36 replies to this topic

[jvl]

Does anyone know how to calculate the travel of a rack gear in a rack and pinion steering system in order produce classic ackermann geometry.

[Ray Bell]

Is that necessary to do?

My understanding is that a line drawn through the kingpin to the centre of the rear axle line should include the centre of the steering arm ball joint to give you true Ackermann effect...

Or was somebody having a lend of me?

[Patrice L'Rodent]

Hi JVL, why do you want 'Ackermann' geometry ? Are you building a car of some description, and what do you want to do with it....becaus thee the requirements are very different.

The classic layout as described by Ray is only true if
1. your tyres run at a zero slip angle, and
2. It assumes all other kinematics are correct. What you are really looking for are the effects of Ackermann geometry, ie toe out as the vehicle steers, and that can be achieved in different ways (Rack position, for instance)
There is also a very good argument that there are times when you should be using reverse Ackermann. This is an extremely complex subject, and a year or so ago, all the race car tech magazines were bubbling over with articles on the subject.

Oh, and the geometry was developed by Langensperger, not Ackermann !!
Regatds
Pat

[desmo]

The subject of Ackermann is a hornet's nest, albeit a rather interesting one. Shall we put the pole down and slowly back away, or give it a good whack and see what happens?

[Keith Young]

i say we have a go, i have some 2d cad layouts of ackerman if i could just dig them up...

[Ray Bell]

Originally posted by rough_wood
i say we have a go, i have some 2d cad layouts of ackerman if i could just dig them up...

Sounds interesting...

Warm up your computer.

[Patrice L'Rodent]

I have the skipping rope out and am in full training.....Just two days out of my most recent Rouelle seminar, so I'm ready to go
Pat D'Rat

[Patrice L'Rodent]

Well, that died in the bum

[desmo]

Rumor is that Mr. Colin Chapman once wrote a paper on Ackermann. Anyone seen or have a copy of it?

[JtP]

Originally posted by desmo
Rumor is that Mr. Colin Chapman once wrote a paper on Ackermann. Anyone seen or have a copy of it?

I do not have a copy of the paper, but Chapman liked using reverse Ackermann on early Lotus racing cars.

[Ray Bell]

Originally posted by desmo
Rumor is that Mr. Colin Chapman once wrote a paper on Ackermann. Anyone seen or have a copy of it?

No, but he's just moved in down the road from here...

Want me to go ask him?

[desmo]

Originally posted by Ray Bell


No, but he's just moved in down the road from here...

Want me to go ask him?

ACB Chapman, who died in the early '80s? That's a tough neighborhood Ray!

[Ray Bell]

Oh?

I thought you'd heard all the rumours... I just wanted to feed them some more...

[BRIAN GLOVER]

This thread don't look like it's going no place.
Chapman was involved with Burt Rutan when he had his heart attack. Seems that he had lost interest in cars and they were designing a revolutionary aircraft together.
For you aviators, the annual EAA convention called Airventure Oshkosh 2003, starts July 29 thru Aug 4. Burt Rutan and John Roncz (Self taught aerodynamicist with an IQ of 200) will be speakers at one of the many forums held during the week at the Forums Plaza Aug 2 @ 2:pm.
Burt will tell you about his latest project, 'Space Ship One' in the 'Tent Talk Show' on Aug 2 @10am and Aug 3rd @ 2:30pm. They are the most popular speakers there next to Chuck Yeager.
There is also the 'Count down to Kitty Hawk Pavillion' comomorating the 100th anniversary of powered flight. This year will be the best ever.

Back to Ackerman, I notice that the tie rods run alongside the upper leading A arm on F1 cars, making the rack pretty high in the nose. What does the steering arm look like? Would there be 0 Ackerman? would there be any bump steer? IE, would there be 0 toe under aerodynamic load going to toe-out in slower sections? Does anybody use rear steer? What is the current thinking in this part of the steering geometory? What sort of scrub do they use if any?
Surely they could lower CG by having rear steer with the tie rod running along side the lower rear A arms which connect in the middle anyway. Instead of the usual rack in this case, they could have two hydraulic actuators pushing the steering arm via the tie rods. In the case of a twin keel design, the actuators could be housed within a wing section.

