Value deviations on ADAMS Car

[Pankaj_Kumar]

Hi,
Im Vignesh from suspension wing of Pravega Racing from VIT Vellore,India
We were iterating our front suspension using the software ADAMS Car.The graphs of Camber vs Roll show deviations from our manual calculations. The deviation was found to be 0.27degree
camber for 2 degrees of roll. I am unable to figure out the cause of this deviation and was wondering if anyone could shed light on the problem?
Also when we increased the IC length, the camber change under roll was decreasing when an increase was expected.

[Tim.Wright]

The problem is either:
Your calculation is wrong
Your Adams model is wrong
They are both wrong.

But unless you explain the calculation and exactly what simulation you ran how can anyone help?

[BillCobb]

Explain what you mean by "manual calculations". Do you mean kinematic analysis of a 4 bar linkage using trigonometry, or the measurement of a mock-up apparatus or of an actual car ? Did you actually meaure the camber by ride of your design or the camber by roll situation ?

[Claude Rouelle]

I am not an Adams fan. Doesn't mean I think it is a bad software, at the contrary. Why am I not a fan? Way to complicated, especially for a FS / FSAE project. I have met in my life many, many engineers who used Adams, nearly on daily basis, probably over 100 of them
I only know 2 of them who use it efficiently. Both of them (one in race car design and simulation - one of the winning Le Mans team -, the other in passenger car / truck design and simulation). Both are real Gods in both C++ and in Vehicle Dynamics.
Both told me that if you start from scratch you will need between 3 to 12 months to acquire all the data that are used as inputs for the simulation.
There are obviously more than these 2 guys in the world who can use Adams properly (Maybe Tim Wright is one of them?) but I do not know them
However in over 15 years of multi FS / FSAE events design judging every year, as well as seminars and university courses, I never met any FS / FSAE team that showed me they made a useful job with Adams. I do not say in doesn't exist, I say I never met such a team.

Some will say that because my company sell software I am biased. I am. But OptimumG software do not belong to the same league; not as detailed but much simpler to use.
As a matter of fact a German car manufacturer bought last year 8 licenses of OptimumKinematics: they told me that they just wanted (for example) to know how much bump steer they have of a simplified kinematics (no compliance) when the car goes down 50 mm. They refine their suspension kinematics and THEN they use Adams with all bushing etc... They told they narrowed more quickly the solutions this way. Lots of time solved.

As far as Pankaj-Kumar I feel that is, again, the kind of a question of somebody crying a bit too quickly for help without having done enough of his own homework. And with enough communication.

That being said Pakaj, are your tire deflections and your suspensions / chassis compliance included in your calculations? If Adams does and you don't that would be a first reason of discrepancy.

Unless you show how you calculate your own camber variation in roll (and, ideally, in heave and in steering too) nobody will be able to help you.
It is as if you were asking what difference there is between an orange and "that fruit" without explaining what "that fruit" is....

[BillCobb]

Post a drawing of your suspension in a few views, list a table of the X,Y, and z coordinates of all the attachment points and the knuckle/(upright) and tierods, etc and we'll have a shootout, so to speak. Winner take all. Best in class will show the evolution of these simple parameters as model complexity increases.

You guys wanted a build forum, lay it on us...

[Tim.Wright]

I'd tend to agree with Claude that Adams is bordering on overkill/too complicated for FSAE stuff unless you have done a training course. It's strength is it's flexibility but from a user point of view it means dozens and dozens of potential gotchas you need to keep under control. Then there are the usual implementation/logical bugs that you find in any good vehicle dynamics software.

I was dropped into Adams/Car without training years ago and it took me about 2-3 years of using it to become a "useful" user. After this phase I can now swear fluently in 3 languages. Perhaps with training you might be able to reduce that to 1 year (or around 20% of your time at uni doing nothing but Adams work). Even with training and experience it's easy to make mistakes. I've seen a few huge fuckups in OEM constructed handling models.

While it's possible to open a demo template, update the hardpoints to your design and run a simulation - if you want the results to be reliable there are dozens of other things you need to check first before running the simulation.

Due to all this, I always use simplified excel models (a kinematic model, a load transfer model and a 7 poster) to calculate several high level parameters of any model which I have in Adams/Car. If the Adams model matches my excel calcs then I trust the model. If not it's time to open the template editor and start debugging it.