mep wrote:Tim.Wright wrote:Until then, try to avoid reading any automotive specific treatments on instant centres and roll centres until you understand the fundamentals. Most of the classic theories are gross simplifications and are misleading or flat out wrong.
Do you have any examples of this?
Yes, there are shirtloads.
[rant=1]
I'm not going to name names because its a small world I work in. I will name the concepts though. The problems are all related to incorrect use of the "geometric" roll centre. This is the one found by projecting so many lines from various points in the suspension that its impossible the the resulting intersection point cannot posses mystical properties.
Many books, websites and seminars will lead you to believe that the geometric roll centre is actually the fundamental mechanism governing suspension roll which is not correct. The geometric roll centre is merely a very basic model, used to simplify all the lateral dynamics of a cornering car down to a single degree of freedom model which is valid under some very simplified conditions. This method is not useless - in fact I use it occasionally to calculate different elastic setups because I've found that its not bad at predicting trends of balance and roll between different elastic setups. But if I want to accurately calculate the tyre forces or roll angle - especially in limit conditions, I use other methods (often a full vehicle simulation).
The main problem of the geometric roll centre (due to its name) is that people think that the chassis rolls about the roll centre and that it acts like a pin joint in determining the tyre forces.
Bill Mitchell showed
here that body movements and tyre forces need to be approached considering the left and right wheels seperately AND knowing the contact patch lateral forces if you want accurate results. Erik Zapletal confirms
here (in his usual condescending style) that the roll movement depends the left/right lateral tyres forces and the force-line or n-line slopes (even with the roll centre in outer space) and also addresses the vertical jacking effect. Danny Nowlan is also using a similar method to Bill Mitchell in his articles and software. S.M. Sincere (SAE Paper 983033) has shown that non-linearities in wheel rates are also a factor in determining the roll angle and move the actual roll centre away from the geometric one. Basically, the industry is slowly moving away from classical roll centre theory, but unfortunately large segments of the education system are not.
The main issue with classic roll centre theory is that it ignores the difference between the left and right lateral tyre forces which can be responsible for a significant portion of the roll angle and vertical load transfer forces. The above people demonstrate this
The fact that there is such a knowledge gap between classical roll centre theory and the actual behaviour of a car has led to various myths and wives-tales being propogated through (often expensive) books, journals, magazines, seminars. I have read, or been told the following during my lifetime:
- Roll centres should never pass through the ground as it causes instabilities and gives give bad driver feedback
- Lateral roll centre migration must be minimised because it makes a vehicle "inconsistent" and difficult to tune.
- Front and rear roll centres must move in the same direction in cornering otherwise handling will be inconsistent in corners
- Rear roll centres should be higher than the front because the "CG axis" of the car is inclined up towards the rear
- Vertical weight forces cause rolling moments about the geometric roll axis which lead to changes in the roll angle and extra lateral load transfer.
- Letting the roll centre pass through the contact patch of a tyre will effectively lock that side of the suspension solid. It also causes cancer.
- Roll centre height variation during roll must be minimised otherwise the LLTD balance will change with roll (fun fact: it will change even with a rock solid RCH)
All of them are complete BS and you can see many of them hide behind the conveniently unquantifiable smoke screens of "bad driver feedback" or "inconsistent handling" which is effectively saying "god works in mysterious ways".
When I was at the start of my career in vehicle dynamics I wasted massive amounts of time trying to understand the phenomena from my list. In the end, as my general engineering skills improved (a process helped by ignoring most automotive texts and drawing loads of free body diagrams myself) I realised/proved to myself not only that every point on the above list is BS, but I have in my time analysed race and road cars which break rules 1, 2, 3, 4 and 6 and show no ill effects. Some of these cars were even analysed at an extremeley deep level using wheel force sensors, slip sensors and comparing them with various calculations. Points 5 and 7 can be proved to be BS with a simple free body diagram.
In addition to that, In all of my (short) career, I've never ever heard of any stories, even from the people who propogate this stuff, about any peculiar handling problems that were ever objectively isolated to one of these roll centre effects (from my list) and then were subsequently cured by changing the geometry to "fix" the roll centre behaviour. Never.
This is why I suggested avoiding automotive texts if you want to learn about suspension. There are some good ones (usually in german) but most of the others will "teach" you rubbish that will take years to unlearn. This was my experience at least, and as you can possibly tell I'm still a little annoyed about it. It took me about 3-4 years to unlearn all the crap I'd "learnt" and to finally understand how a vehicle rolls and pitches. It turns out that it just follows the normal laws of physics.
[rant=0]