Originally posted by Patrice L'Rodent
Well, that died in the bum

[red300zx99]

I'm sure doing some calculations using tire data would show which would be better, ackerman or reverse ackerman. Whatever the case this may be a hornet's nest, but I think the many minds of this board could figure something out.

[desmo]

The effect of downforce loadings on the tire curves- specifically the migration of the friction peak value relative to slip angle as the downforce component of the normal force at the contact patch changes by roughly velocity squared- means that one cannot calculate optimal Ackermann geometry without knowing the real tire curves doesn't it? Since no one this side of God has access to the actual tire curve data, we cannot run the maths and must talk in generalities or assume Pacjecka (spelling?) equation calculated values or some other general assumptions about tire behavior.

Didn't Racetech offer some free Ackermann calculating software on its website? Perhaps they still do and some rough guesses could be arrived at by eyeball regarding the Ackermann values used in F1.

[Ray Bell]

Originally posted by BRIAN GLOVER
Back to Ackerman, I notice that the tie rods run alongside the upper leading A arm on F1 cars, making the rack pretty high in the nose. What does the steering arm look like? Would there be 0 Ackerman? would there be any bump steer? IE, would there be 0 toe under aerodynamic load going to toe-out in slower sections? Does anybody use rear steer? What is the current thinking in this part of the steering geometory?

Doesn't matter what anyone thinks, I'm sure it's not allowed under the F1 regs...

And as for the bump steer question, bump steer is eliminated by using steering arms parallel to the main links and making them the same length. I suggest that this is the normal practice, and harks back to the Lotus 25.

[BRIAN GLOVER]

Sometimes they are not parallel like when they are standing in the pits with wheels straight ahead or in slow turns. I assume they become parallel under load to keep the section that they form at its lowest drag profile. Something must happen in those 10mm of travel.
In a turn, with caster coming into play, there must be a toe change even with 0 Ackerman and I would suspect that they can even adjust bump for tire curves and other conditions.
Ive seen them close up on the McLaren, and I wondered why they where not parallel and this is what I came up with.
The link and the tie rod may be the same length, but the position and shape of the steering arm would determine bump steer.
I suspect that there is toe out at speed under load or in roll in the front and toe in at the back wheels.

Originally posted by Ray Bell


Doesn't matter what anyone thinks, I'm sure it's not allowed under the F1 regs...

And as for the bump steer question, bump steer is eliminated by using steering arms parallel to the main links and making them the same length. I suggest that this is the normal practice, and harks back to the Lotus 25.

[RDV]

Discussing Ackerman, or epure de Jeantaud if you are french , or Langensperger steer to be accurate is akin to different branches of the same religion discussing the interpretation of holy relics.

Main effect would be to ensure no scrub going around corners as for a given radius of trajectory (mean car axis distance to center of curve) , and assuming no slip (wheels track and have no slip angle relative to ground movement) outer wheels would be describing a circle of bigger diameter than inner wheel circle (by exactly the track width) . Analogous to reason for differential on rear axle. Much used on road cars to minimise tire wear and squeal.

As tires in racing conditions are always at a slip angle ( difference between front slip and rear slip angle defines over/under steer) ideal geometrical point for Ackerman steer will vary during a corner depending on slip condition, which in itself is dependent on vertical load and speed and camber and tire pressure... Tire cornering force versus slip angles are never linear, and vary also to parameters mentioned before.

