Chassis Torsional Stffness

[johnc]

From the latest Mark Ortiz newsletter:

 

FRAME TORSIONAL STIFFNESS

 

Recently I was asked to assess a northeastern dirt modified car which has the tendency to consistently crack one particular welded joint in the tubular plus 2 x 4 mild steel chassis.

 

Attached are two bitmap image files from the two leading northeastern modified chassis builders

[not included here as they are copyrighted catalog images]

. The car in question is not one from either of these builders but is essentially the same in design. It is actually designed to be more flexible, or that is the claim of the builder.

 

I must confess that even though I have worked with these types of cars recently I had never really taken a step back and studied just the 'naked' chassis structure. All I see is a 'wet noodle' that only resists torsional loads on the sheer size of the material used. The 2 x 4 frame rails are required by rule but the rest of the chassis is just a 'stick built house'.

 

So my question to you is why do the folks who build and race oval track dirt cars still insist, rather strongly, that a flexible chassis is the 'hot ticket' to get the car to turn in to the corner? That is the claim I hear the most in defending flexibility.

 

I was trained to believe that you go as stiff as you can go up to the point of incurring a weight penalty and adhering to the required safety rules.

 

The claim is that the flexible chassis is easier to tune. I agree because to me it is to a certain degree un-tunable with an undamped fifth spring doing the work.

 

The only drawback to going as stiff as possible with a dirt car chassis that is required to use a beam axle front and rear is that, in my opinion, it would narrow your setup window and in effect, yes, be a little harder to zero in on a fast setup with just springs, shocks and anti-roll bars.

 

One additional piece of information: the engine is not used as a structural member and in fact

 

besides the engine plate which fills the bay below the A-pillars, the mounts are designed to promote engine movement as well.

 

I really need a sanity check on this one. What is my counter argument to the 'flexi' folks?

 

The frames shown could be less triangulated than they are, actually.

 

Dating from carriage building, vehicles with beam axles at both ends have used torsionally flexible frames with decent results. Most trucks are still made this way. They don’t even use poorly triangulated space frames; they use channel-section ladder frames. The stresses in the frame rails are such that they are heat-treated, and frequently carry cautionary labeling discouraging welding on them lest the heat treat condition be altered.

 

Some people may be convinced that there is a performance advantage in a torsionally flexible frame, but another reason frames aren’t triangulated better than they are is packaging. The diagonals required are hard to find room for. Things like the engine, the driver, and the exhaust system tend to get in the way. With more diagonals, the car becomes harder to work on and to get in and out of.

 

I am not convinced that a torsionally flexible frame is an advantage, for turn-in or otherwise, but with softly sprung beam axle suspension at both ends, it doesn’t make much difference because the torsional loads on the frame are small.

 

A torsionally flexible frame does make the total chassis more warp-compliant, but as the questioner notes, the compliance is largely undamped.

 

Excessive torsional flexibility can give rise to torsional oscillation. One commonly accepted rule of thumb is that for the suspension to work as it should, particularly as regards damping, frame twist should be no more than a tenth of the total compliance in the warp mode, suspension and tires included.

 

With an existing car, we can test this. We need a flat support to place under one wheel, or alternatively two of these for two diagonally opposite wheels. These should be large enough to fully support the contact patch, and tall enough to create significant suspension displacement, yet not exhaust the suspension travel at any corner of the car. When we set the car on the support(s), we create a known displacement at the contact patches. If we are working in inches, we can get this in angular terms by dividing 57.3 by the track width, then multiplying by the height of the support. Alternatively, we can measure the angle directly with a long straightedge and an angle finder. If we do this at both ends of the car, the angular warp displacement is the sum of the front and rear displacements. We then know how much warp, twist, or cross-axle articulation we have at the contact patches.

 

We can measure the frame twist by measuring angle from gravitational vertical or horizontal at any reference surfaces at the front and rear of the frame, first with the car on level floor, and then with

 

the car on the support(s). The

change

in the front/rear difference is the amount of frame twist under the load imposed. The frame twist should be no more than a tenth of the warp at the contact patches.

 

It will be apparent that stiffer springs, bars, and tires will call for a stiffer frame, to meet this requirement. A suspension that uses high roll centers to get its roll resistance, rather than relying on the springs and bars, will not require as much frame stiffness.

 

Is a car with a flexible frame easier to tune? Well, it is less responsive to changes in springs, bars, and even roll center height. It is therefore less prone to being overly touchy to small adjustments, but cars with compliant tires and soft suspensions tend not to have such a problem anyway. Indeed, even pavement cars with really stiff suspensions are also more often under-responsive to changes, due to insufficient frame stiffness, than over-responsive due to excessive stiffness.

 

A flexible car will respond differently to changes in track conditions than a stiff one. That doesn’t necessarily mean it will change less, but it won’t change the same as a stiff car. If the car corners on three wheels, like the one we discussed last month, frame stiffness does not affect wheel load distribution, past the point of wheel lift.