Roll center, weight transfer and confusion

[craigsimon]

Hi All,

I am reposting what I posted on v-eight.com to reach a wider audience

This relates to a chapter in "Competition Car Suspension" by Allan Staniforth

I have been reading the weight transfer chapter written by David Gould and I can't get my head around his use/assumption of a front and rear CG of the sprung mass to calculate moment arm for the front and rear roll center (and front and rear lateral weight transfer). The centrifugal forces will react through the CG and this can't be in multiple positions? can it? I can cope with his later calculations with a mean RC height at the CG location> I can understand how the different RC hights will have different resistance to lateral weight transfer with respect to "The" CG but not how you can have independent front and rear weight transfer components AND a mean weight tranfer component, that he seems to use in his calculations. Yet Gould is a highly successful car builder and I find it hard to doubt his thoughts based on his track record.
Carrol Smith also has a chapter on this concept in "tune to win" with a similar drawing of multiple CG's along the wheelbase of the car and his concept of CG wedge, is this now a superceeded view?

Comments, Help?

Craig

[MWP]

.... Head explodes

[ed]

courtesy link to other thread

http://www.v-eight.com/tech_forum/viewtopic.php?t=476

[RobertoX]

Hey craigsimon,

Its just a model, like most of our physical models that we use in physics and engineering etc. like Newtons laws, they work under certain conditions but do not tell the whole story.

It would be more accurate to model the sprung mass of the car with 1 centre of gravity and an axis of rotation as it rolls and moments of intertia around each axis etc but how far do you want to go? You could include the flexibility of the chassis, all the members attaching to it, the fact that the axis of rotation that the chassis takes wrt the ground will change throughout its roll, the rotational moments of inertia will change as driver/fuel etc get thrown around blah blah blah...
The chapter in Staniforth's book has obvious flaws in it but the approach has worked for many good race car designers. I suppose the thing to realise it its a good starting off point but there is more to it. F1 designers wont use this type of approach but its better than lowering the buggery out of your car and saying low CoG = fast.

I don't think Staniforth's book is that good, Carroll Smith's is a little better (but don't even go near Herb Adams ). If you want a good read check out Milliken & Milliken - Race Car Vehicle Dynamics or "The Multibody Systems Approach to Vehicle Dynamics" (I can't remember the author but it is good). But you can't really use the methods in these books with a calculator.

[craigsimon]

Thanks Roberto,

I have the Milliken book it is dense and is taking a while to digest. So that chapter is to be taken as a way of getting a decent base line geometry?

thanks
Craig

[RobertoX]

Yeah I think it is good for getting an understanding of what the different things do. If you want to play around with geometry a bit there is a useful package called susprog3D . If's fairly basic but easy to use and there is a free demo as well...

You need to consider where your roll centre is, where the roll axis is, the tyre camber curves of your linkage, steering geometry... this is all after deciding/calculating track, wheelbase, centre of gravity height etc...

Also it is useful to note that the stuff in that chapter is just steady state not transient, which doesn't happen all too often in real life (or in real track conditions anyway).

[merc-blue]

Is you problem that his uses a different COG for the front and rear?
I havnt read the article you are talking about so not sure exactly what your asking

[craigsimon]

he has an assumed/estimated COG for the rear of the sprung mass and a assumed COG at the front of the sprung mass and then generates a mean COG of the sprung mass which to me is the only place the COG can be? or if you were to estimate a front COG what components do and don't you include? and rear etc..?

Craig

[merc-blue]

Without reading the chapter I cant really help clarify the problem,


I am guessing that he has devloped a model/method that aids his designing process,

Look at other physics examples eg light, 2 different models that both work in there on conditions

[Crazyracer]

Its just a way of isolating the front and rear suspensions, so that the modeling for them can be simplified.

If you tried to recreate the real world situation you would be recalculating the position of the COG for every incremental shift in roll, dive or squat as well.

As MWP said "Head explodes". You'd need some very complex computer modeling to get a better picture of what is happening.

[craigsimon]

Thanks Crazyracer,

Back to my original question, I can understand isolation each end but won't they see the same sprung mass CG hight as there can only be 1 CG right?

[craigsimon]

Thanks to all that posted in this thread some excellent additional info on suspension design has been posted on this thread over at v-eight.com

link above but here it is again

http://www.v-eight.com/tech_forum/viewtopic.php?t=476

Craig