So classic Ackerman steer or reverse ackerman would be correct at only one point of trajectory through corner.. hence feverish arguments over merits.. Everyone has his pet theory and rules for best performance, but you will find it so dependent on other factors as to make it almost irrelevant. whichever choice you make ,initial toe out or toe in will influence car behaviour in transients much more than ackerman effect itself , for unless you have a fork lift truck or London cab angles of steering the divergence of steer angle between the tires is very small at usual lock angles, at Loewes certainly , at la Source in Spa also. Elsewhere definitely a second order effect.

As to bump steer very few cars have perfect bump steer in lock , even if zero/zero in straight ahead condition, if you try to do a three bar compatibility graph , then move the third bar as a rack end would move you can see the problem.

Closest solution is indeed top link (or bottom ) in same horizontal plane as track rod. You must also bear in mind effects of elastokinematcs ( as mechanical components deform under load, thus changing their geometry), the Porsche 917 had enormous bump steer (measured statically) but when under load during corner this compensated for chassi deflections (rack mounts, pickups etc).. use of bump steer can help tune car behaviour the same way , by modifying steer( hence slip ) angle of each tire (ft & rr) to desired characteristic.

[Patrice L'Rodent]

A simple way to determine whether pro or anti Ackermann is required is to test with toe in and toe out. If the driver reports toe in is better in a fast corner and toe out is better in a slow corner, then the car needs positive Ackermann (and vice versa)
The logic behind this is that the car requires more steering angle in slow corners than in fast corners, so in the example above, the car would be set with some toe in (to suit the fast corners at small steering inputs) and positive Ackermann so the car develops toe out in the slow corners.
This is a quick and dirty way to determine what you need without knowing the tyre curves.
In fact, a little reverse engineering would then allow you to get a pretty good handle on the tyre curves.
Incidentally, Langensperger developed the geometry to prevent the gentlemens carriages of the time from tearing up their manicured gravel driveways. No one seemed to worry about the hoofprints, or the 'apples' the horses leave behind.
Pat D'Rat

[LMP900]

One important effect of lots of toe-out (which can be generated by Ackermann) in a corner is the direction of the force vector at the contact patch. With increasing angle of steer on the inner tyre as compared to the outer tyre comes increasing drag, which gives a yaw moment into the corner. That's probably the biggest effect of Ackermann and static toe-out on the front wheels. It's useful to improve transient response, but is not as good in steady-state cornering. However, there's not too much off that in circuit racing.

[red300zx99]

First it must be stated that the ways to create ackermann geometry are numberus from tie rod lengths to spindle dimensions and im sure probably a few more ways. But I have to question Mr. Glover in asking how caster would effect the toe. Yes caster creates a self-aligning torque with the tendency to straighten up with the line of tragectory, but this torque doesnt change the heading of the wheel if the steering angle is kept constant. So I guess my question comes from how would a tire change direction(not slip angle) if the tie rod is kept in the same place?

OK now to more ackermann. The word according to Milliken.

For low lateral acceleration usage it is common to use ackermann geometry. This geometry ensures that all the wheels roll freely with no slip because the wheel are steered to track a common turn center. Note that at low speed all wheel are on a significantly different radius, the inside front wheel must steer more then the outer front wheel.

High lateral accelerations change the picture considerably. Now the tires all operate at significant slip angles and the loads on the inside track are much less then on the outside track. Looking at tire performance curves, it is seen that less slip angle is required at lighter loads to reach the peak of the force curve. If the car has low speed geometry (ackermann), the inside front tire is forced to a higher slip angle then required for maximum side force. Dragging the inside front tire along at high slip angles(above the peak lateral force) raises the tire temperature and slows the car down due to slip angle(induced) drag. For racing, it is common to use parallel steering or even reverse ackermann.

[desmo]

More random thoughts re Ackermann, epure de Jeantaud, Langensperger angle, dynamic toe out etc:

Do we really want slip angles for both front tires to be at the peak of the tire curve simultaneously? Might it make the car easier to drive- and perhaps faster in the real world, even at the expense of some ultimate lateral grip, to have one front wheel at slightly below it's lateral force peak to 'soften' the car's response to breakaway?

