Hello,
Great post. I just wanted to contribute a little (as much as my poor English will allow)
First I think that it should be beneficial to outline why as stiff as possible chassis is desired.
Under cornering conditions, when vehicle rolls weight transfer distribution (and so tire contact patch loads -> handling characteristics) is determined by 3 springs in series: Front axle roll stiffness, rear axle roll stiffness and a torsional spring in between them - chassis.
We do have control over dynamic weight transfer dist. via altering front/rear roll stiffness by changing springs/bars motion ratios etc. etc. Sad part is we don't have control over stiffness of mid torsional spring (chassis) - unless we redesign it. Also when 3 springs work in series it's not very efficient to increase stiffness of 2 of them if one is weak. Weakest spring will always limit the final stiffness.
We want dynamic WT distr. to be as close as designed (front to rear stiffness) but when middle torsional spring is added than designed weight transfer percentage is altered. Simply said car is becoming less sensitive to setup changes and so it's harder to find optimal setup.
This was about steady state cornering. How ever during real life dynamic events car very seldom if ever will see pure steady state condition. It will more likely see a combination of lateral and longitudinal weight transfer. Also there are various bumps and all other sorts of surface irregularities. Car with flexible chassis may handle very unpredictably and it's very hard to quantify needed setup scenarios to achieve desired handling characteristics.
Another question - how much is enough?
I personally believe that there's no such thing as too stiff a chassis. How ever I do believe that there can be to heavy chassis. So usually designer is constrained by minimal allowed weight (in race series they run) and all other sorts of limitations - materials allowed, CGH, safety regulations etc. etc. (in no particular order of importance)
Usually it's said that desired stiffness is dependent on difference in front/rear roll stiffness. How ever modern studies show that this is not the case. More so needed stiffness is more of the function of overall roll stiffness and front/rear weight distribution ratio. The bigger are those the bigger stiffness is needed so to have real dynamic WT distr. close to what would be governed from front/rear roll stiffness and absolutely rigid chassis. There's a really nice SAE paper that describes this in significant detail (I'll look for it's number in case someone is interested)
I have to add here that not only chassis torsional stiffness is important but also bending stiffness and suspension installation stiffness.
I'd like to add a couple of cents to great post from Kalun_D. To get realistic numbers out of your physical test is very important not to overconstrain the chassis - this will cause it to appear stiffer than it is. Chassis should be constrained that way so only 6DOF (degrees of freedom) are constrained.
Also numbers indicated for NASCAR chassis are right - 8-9000ft/lb. deg is usual numbers for chassis "as is" bought by the team from chassis manufacturers - Hopkins, Laughlin etc. How ever teams (at list those running at the higher end of the field) rarely race it "as is" There is another very nice SAE paper about mods made to Hopkins chassis and it showed that it was possible to achieve as much as 232% increase with only 40lbs of additional weight.
Also I agree with numerous posts that monocoque is the way to go (if more complicated to design and manufacture) because loads are distributed much more efficiently. There where problems with feeding concentrated loads but techniques to deal with it are well developed. There was also problems with skins "oilcanning" and even buckling under load but after people realized that if a core (ally honeycomb gives best weight/stiffness ratio) is placed between to skins it helps enormously by stabilizing them - stiffness increases dramatically with very little weight penalty.
Also reason to use fancy materials like carbon fiber is clear - stiffness of woven HS carbon is close to that of Ally - 70000 Mpa. Density is much lower. Ultimate strength (of most importance in terms of safety) is higher than that of steel Av. yield strength of carbon steel is roughly 280 - 350Mpa and Epoxy woven carbon is 700 - 800! And when UD (unidirectional) carbon was introduced it opened an all new world of tools - stiffness was up to 130 - 140Gpa, and typical yield is about 2000/1300 (tension/compression). And than there's intermedium modulus carbon, high modulus and ultra high modulus with stiffness up to 600Gpa! (steel is about 200Gpa) How ever use of those high specialized materials is not as straight forward because strength goes down at about same rate stiffness goes up.
How ever an ally faced/ally honeycomb core monocoque chassis still could be made extremely stiff (and much cheaper) - only major drawback is one would struggle to pass current safety/impact regulations because of low yield strength of ally (150 - 250Mpa)
Sorry for such a long post - may be it could be interesting for some. Also sorry for my English and numerous repeating same terms/words.
Thank you
Ted
