beyond ackermann, an of the wall idea

The nice thing about this forum is that you put up a stupid idea and people will usually be polite shooting it down. I hope that stays true this time!

My idea is the "forward pivot axle front axle" which is shown in the two sketches linked below. Before describing it I'll try to explain the problems it is trying to solve.

- narrow track can help speed but can limit lateral G due to high weight transfer onto the outer tyre but wide track may raise drag. So it would be nice if you could have wide track on corners but narrow track on straights. In fact all you actually need is a wider track on the outside cornering wheel.

- It would be nice if there was away of automatically raising the spring rate at the wheel as weight transfer increases vertical load.Then you don't use up outer wheel movement ( like active ride apparently could do).

- As the inside wheel hitting a kerb upsets the balance it would nice if the inner wheel could move away from the kerbs as the car goes round.

The forward pivot axle gets round these problems by swinging the wheels and axle around a forward pivot , in this way the outer loaded wheel moves away from the car and steers the car around. At the same time the outer wheel pull rod swings the spring / damper into a less effective leverage position so raising the rate. At the same time the inner wheel moves in towards the car so allowing it to give a little more kerb avoidance. The inner wheel rate is reduced by the opposite effect of the outer wheel pull rod geometry which helps bump control.

The sketch is here


As you can see (hopefully) the axle assembly is steered by the rack which also acts as the lateral location and thus takes the cornering load. To allow thiis it has a spiral thread rotated by the steering wheel via a step up gear. The screw thread or spiral makes it non reversible so allowing it to both steer and take side loads.

The spring pull rod arangement changes the angle of attack of the spring as the axle pulls it towards the outer, loaded wheel.Thus as the cornering load on that wheel rises the spring rate climbs to match it.
So automatic matching of spring rate and load.

There are some practical difficulties that even I can see.

1) the forward links react the brake load so will cause dive.

2) the front pivot link length is short so a large steer angles are required to move the outer wheel very far to the left ( in the sketch).

3) The pull link/damper/inner link geometry as shown would cause the rate to rise too quickly.

These problems are (hopefully) fixed in sketch 2



By changing the front linkage to four semi parallel links any desired anti dive can be factored in and by using four links a virtual axle pivot point can be set far enough aheasd to allow the right axle out - swing versus steering angle ratio to be created.

By using a similar virtual linkage for the spring inner link the rate rise on the axle swing out can be controlled to a desired level.

So I hope have explained it reasonably well, now you can shoot it down!

you won't get much steering lock I think, or at least you'll need a very long travel rack (not impossible)

Why not use two steering racks, as forward and rearward lower links, then you can control the lateral shift and toe independently?

High steering effort (and increasing with latacc) since all that wheel translation is opposite to cornering force.

High steering effort (and increasing with latacc) since all that wheel translation is opposite to cornering force.

Yes, but for a light car it might make sense. Say 100mm rack travel, in one second, so 50mm in half a second, s=1/2*a*t^2 a=0.4 m/s/s. If the sprung weight at that end is 300 kg that'd be rack force of 120 N which is not that much (600 N is about the limit without PAS). Admittedly that is pushing the sprung body inwards rather than the wheel out, in reality you'll get a bit of both.

Greg, I think you are talking about something else. The plan-view IC for the steered wheels is well forward of the axle line. This is equivalent to a large trail - perhaps hundreds of mm depending on how much the track is to be offset. The steering work for a 100mm offset, 300kg sprung weight, cornering at 1G is
300 x 9.8 x 0.1 = 294 J I guess you can halve that if you assume latacc is zero when you start turning the wheel and increases linearly with steering angle.

Greg, I think you are talking about something else. The plan-view IC for the steered wheels is well forward of the axle line. This is equivalent to a large trail - perhaps hundreds of mm depending on how much the track is to be offset. The steering work for a 100mm offset, 300kg sprung weight, cornering at 1G is
300 x 9.8 x 0.1 = 294 J I guess you can halve that if you assume latacc is zero when you start turning the wheel and increases linearly with steering angle.

Newton

What would he know about cars?
Need more assumptions to get a force. Let's assume the 100 mm offset occurs at 10 degrees of steering angle. That requires 572 mm of trail. If steering ratio remains the same, steering effort would be about 28 times higher than conventional steering with 20mm of trail.

What would he know about cars?
Need more assumptions to get a force. Let's assume the 100 mm offset occurs at 10 degrees of steering angle. That requires 572 mm of trail. If steering ratio remains the same, steering effort would be about 28 times higher than conventional steering with 20mm of trail.

I'm talking as if the car was on rails, so the sprung mass moves inwards, you are talking as if the body was on rails pushing the wheel out. I acknowledged that there was a mixture, in practice the body moves far more than the wheels.

