Chassis roll and yaw rate, lateral velocity

[WilO]

Wondering if anyone can shed light on how chassis roll might affect a vehicle in terms of yaw rate and lateral response....Maybe roll inertia is a better term. It would seem to me that roll angle and roll rate could actually have an affect on these characteristics. Any help would be greatly appreciated. Apologies if this a dumb or obvious question.

Wil

[Tim.Wright]

Since roll is a "movement" quantity (i.e. angle), the primary effects of it are also movements elsewhere in the vehicle. Secondary effects are force based.

Taking the primary effect first, for handling the most important of these are the wheel angles, particularly camber. Generally speaking (of course not always) roll moves the wheels in away from the ideal position in a corner in terms of camber. This will have an effect on the overall grip capability of the vehicle and possibly the balance too. The obvious solution here is to reduce roll to nothing. While this will improve your wheel angles, it will greatly increase the warp stiffness if not the overall stiffness of the suspension to the levels of that found in a go-kart which is no good for grip on anything but a completely flat road. In fact it will defeat the purpose of having a suspension in the first place.

Secondary effects are to do with the wheel loads. As the vehicle rolls it moves the springs and dampers which in turn change the wheel loads. This is how balance is altered using the springs and dampers. If a vehicle didnt roll, then changing spring or damper settings on one axle will do nothing to the balance.

For roll rate, do you mean roll stiffness or roll speed?

For roll inertia I would say it is not worth worrying about for a few reasons. 1. Nobody know the roll inertia of their chassis. 2. On a tightly packaged race-car, there is no way to change the roll inertia as a tuning tool 3. With the low roll angles and low weight of a race car, the transient phase is over so quickly that roll inertia effects would be insignificant compared to other characteristics such as springs and dampers.

Tim

[WilO]

Thank you for the reply Tim.
Unfortunately I worded my question poorly; I was thinking more along the lines of the relative phasing of these events, and how yaw rate and lateral velocity might be affected by roll movement, particularly if a vehicle is underdamped in roll. If we had an appropriately instrumented vehicle (LVDTs on the dampers, yaw rate sensor, etc.) how might the graphs look overlaid on each other.

The other effects you mention (camber gain) are things I am aware of. One clarification I'd ask for: Assuming a vehicle of finite track with a CG above the ground plane, I'm assuming that the wheel loads would vary whether the sprung mass actually rolled or not. The only way I can imagine it NOT rolling is the presence of an active suspension, which would be required to provide a support force to prevent chassis roll. Wouldn't this support force would change the wheel loading?

Thanks again for taking the time to reply.

Wil

[DaveW]

Once, many moons ago, I was involved in setting up the suspension of a (Passive) reasonably high performance luxury vehicle. The initial set-up was considered acceptable, except that (subjectively) the front loaded wheel collapsed under sharp cornering inputs. My initial reaction that the front compression damping needed tightening (not good for ride). One of my more able colleagues suggested, as an alterative, we should try stiffening the rear rebound damping. We tried both. The interesting thing was that both worked, & both (again subjectively) reduced body roll significantly. No guesses about the option preferred....

[Jersey Tom]

Chassis roll does impact the responses of yaw rate and lateral acceleration... through a variety of mechanisms. Roll-camber, roll-steer, relative delay of load transfer are the big ones.

That said, all of these are generally second order effects. Other things tend to dominate... until you get extreme.

A diabolically under-sprung car would be one example... where the response rate of load transfer is on the same order or slower as the driver control inputs. Will have a very nonlinear, unpredictable feel as the driver chases it around. Might even scream that the car is loose or oversteering everywhere... even if the steady state trim is understeer.

Of course you can throw all the roll stiffness in the world at a car and still have --- yaw response because you've completely missed the boat elsewhere - particularly in having crap or mismatched tires.

Which is why it is critical to have drivers that can separate all the responses in their head, and why a good race driver does not necessarily make a good test driver.

[Tim.Wright]

Thank you for the reply Tim.
Unfortunately I worded my question poorly; I was thinking more along the lines of the relative phasing of these events, and how yaw rate and lateral velocity might be affected by roll movement, particularly if a vehicle is underdamped in roll. If we had an appropriately instrumented vehicle (LVDTs on the dampers, yaw rate sensor, etc.) how might the graphs look overlaid on each other.

Ah ok, you are talking about transients here. I dont understand transients very well. Best someone else answers that.

The other effects you mention (camber gain) are things I am aware of. One clarification I'd ask for: Assuming a vehicle of finite track with a CG above the ground plane, I'm assuming that the wheel loads would vary whether the sprung mass actually rolled or not. The only way I can imagine it NOT rolling is the presence of an active suspension, which would be required to provide a support force to prevent chassis roll. Wouldn't this support force would change the wheel loading?

Wil

Yes, the loads will change regardless of whether the chassis rolls or not, but you cant control the distribution of the load transfer front/rear unless there is some movement to displace the springs. Since roll angle is the same front and rear, that means Xdeg of roll will distribute load depending on the relative roll stiffness of the front and rear axle.

