Chassis Twist

Hello RC,

with linear potentiometers in parallel to the damper there are obviously a number of purposes to use them for. For this question I am referring to determine the modes of the sprung mass, meaning pitch/heave/roll and chassis twist. Latter is where I struggle at the moment because of two quotes I got from reliable sources:

1) "You can have different front and rear roll without having the chassis twist"

2) "Assuming a rigid chassis front and rear roll angle are the same"

Could somebody please clear that up for me?

How do I determine a "approx. rigid" chassis based on potentiometer data?

Thanks in advance.

- Seb

Hmm, would you be able to link the the sources somehow?

For 2) If the chassis is infinitely stiff, it's not really possible to have different roll angles under generally normal circumstances. If you transition to a cambered road or something I suppose it's technically possible because the reference plane is forcing a different angle, but I doubt that's what you're asking.

For 1) Are they referring to different front and rear roll rates? That doesn't sound correct

To determine "approx. rigid chassis" I would assume you want verify that left and right, or front and rear stay parallel to eachother when loaded from braking, cornering, or accelerating. So I would 0 it all on flat ground, go around a skid pad as smoothly as possible, and verify that both insides, and the outsides produce the same results. I've never tried this, but it sounds like a really neat idea to get an idea of how stiff the chassis is.

There may be some translation problems, or possible misunderstanding or confusion between roll resistance, body/chassis roll, roll centres, etc.

As Robbie said, for a truely rigid chassis the 'roll' at the front and back must be the same.

Been a while but, as I understand it (and I strongly recommend you checking further), torsional (twist) resistance in a chassis is measured in the couple (torque) required to twist it through 1 degree, relative to the spring loading points on the chassis. As you can imagine, 1 degree is quite a lot, and may be damaging to the chassis if there's a weak point.

I'm not sure how to do any useful measurements with a potentiometer (I assume you mean as used for suspension travel logging?) - strain gauge, maybe - but if you have some corner weight scales, or even a good bathroom scale, you may be able to get an approximation.

NOTE, not sure how well this will work, so very open to discussion!

Method A.

Remove suspension and support vehicle on stands on the heavier end and one light end corner, with a small screw, or bottle, jack under the fourth. Level the heavy end of the chassis, doesn't need to be precise. You will also need the distance between the chassis spring mounting points on the light end.

Make the initial measurements of corner plate loading with jack lowered so chassis supported on three corners. Note level angle, or measurements, at the partly supported end to establish initial flex, if any, and the jack' plate loading (there should just be the weight of the jack) . Now use the jack to lift that corner until the other side is just touching/lifting from it's support and repeat the angle/distance measurements and load on the plate.

You now have either two angles, or measurements (with the chassis points distance), which will tell you what the deflections are to calculate the change in angle, which is what we normally want. You also have the forces applied to the chassis points at both angles and the distance between them which means you can calculate the torques applied and so the actual torque (moment) for the angle.

You can now correct to what moment is required to twist through 1 degree.

Method B uses the bathroom scales instead of the corner plate.

If one has access to a chassis machine, or large rigid platform, one might wish to fix the three points to that, as more force can be applied and/or the chassis can be measured at more points, to establish where the main flexing is occurring.

Hi Sebastian,

This is a fair question you're asking, here are my thoughts on this.

For a theoretical case of a simplified rigid plane, you're right in thinking the front and rear roll angles must be the same. However, in reality there are plenty of reasons you can and will measure different roll angles front and rear, the ones I can think of right now:

-Torsional deflection

-Compliance in chassis/suspension/hardpoints

-Measurement and calibration error in the damper pots

-Vertical deflection of the tyres

The last point about the tyres is probably the most significant of these points I've listed for most situations. Remember that when we "measure" the roll angle with damper pots, we're only roughly approximating it as we're only measuring the deflection between the sprung and unsprung masses. In many cases, the vertical deflection of the tyres is quite significant and will absolutely be different on each corner or the car. So even just with the effect of differential tyre vertical stiffness alone, you'll get different "measured roll angle" at each end of the car when using damper pots.

For all the reasons I listed above, there's no way to measure chassis torsional stiffness using damper pots. Any twist in the chassis will be swamped by these other effects. To properly measure torsional stiffness, you need to attach one end of the chassis to something extremely stiff and large (usually a large steel I-Beam or similar) then you put a pivot at a known point and twist the chassis around this point. You measure the deflection with a micrometer. There are different variations of this type of testing, using suspension in place, rigid dummy dampers, no suspension at all, but it's all a similar idea.

Below is a diagram showing a setup like this.

Hope some of that info helps you out,

Tim

Thank you guys for the great input! It really helps a lot.

So in conclusion laser ride height sensors are actually an essential tool when measuring roll whereas I can imagine road surface imperfections are going to play a major role that needs to be considered.

@Robbie: The sources I was referring to were Optimum G and Milliken RCVD.

Regards

Seb

No problem Sebo,

Yea laser ride height sensors are definitely a great tool. However, in my opinion, they still most useful when used in conjunction with damper pots. Using them in conjunction helps you understand the relative contributions of the sprung to non-sprung deflections as well as the vertical stiffness of the tyres. Both are important!

I would also add, in most series laser ride height sensors aren't allowed, so making use of your damper pot data is often still critical as it's all you've got!

Tim

Yup, you've got to use the tools you have available.

Tim brought up the suspension and that's a very important factor that's often overlooked, not just in the bushing deflections, if used, but the actual suspension parts themselves. I overlooked it myself, as the loadings through the hard points (where the suspension is fixed to the chassis) is not a simple vertical force, but will be applied through them all, at different force levels and angles. On that, don't forget the lateral and longitudinal loadings can also be much greater, as the 'G' loads through the tyre contact patches will be loading the chassis in those dimensions as well.

In the illustration above it is also checking the torsional variation of the chassis AND the suspension - if used with a chassis-only set of measurements it may indicate suspension components that need to be re-designed or modified. You may even want to set up a test fixture for the other loads, perhaps with a Porta-power, as with some vehicles parts can flex enough to cause geometry variations - IIRC, this was an issue with Escorts running slicks, back in the day, as the strut flexed, and was the driving force behind the redesign of struts with larger diameter, hollow shafts with internal dampening. This is most likely going to be a concern with high wheel to spring/damper ratios with higher rate springs/damping.

Oh, and don't forget the ARP loads, too.

Alright so basically we are talking about the total compliance of the entire construction here while identifying the weakest points or rather the elements influencing the measurement of chassis twist angle the most.

I assume calculating the exact loads should be conducted with a suspension design software as different 'Gs' will result in a variation of the force application points (see following slide by Danny Nowlan of Chassis Sim).

By experience OEMs make use of K&C rigs and strain measurement strips for validation purposes.

However my intention is to get a good idea of chassis twist / suspension compliance without getting too expensive.

So, after determining the x/y/z-direction loads, wouldn't it be easier to mount the suspension assembly on a fixed plane (representing the chassis) and measure the deflection?

Regards

Seb

Seb, calculating the internal forces inside each suspension arm to a useful level of detail is relatively simple using a few simplifying assumptions and some vector maths. As you say, you would come up with the different loading conditions - different levels of lat/long G force and different combinations that represent the different worst-case scenarios. You can then calculate the resultant loads in each direction for each component/hard point.

After that comes the validation. So this is where K&C rigs come in, certainly, these are using in the OEM world but also in high-end motorsport very extensively. For anyone without access to a K&C rig, the next best is to instrument the car with strain gauges and record the real deflections/stresses you get on track.

Thanks for clarifying, Tim :)