Newton says the force on each is equal and opposite.

I am talking steady state cornering (my last post at least), the cornering force is trying to re-centre the wheels via the massive trail.

AH, ok yes.

Not that i believe in this but you could possibly offset some of the forces with dynamic toe settings.

Thanks for the comments guys. I had sort of assumed that power steering would be needed to move the rack far enough quick enough against full lateral G.

I envisage a rack with an external screw thread matched by a rotating barrel with an internal thread. This would be physically driven by a step up gear from the column for safety but would also be rotated by an electric motor. I think power steering has been in single seaters before. The power could an issue but I am only trying to do something similar to active rams,but one not four.

I am too weak at physics to articulate this well but one thing about this idea is how it would affect turn in and yaw. In a normal set up the weight transfer at the wheel rim is the result of a chain of events. Turn steering wheel, generate steer angle then slip angle then move car mass laterally. In the "front pivot axle" the lateral movement of the car is the first thing that happens because you are physically moving the mass of the car's front end over into the turn direction - bit like a rider steering the bike with his/her knees.

Whether that would change the yaw response time and if that is good or bad I am not quite sure. I think it would and whilst it would make the car more "nervous" it might also allow some gain in not having to initiate the yaw vis a change of tyre related events, so you could use the tyres for other things .... if that makes sense

Edited by mariner, 28 October 2011 - 09:39.

I can see some benefits for sure but my main concern is the peak steering force, the high level of power assist that would be required and the resulting "loss of feel". At the very least the power assist woulod need to be non-linear eg 50% assist at low forces and 90% at high forces.

With a secondary steering adjusting the "toe" you can set the car up prior to the corner.

what you do is that you make the rack go out to the side by steering the wheels there. the car will not move or move marginally because the rack will offset the wheels going to the outside.

Kinda like a motorbike where you lay it down "before" the corner.

Edited by MatsNorway, 28 October 2011 - 14:58.

- In fact all you actually need is a wider track on the outside cornering wheel.

- It would be nice if there was away of automatically raising the spring rate at the wheel as weight transfer increases vertical load.

- As the inside wheel hitting a kerb upsets the balance it would nice if the inner wheel could move away from the kerbs as the car goes round.

- No, you need wider to the inside as well, the last 20 years have been all about using both front tyres not just the outside. Oddly Nascar, seen as a low tech sport, has actually been leaders in this area and a reason for their appreciable decrease in lap times. That said of course an offset is going to be an improvement.

- Lotus do it in F1 now where the suspension pushrods mount to the uprights eccentrically and jack up the outside of the front, we had a thread recently about it.

- Citreon DS19 had a narrower rear track to negotiate city corners better with it's long wheelbase and to this day I am surprised no F1 team has run a special narrow rear tracked car for Monaco.

[/url]

By changing the front linkage to four semi parallel links any desired anti dive can be factored in

You have no idea how close your drawing is to my front end right now!

Edited by cheapracer, 29 October 2011 - 03:26.

Are you not pulling the inner wheel spring/damper unit into a more vertical position - hence jacking and raising the rate
and pulling the outer wheel spring/damper unit into a more angled position - the opposite of what you want - I see that it would need compressing as you turn
There are some extremely weird dynamics going on with the spring/ damper units and the front of the car going up and down as you turn

Very interesting idea/concept though

Carl , you are quite right about the spring/damper angles. The dampers should slope outwards from the pull rod joint to the chassis to get rising rate. I had the original sketch right but made a misteke when I redrew it, sorry.

Having gotten peoples feedback I have thought a bit more about the idea and why you might do it.

The big negative is obviuosly the rack power but if a Moog strut can push up a loaded rear tyre I guess it is possible ( oh, for Dave W's knowledge).

However much you play with lengths and arcs there will always be some compromise on the rates as you are plotting arcs vs straight line axle/wheel dispacement.

However I think I have seen why you MIGHT use this ( at least in theory). The lateral displacement will always be more on slow cornrs ( high steer angle) than fast corners. So it's benefit ( if any!) is to take load off the rear wheels in slow corners and hence improve traction. It does that by physically shifting weight onto the front inside wheel and by adjusting the outside wheel rate in line with load. So you have another mechanical way of balancing the car without using roll bars which operate equally at all speeds. I think that is the key potential gain apart from drag.

I am still struggling to see if the instant turn in benefit exists but that's beacuse my non-maths brain can hardly climb even the foothills of tyre modelling


So it might be best suited to hillclimbs where some very sharp corners exist and the steering loads woul;d not be too excessive as the cars are very light

Steering 'feel' might be the biggest failing in this
But if all the math works , then the only way to find out is to BUILD it
Plenty of dirt cheap 'almost finished' Locost projects on ebay these days !

