Steering Shaft Design

[Chris Hagen]

Hi all,
I wanted to ask for some opinions. The shaft from the rack to the wheel generally is broken up by u-joints, slip joints or similar on most production cars. On high level formula or LMP cars I have not seen any details of the shaft and felt that if the shaft is lined up, short enough and the wheel mount and rack mounts are rigidly located relative to one another there may not be a need for joints. In most cases the shaft is long and even if you have a driving position/wheel orientation lined up with the rack input there is some chassis deflection that may cause some minor bind or stressing of the shaft at the ends.
The goal is to create the most direct input to the rack. Possibly a joint at the rack to account for chassis deflection and a strait shaft to the wheel with the mount near the wheel being flexible rotationally like a bushing or spherical? Less is more here, so what have you seen?

Chris

[Greg Locock]

Putting Hookes joints in the column has a couple of advantages. The first is that spear aimed at your chest is no longer an unbroken shaft. The second is that by phasing the U joints correctly you can create a fast steering ratio on centre, or a slow one. This allows you to trade what you might loosely call nimbleness for stability.

[Tim.Wright]

Looking at these cutaways of a Sauber F1 and Audi LMP1 it seems that they do use joints with very shallow angles.

The biggest advantage of having joints is that you can adjust the steering wheel height.

[Tim.Wright]

And regarding binding, if you can make the shaft stiff in torsion and soft in bending then the shaft itself could provide the required alignment degree of freedom to minimise binding.

Alternatively, I've also seen the upper steering column passing through the eye of a monsterous ball joint to create an auto-aligning mechanism. It's a bit of a dog's breakfast IMO as it's really heavy and bulky. A rubber/plastic bushing will probably have the same effect for a lot less weight.

[Smokes]

I wonder if a cv joint would give better feed back to the drivers; but, looking a the small angles the UJ would change the angular velocity too much.

[Tim.Wright]

Main issue with CV's (like the ones in driveshafts) is that they have backlash which is no good for steering systems.

[Greg Locock]

If you want a CV use a double cardan.

[Inertia Lab]

Had to change usernames for some reason. This is Chris Hagen the OP.

My thought was to make a large OD (50-75mm) thin wall (1 mm-ish) shaft with waves like a bellows/rack boot near the rack input and steering wheel with a similarly thin walled large OD shaft between them. I have not found any source for making such a part, but to provide a tiny misalignment angle and a mm or so extension and contraction with chassis flex may be possible and formable with a hydro-form of a sine wave along the length of the tube with a nearly equal period and amplitude (10-15mm) for 150mm or so at each end.

[Inertia Lab]

This would also crumple up to 300mm and likely can be made to induce the intermediate tube to crumple as well if need be while maintaining a direct strait path.

[Greg Locock]

You need to be careful with a feature like that in a steering column. Fatigue risk is high. Columns are zero redundancy systems, you have to engineer in robustness right the way through (yes I have worked on columns).

[Inertia Lab]

Thanks Greg,
I’ve had some proximity to steering design, but was not aware of the details of the constraints. Where were you involved with the column design if you don’t mind me asking?
As far as mechanical design I feel up to the task, but Im not aware of the margins and industry methods used to detrmining a safe system. I know where I would start, but I don’t know what I don’t know here. Im curious what kind of constraints were the columns designed to in your experience?

[Greg Locock]

I took the Mercury Capri column through FMVS signoff, after the (Audi unintended acceleration) interlocks with the T bar were fitted. My excellent predecessor had negotiated a series of tests with NHTSA ( I guess) that meant that we could fit the new column without a full crash program.

The standard collapse feature these days is lots of plastic balls buried in pits in telescoping shafts. I'd take a mad guess that you need a FoS of 10 if you are designing blind.