And along these lines, given that most road circuits are primarily either clockwise or anti-clockwise, which is to say that the sums of the turning angles for a lap left and right will be unequal, might it make sense to perhaps have laterally asymmetric steering geometry to suit? And might it make more sense perhaps to use designed in roll steer geometry to achieve this?

I've got more crazy questions, but I better stop here for now...

[RDV]

If the car has low speed geometry (ackermann), the inside front tire is forced to a higher slip angle then required for maximum side force. Dragging the inside front tire along at high slip angles(above the peak lateral force) raises the tire temperature and slows the car down due to slip angle(induced) drag

.

Precisely, in this case the drag vector of inside wheel is giving you a very strong yaw couple that makes car turn in ( same reason we use toe out to improve response) we are effectively not using maximun cornering power from this wheel (inside) but are using it to sharper response at turn-in and mid corner phases, as succintly explained by LMP900=

One important effect of lots of toe-out (which can be generated by Ackermann) in a corner is the direction of the force vector at the contact patch. With increasing angle of steer on the inner tyre as compared to the outer tyre comes increasing drag, which gives a yaw moment into the corner. That's probably the biggest effect of Ackermann and static toe-out on the front wheels. It's useful to improve transient response, but is not as good in steady-state cornering. However, there's not too much off that in circuit racing.

..as for laterally assym steering lay-outs, yes!! incorporate computer controled closed loop steering strategy responding to tire perfomance into the AtlasF1 blue sky F1 car, or is that thread dead now due to lack of sponsorship? especially as we can also use this on rear axle too, effectively having this on all 4 wheels.... have worked with this in ice racing cars (4 wheel steer)and it brings some interesting results

[BRIAN GLOVER]

By all means, call me Brian. I'm not sure what you mean by keeping the steering arm in the same place.
Depending on where the steering arm is located on the upright and its relationship to the location of the wrack and the length of the tie rod, the caster angle will have more or less of an effect on toe when the steering wheel is turned. With 0 caster, the heim joint for the tie rod will move up and down when the upright is turned from lock to lock because of ball joint inclination. Change the caster and the vertical movement becomes greater. In the +or - 70 degrees of travel from left to right, the heim may move 2" vertically. The loci of the heim joint can be quite complex also. On each plane it goes thru a radius and will cause a toe change. It is possible to alter this steer by design, but you wont be able to eliminate it completely unless you incorporate an additional coupling on the lower A arm like on the Honda NSX and the rear suspension on the Renault F1 car or have a steering system like an ox wagon.
We are not talking about bump steer here, however, all the geometry of the suspension is coupled some way or another. In the straight ahead position in full bump or droop, it is possible to almost eliminate bump steer, but suspension travel off centre is a different story and will be effected by caster amongst other things. Thats why I asked what the steering arm looks like on a F1 car with the tie rod that high.

There are so many variables, it would be better to plot it out on your Mechanical Desktop and move it around and see what happens.


A different note: I looked at the new Indy cars at the Indy Museum and noticed that the centre of the spindle or wheel bearing was far forward of the ball joints by about 2"and about 11' caster. Turn the steering and the car moves from left to right. Ive heard of scrub radius, but what would you call this?

[QUOTE]Originally posted by red300zx99
[B]First it must be stated that the ways to create ackermann geometry are numberus from tie rod lengths to spindle dimensions and im sure probably a few more ways. But I have to question Mr. Glover in asking how caster would effect the toe. Yes caster creates a self-aligning torque with the tendency to straighten up with the line of tragectory, but this torque doesnt change the heading of the wheel if the steering angle is kept constant. So I guess my question comes from how would a tire change direction(not slip angle) if the tie rod is kept in the same place?

