle roll motion. "Loaded" motion ratio & installation stiffness (of the whole stack - wheel, upright, links, rocker, rocker post & damper attachment structure) are identified from the measurements. (Apologies for the detail, but you did ask.) Theory was that the pull rod (carbon strap & Ti end fittings) contributed the major part of the 2K F1 compliance. The IRL pull-rod looks more likely, but is still considerably less substantial than the equivalent push-rod.

But why did they hang on to it for so long, I mean, there was hardly any aerodynamical advantages with it, or was it?

I guess that aerodynamicists ruled then, as now.... Logic wouldn't necessarily contribute (then, as now....) Lotus (for one) was usually successful when innovating, rather than refining.