ANTI DIVE/SQUAT AS MOTION ABOUT PITCH CENTRE 2

[autogyro46]

First, I'm a newbie here. But, as a teenage lad, I constantly took out Costin and Phipps (first edition! when it was new!)from the library and never quite got over it. As an EE, I tend to look at suspensions in terms of common-mode and differential movements, and my questions may sound at times naive.
My question is the following; it strikes me that one can look at the behavior (OK, I'm a Yank) of anti-dive and squat geometry as geometrically governed motions of the CG about a pitch centre (the UK spelling tends to get better answers) much the same way that unequal length non-parallel control arms govern CG movements about a (notional) roll centre. If that is so, it should be possible to arrange the intersecting planes of the control arms so that they intersect well past the CG at a height at or perhaps above it.
One could then further arrange the two intersections so that the lines from the contact patches to each of them from the tires intersect at a predetermined point in relation to the CG thus forming a pitch centre that could be made stabile over some range of deflection, much like a roll centre
By moving the plane intersections way out, the effects of binding could be reduced, and one one could have a more stable geometry. In addition, one could place this centre very near the CG
Oh, I forgot the question: am I out to lunch?

 

[GregLocock]

I don't know about out to lunch, perhaps I'd be more inclined to say worrying overmuch about something that doesn't seem to be a big deal, at least so far as antidive goes.

The curves I see of pitch angle vs braking g, or traction g, tend to be very linear.


Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[autogyro46]

Aha! you've uncovered my second bugbear, which I thought was cleverly concealed by this one.

After mucking around with a low end suspension program, I discovered the magic of placing the control arm intersections at the centre line of the opposite tire: virually no roll centre migration. (childsplay?)

The cost, of course, is high camber gain: fun for corners, not so much fun for braking. Hmmmm. How to preserve the first and minimize the pain of the second.

Fight geometry with geometry! It's all just looking at the problem sideways. (What was it Maxwell said?)

I used to work with high-end loudspeakers, and it attracted the same kind of obsessional nutcases (get a life!)that I suspect I am being here.

Thanks so much.

 

[BrianPetersen]

I think you need to go out and crawl underneath a few real vehicles (and some examples, like open-wheel race cars, have everything exposed for all to see).

Having the instant-centre close to the wheel on the opposite side approaches Ford Twin-I-Beam truck suspension. It was a great innovation in the 1960's and was better than the solid axle leaf sprung front ends used up to that time. Nowadays, it's not exactly the greatest thing since sliced bread, and Ford has gone away from it (I'm not sure if the Ford Econoline vans still use it - if so, they're the last one). The Twin-I-Beam can be regarded as a less-extreme example of swing-axle suspension, which was more commonly seen in the rear (original VW Beetle, original Corvair used this).

Modern designs have tended to use much less camber change with wheel travel and control the body roll with spring rates and antiroll bars so the camber that occurs because of body roll is less pronounced.

Having a strong caster angle in the front end geometry tends to tilt the outer front wheel in as the steering turns in.

Careful location of pivot axes can achieve an anti-dive effect when the brakes are applied, so the camber change and roll center migration when on the brakes aren't as big a deal anyway.

The vertical migration of the roll center is not a big deal as long as the vehicle is close to its nominal ride height anyway. This is why extreme lowering of vehicles with MacPherson or A-arm suspensions causes a lot of trouble unless the geometry is corrected by altering the location of the pivot points. Plenty of import-vehicle "tuners" want to achieve the race-car look and go-kart handling feel, but they do it without relocating pivot points (too expensive) so it ends up with crap ride and handling.

 

[autogyro46]

Brian;
Believe it or not, I actually remember the twin I-Beam (it's long gone)and I recognized the similarity instantly.

Apparently the original Porsche 911 had a camber gain of 0.8 deg/inch, and I saw somewhere I saw that at least one of the Miatas had 0.91. I'm sorry, I can't see these gain sensitivities by looking at them directly; my eyes aren't that good.