Also, its possible to have a car that doesnt roll. If you set the IC of each suspension high enough you will have a jacking force that equals the load transfer and the end result is no movement of the suspension. This is commonly seen as "putting the roll centre at the CG". It is never done. Ive never heard a decent explanation why though.

Tim

[WilO]

Dave,
thanks for sharing that tidbit; gives me more to think about with my present road/track vehicle...

JT, thank you for the clarifications.

Tim,
Thank you for your clarifications, and setting me straight regarding the distribution of load transfer, and for pointing out what I should have remembered regarding IC height and jacking forces.

My own road/track vehicle was the inspiration for the thought experiments that lead me start this thread. I suspect there remains a fair amount of compliance that contributes to roll steer, and other nonlinearities that make the car handle like a snow blower....

[olefud]

Also, its possible to have a car that doesnt roll. If you set the IC of each suspension high enough you will have a jacking force that equals the load transfer and the end result is no movement of the suspension. This is commonly seen as "putting the roll centre at the CG". It is never done. Ive never heard a decent explanation why though.Tim

I’ve heard that drivers dislike the feel without roll, or when the front roll center is higher than the rear. More importantly, with out roll couples and the resulting roll, it wouldn’t be possible to tune for over/understeer by adjusting roll resistance at the axles. Weight transfer feed in over the period of the roll rather than as a step function without roll is probable a good thing.

None of these considerations are drop dead concerns that couldn’t be worked around. But it’s a lot easier with roll.

[Tim.Wright]

I’ve heard that drivers dislike the feel without roll, or when the front roll center is higher than the rear.

Be careful what you believe. When it comes to roll centres, there is no other field in vehicle dynamics so steeped in bullshit.

Everything else you said makes sense

Tim

[olefud]

Be careful what you believe. When it comes to roll centres, there is no other field in vehicle dynamics so steeped in bullshit.

I wouldn’t say bullshit so much as stating a very special case situation as a general rule. I’ve worked with teams and drivers that did (by my lights) strange things yet went fast. Compensating errors are wonderful things.

[Jersey Tom]

I'd say the vast majority of talk regarding racecar engineering and vehicle dynamics is heavily steeped in BS.

[RideRate]

Didn't mean to double post.

[RideRate]

Thank you for the reply Tim.
Unfortunately I worded my question poorly; I was thinking more along the lines of the relative phasing of these events, and how yaw rate and lateral velocity might be affected by roll movement, particularly if a vehicle is underdamped in roll. If we had an appropriately instrumented vehicle (LVDTs on the dampers, yaw rate sensor, etc.) how might the graphs look overlaid on each other.

First off, good question. I think it shows a lot of thought.

As stated this is a concern of transient behavior. I think the conceptual problem here arises from the aforementioned standard racecar bs and the overly obsessive concept of chassis 'roll'. As I approach the problem, roll itself is nothing more than a resulting event (effect), never a cause. Roll itself isn't going to drive your yaw or lateral response. Wheel loads do drive these things which are affected by dampers, springs, geometry, distributions, etc. Oddly enough beyond the car's fixed roll inertia these same components that drive wheel load responses also drive transient roll response. So the roll itself isn't driving anything, it's just a response, however the important components that drive the important things also drive roll. Hope that makes sense for your transients.

The other effects you mention (camber gain) are things I am aware of. One clarification I'd ask for: Assuming a vehicle of finite track with a CG above the ground plane, I'm assuming that the wheel loads would vary whether the sprung mass actually rolled or not. The only way I can imagine it NOT rolling is the presence of an active suspension, which would be required to provide a support force to prevent chassis roll. Wouldn't this support force would change the wheel loading?

Just to drive home my point. Yes the wheel loads would be different in transients only for the two cases, but it wouldn't be BECAUSE the vehicle didn't roll, it would be because of the components you changed to achieve zero roll. Those changed components caused the zero roll and different wheel responses, the no roll didn't directly cause any new response.

Remember both cases do however have the SAME steady state loading considering a finite track length and cg height. The only effect roll can have on the steady state loads is if the motion causes your track width to change or cg height to change. This goes for the steady state loading with active suspensions too if it doesn't cheat by moving the cg. For clarity the constant loading is left side vs right side for a steady state corner, the roll distribution can make the individual wheel loads different, but I'm assuming roll distributions also remain equal between the cases.

No roll from large jacking forces has been done but it wreaks havoc on system response and wheel compliance while introducing oscillations. Same as lots of anti-squat or anti-dive which tends to promote wheel hop and brake chatter. You can put all that jacking in and then fight the problems to make it better, but question is always, "why?" There's a reason it's not done successfully.

[WilO]

Thanks for the help RideRate, appreciate your input and it helped me to 'visualize' whats occurring during transient roll.

[olefud]

Transient roll response can be tuned and shouldn’t be overlooked. For instance, just as axle roll resistance can be tuned for steady state balance, dampers can be tuned for transient balance. Steady state dominates by duration but a good turn in sets up the steady state.