twohoos

Discussion in another thread quickly went out of scope, so I'm starting this thread.

The question at hand is how to quantify the effect of suspension compression (including lowering springs) on the S2000's roll axis, and thereby its potential effect on weight transfer.

When we left off, nmrado had noted that:
To continue...

Hmmm...I was expecting the lever length to change more than the amount the car is dropped (e.g. if lowered 1 in., roll couple increases 2in. or something).

Maybe we're talking about different things? I'm thinking of the difference between the CG height and the roll axis height (CG height is a single value, but we want roll axis height at 3 locations: front track, CG location, and rear track). Lowering the vehicle by an inch lowers the CG height an inch, but typically lowers the roll axis significantly more.

Illustrations by J's Racing, taken from Rob's (maxrev's) site: Point A is the CG, point B is the roll axis (point P is the "instantaneous roll center" which is needed to find point B):


When the car is lowered, the difference btw A and B can be much greater than just the amount of lowering (figure exaggerated for clarity):


The reason we care is that for a given lateral acceleration, the amount of weight transfer is directly proportional to the height difference between A and B (see Milliken & Milliken, Puhn, or similar).

Note that the same logic illustrates the general need for a progressive spring rate: as the suspension compresses, the chassis gains leverage, thus reducing the springs' effectiveness. (This effect can be countered if, for example, the springs' "angle correction factor" increases during compression...but in our case the ACF is already 0.94 or so, so it doesn't have a lot of room to grow.)

Anyway, to the point: I think we need some measurements of CG height and front & rear roll center heights. Volunteers? Discussion?

twohoos

P.S. Determining the CG position is *fairly* straightforward if a set of chassis scales is available...see Milliken/Puhn again, or:
http://www.unece.org/trans/main/wp29/wp29w...RSG79_inf11.pdf
http://www.jameshakewill.com/cg-height.pdf
http://www.longacreracing.com/articles/art.asp?ARTID=22

robinson

I'll be contributing here eventually with data from WinGeo. I just need to get another corner balance and an alignment first.

User 121020

Twohoos, we're talking about the same thing. The graph I posted in the other thread was just for the front geometry and it was just a rough stab at determining the roll lever length (CG height - roll center height) for the front.

For the CG height, I used the number (~26.5") from Rob Robinette's website, S2000 CG Height, as my starting point.

I was also surprised that the roll lever length was fairly insensitive to chassis height changes.

As mentioned in the other thread, I plan to take measurements of the chassis once it warms up a bit. The idea is that the model I build with the measurements and the WinGEO model that Robinson builds should produce the same results, which should improve the validity of the two models.

robinson

With one exception, I have SPC ball joints on each corner, so that will change the results a bit.

User 121020

Right, our respective modifications will change the results a bit.

Robinson, are you going to take measurements off of your chassis or do you just want to tweak the numbers I come up with so they're consistent with your setup, e.g., the SPC ball joints?

As of right now, I'm planning to put my car on jack stands, tape down some quilters grid to my garage floor and use a plumb bob to measure locations on the grid. The plumb bob will also be used to capture the vertical distance from the ground. I know my garage isn't level, but it appears to be fairly flat. The induced error shouldn't be too bad. Anyone have any better suggestions as to how the suspension points should be measured, short of using an articulated-arm CMM?

twohoos

nmrado, interesting! So you estimated the front roll center height to be just 1 to 2 in.??

robinson

The program uses 14 decimal accuracy, so even having different alignments will change the results slightly. I think I will find out the diffenece at the spindle with and without the SPC ball joint, so that I can analyze both setups in Wingeo.

ZDan

This is not true. The amount of "weight transfer" is simply lateral acceleration multiplied by c.g. height above the GROUND, not the roll center.

Total lateral weight transfer is weight multiplied by c.g. height above the GROUND, divided by the track.

If the roll center is made lower, then you will get more body roll, but you WON'T get greater weight transfer at the same lateral acceleration.

User 121020

A very rough estimate, yes. Finally found the image I was talking about...it's in this thread: Front Suspension Geometry.