Anyway, that wasn't the central point. When you have two moving points which can articulate around a common structure in one plane, their respective articulation geometries in another plane can describe the movement of the sructure in that plane in a manner similar to the first

Anti dive traditionally describes the behavior of the front suspension in the longitudinal plane with respect to the CG only, and neglects the rear. Likewise, in the reverse case, anti squat. This is simply another way of looking at the phenomonon.

 

[NormPeterson]

"After mucking around with a low end suspension program, I discovered the magic of placing the control arm intersections at the centre line of the opposite tire: virually no roll centre migration. (childsplay?)"

Quite possibly the program is based on the assumption that the sprung mass actually rolls about the "geometric roll center". Quite possibly that's an inaccurate assumption.

Unless driving fast through significantly banked turns is a big concern, lateral migration of the "roll center" may not mean very much.



Norm

 

[autogyro46]

Thanks.
The program isn't even that smart. It doesn't know about physics. It only knows about "geometry". (That's why I used the term "notional".)

 

[BrianPetersen]

It sounds like the program you are using is too simplistic for what you are trying to think about.

The concept of a "roll center" isn't even valid as soon as you have uneven loading of the left and right sides - which happens as soon as you have any significant cornering at all. At the extreme cornering limit, the instant-center of the outside wheel can be regarded as "mattering", but there can be so little loading on the inside wheel that it scarcely matters what happens on that side.

MacPherson struts have a disastrous camber curve and instant-center movement; many of them start tilting the wheel the wrong way beyond a certain position of bump travel. But, if the geometry is correct, they work pretty good, at least for the front end of an ordinary grocery-getter vehicle.

 

[autogyro46]

Not to mention a lot of current Porsches (including the Cayman S)

Again we're only talking geometry here. Lateral weight transfer is a whole 'nuther ballgame, especially when it is tranferred via combination of various compliances, which can act differently, especially in the time domain.

Re inside wheel loading, I'm about to start another thread
on ARB's, specifically "U" bars vs "T" bars, which work completely differently.

 

[peuteu]

one good idea is to use hydraulic suspension self levelling suspension; for exemple citroen activa

http://activaclubfrance.free.fr/brevets.html

 

[autogyro46]

Ahh...ignorance is, well, if not bliss..ignorance.
Little did I know when I started this thread, that almost simultaneously, there was a Battle Royal (well, a lively discussion) on this very topic in the recent pages of Racecar Engineering, of all places,(March, Sept, and Oct '09) between the august Mark Ortiz, and the equally august Danny Nowlan of CarSim, of all people.
I haven't completely digested it yet, (I just downloaded the articles through Zinio) but the debate seems to about whether Nowlan's Force-Based model accounts for the simultaneous interactions of both front and rear force lines with the C.G.
As near as I can make out, my neophyte's view of the behavior seems to be closer to Nowlan's: I confess I have difficulty with Ortiz's "resolution lines".

Any thoughts, guys?

thanks

 

[GregLocock]

It's only RCE, don't fret. My take on having been involved with modelling the roll behavior of cars, is that any sensible scheme of analysis is good enough for predicting changes, but there are many gotchas if you try to predict accurate (say to better than 0.3 deg/g in 4 deg/g) values from first principles. I suspect strongly that the same applies with the longitudinal acceleration effects.


I'd expect a successful approach would consider compliance effects (including the body), delumping of unsprung mass, and of course changes in geometry as the vehicle accelerates.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[autogyro46]

Thanks
On of the results to come out of my mucking about with roll centers, etc. ((with Performance Trends) is my belated appreciation of how "hard" the hardpoints need to be, or, conversely, that one should regard everything as a potential compliance, until you know they aren't.
There is a saying in optics that one must assume that everything one deals with is, in fact made out of rubber.

I assume that "delumping" starts with the pneumatic compliance in the tires, followed by the tire compound compliance, followed by bushings, followed in turn by the load bearing elements in the suspension.

A side note: I was going to a link a friend to your website's posting about using many Olympus jet engines as a booster stage for space vehicles, but Yahoo has apparently yanked the site. Is this page accessible somewhere else?

 

[GregLocock]

No, not yet. Yahoo killed geocities for their own inscrutable reasons. I will slowly add stuff onto my new website but it is slow going.

http://www.islandone.org/Propulsion/LACE.html

is a better bet



Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.