[Terrance W. Peterson]

Castor does effect toe-in it is a simple geometric relationship. ackerman is usualy acheived by the relationship of the steering arm position to the rack and the vertical link the most important relationships are castor, steering arm position ,rack position. Michael Costin,David Phipps in there racing and sports car chassis design 1961 page 95 lay out the basis of ackerman, and suspension layout. Offfield

[LMP900]

For racing, it is common to use parallel steering or even reverse ackermann.

It may be common, but it's by no means universal. I've seen more than 100% ackermann on single-seaters: it's certainly a blunt instrument to improve turn-in, but if it gets you a faster lap-time who cares?

[red300zx99]

Brian haven't thought much about your caster info yet. But back to the ackermann as referred to LMP900 and saying that he's seen 100% ackermann in some single seaters. Yes it may be widely used, but that doesn't mean it right. Desmo said

Do we really want slip angles for both front tires to be at the peak of the tire curve simultaneously? Might it make the car easier to drive- and perhaps faster in the real world, even at the expense of some ultimate lateral grip, to have one front wheel at slightly below it's lateral force peak to 'soften' the car's response to breakaway?

Sure it may be easier to drive, but the better the driver the better he will be able to drive this car at the limit better and produce quicker lap times.

[LMP900]

Yes it may be widely used, but that doesn't mean it right.

The great thing about motor racing is that it's easy to judge what's "right" - if it wins, its right: and I've seen 100%+ Ackermann win. Sad but true.
Having said that, I don't use that much!

[red300zx99]

Factor in orther things such as luck and other chassis differences and the win might be a loss. I guess is sort of like the dog turd theory, after 3 weeks of winning everyone will have a dog turd on the hood. But from the perception of physics and the possibilities of physics, which is better? Mosh Pit was a proper name

[LMP900]

If the dog turd helps your turn-in, why not?

[red300zx99]

But what if the dog turd makes turn in easier, but reduces the limit, screw the driver, if the driver were perfect and could control such a beast, which makes turn in better?

[RDV]

...... I did say this discussion would be akin to decyphering ancient religious scrolls.... strong positions will be taken and harsh words exchanged...

LMP900 -- quite agree , pragmatism rules, I would carry a elephant turd on bonnet if it goes faster

red300zx99

But from the perception of physics and the possibilities of physics, which is better?

and

, screw the driver, if the driver were perfect and could control such a beast, which makes turn in better

?

..I have always been a strong advocate of using correct engineering, which is a subset of physics, and also agree on knowing the reason WHY things happen, but a t the race track , with the driver and car you have , and also given that ackerman is but a small item of a very complex phenomena ( it involves vehycle dynamics,(including tire carcass effects,of which we never have enough info) driver physiology and psychology, sometimes putting the dog turd on does make it have a faster lap time....

[LMP900]

RDV and I are singing absolutely from the same hymn sheet. At the design stage, you aim for the theoretical optimum, but as a race engineer you know that circumstances alter cases. It's at that stage that the dog (elephant, wombat, etc.) turd becomes an attractive option.

[red300zx99]

So what, still no concrete answer?

[Greg Locock]

One of the people I know has built and raced many vehicles over the years. His conclusion re Ackerman was that it was almost the least important kinematic characteristic for a circuit car. He now builds his cars with parallel steer.

In contrast for a road car that sells widely as a taxi the Ackermann is crucial, but, to back up some of the comments above, the actual Ackermann we get is not just the simple ' point the steering arms at the back axle'. Having said that, I haven't actually had to do that bit of the design yet (and the way things are going I never will) so i don't know how the Ackerman gets set.

[Greg Locock]

Well that one came back and bit me on the bum.

I've just spent four weeks setting up hardpoints for a new geometry, and Ackermann has been one of the hardest things to get right. I've been looking at plots of steering wheel torque vs angle trying to get them right at all speeds.

In a race car what you might find is that Ackermann affects the frequency or speed at which you get torque reversal .

Pat's way of evaluating it seems spot on to me.