I would love to build all these things but I don't have enough time left, money or space to do them all.

Maybe if I win the Eruo Millions Lottery I can rent a workshop unit, give my wife a space for her quilting and sub contract a lot of it to do these things!!

Wouldn't fancy trying to reverse the car out of a tight space, or drive down a narrow spiral ramp, you'd be rubbing the outside wheel against the outside wall and the doors against the inside one...

I would love to build all these things but I don't have enough time left, money or space to do them all.

Maybe if I win the Eruo Millions Lottery I can rent a workshop unit, give my wife a space for her quilting and sub contract a lot of it to do these things!!

prob against rules ?

http://cgi.ebay.co.u...3#ht_881wt_1043

I believe Lotus did something similar for a soap box style racer.
Lotus Gravity Racer

I believe Lotus did something similar for a soap box style racer.
Lotus Gravity Racer

No evidence of it, completely normal upright pivoted steering.

The technical reason is there to suppoert what I said in post 15, first paragraph though not that rocket science is involved with understanding what a wider track does.

No evidence of it, completely normal upright pivoted steering.

The technical reason is there to suppoert what I said in post 15, first paragraph though not that rocket science is involved with understanding what a wider track does.

I believe at least one version of the Lotus 119 design translated the front axis laterally in response to steering input to reduce the weight transfer to the outside wheel, which Lotus calls active mass distribution on one of the graphics. The wheels in a gravity racer are narrow and highly loaded per unit area and as such grip saturates quickly under addition loading, so a balanced weight distribution provides more cornering force, which is another benefit of the offset design. Since gravity racing has a very small frontal cross-section they probably don't benefit much from a narrow track. I have seen a page that better describes the details, but neither Google nor I cannot find it now.

It would be great to see it built at least in a simulation. In addition to the concerns raised so far, steering force and steering feedback, I am concerned about mechanical rigidity vs. weight, packaging for wide tires and how the offset will interact with the rear tire slip angles.

In one way it is a little bit like 4 wheel steer, it upsets the relationship between the average steer vectors of the two axles.

Not sure if that is good or not.

As to simulating it, I'd be interested to see it but it isn't a standard template.

I think Mats' idea of steering the front axle outboard of the car and then turning the wheels in is elegant, if rather hard to program.

I think Mats' idea of steering the front axle outboard of the car and then turning the wheels in is elegant, if rather hard to program.

And you can never ever get a counter.. then you will probably roll overto the other side.

Kinda like a motorbike only that you have a bigger margin for error before you actually roll.


When i think about it. it all depends on the reaction time of the driver. if he manages to get the steering rack to the other side in time he will be fine. if its a old granny she is in deep ****. But i guess she would have been that regardless of steering system.

Edit2:

This system should be a blast in chicanes. you just throw the wheels around the curbs instead of going over it.

Edited by MatsNorway, 03 November 2011 - 08:30.

I believe at least one version of the Lotus 119 design translated the front axis laterally in response to steering input to reduce the weight transfer to the outside wheel, which Lotus calls active mass distribution on one of the graphics. The wheels in a gravity racer are narrow and highly loaded per unit area and as such grip saturates quickly under addition loading, so a balanced weight distribution provides more cornering force, which is another benefit of the offset design.

I would probably call it "lots of trail".

This system should be a blast in chicanes. you just throw the wheels around the curbs instead of going over it.

The rears are a bigger problem than the fronts as they run a tighter radius.

Edited by cheapracer, 03 November 2011 - 09:30.

Isn't this a bit like the one-time off-set suspension on Indy cars, but with the bonus of being able to shift the off-set from left to right as required?

well if Lotus considere the idea of axle weight transfer I am cheered up.

There is another way to play with this sort of concept without high rack loads.

Attach the two front wheels to a simple frame with a fifth wheel at the forward apex, connected to the steering wheel , and pivoted somewhere in the middle. Then use the front " steering" wheel to move the rear " cornering" wheels on the frame.

I am not suggesting that as an actual layout so much as a way of modelling the general concept using existing techniques

sort of do the steering whel , then treat teh front "cornering" wheels as an " axle" , then do the rear wheels last. As the steering wheel takes virtually no lateral force all the lateral stuff can be trested as normal.

I am now almost certainly way out of my depth on dynamic/tyre modelling so I will just shut up!

Isn't this a bit like the one-time off-set suspension on Indy cars, but with the bonus of being able to shift the off-set from left to right as required?

yes

Cheapy

One less wheel on the curb is still better.

And espesially on corner exit.

then only the front wheels would hit the curbs. instead of oversteer it would perhaps induce understeer.