Using Rake to Tune Car's Handling

[Tennisbiki]

I suspect not many people out there is interested in tuning suspensions for the downloaded mods to behave more "believable". So this is more of a journal or reflection than a how to tutorial.

I do not have much credential for modding cars in Assetto Corsa. The extend of my messing around is limited to suspension.ini, car.ini, setup.ini for about a handful of cars posted in RaceDepartment. Though I teach science in a K12 school and have an BS in engineering.

The recent obsession is on a Mercedes Benz 190E 2.3-16 model posted in RD. Though there is a 190E EVO 2 model from Kuno, there has not been any stock spec. W201 models out there until this one in October, 2023. Digging a bit deeper, I found the same suspension and aero as the Kuno’s 190E EVO 2 car. So essentially someone, a group of people, created or rip a 3-D model, did a bit of tuning on engine, drivetrain, suspension rates and got released, by multiple people. The mod itself, came with little suspension tuning option, like a proper stock car. Although the wheelrates are reasonable at around 1.4/1.5 Hz, the handling is ..edgy. The first obvious problem was the suspension travel of course, sitting in bumpstop or almost. Even after taking cared of that, the car looses grip near the limit. Using camber cannot dial that out. I then put in the proper front Strut and edited the rear to be close to the multi-link suspension. With much tuning it got better. I can see it in the suspension app, and how the car break away. At my better tune, the rear break away almost at the same time as the front and the car makes a slow spin toward the inside of the turn. So I struggled to give the rear more grip. The car was supposed to understeer, according to multiple reports, and can tighten with throttle, my goal.

I tweaked the aero.ini. It is an estimate, but I give it more lift in front so that the front grip is reduced at speed. It was bettered. And weeks after weeks, I was still stifled by the handling. The car would be fishtailing going through chicanes. Finally I was so desperate and about to slap on a rear DBW from a E36 M3.. but I did something else first, small thing, I changed ride height balance. Going out of “reasonable measure” I raised the front rideheight by 30mm and lowered the rear by about 20 mm. Just like that the rear was tamed.

Why? Allow me to think out loud for a bit. I think this CG height/roll center front/rear balance is probably something more fundamental than the camber/toe/wheelrate/ARB. All those forces work through CG, ALL forces. The most basics are mass, CG front/rear location, inertia and then CG height. Everything is locked in the chassis/body, but CG height can be adjusted by rideheight or suspension geometry. No, they are not the same thing CG height and roll center but in my understanding they are connected. So basically the soul of the car lies in these parameters, mass, CG location, inertia and f/r roll center. And then you can make the car stiff or soft, quick to respond or more forgiving with the suspension and tires.

Anyway, I know tuning the rideheight/roll center (rake) is probably tampering the data or reality. But creating a believable driving model is ultimately an art, so I am taking the liberty to be creative to achieve the my version of “reality”.

Thoughts?

 

[Stone16v]

I could help with measure some things from the real car, if it would help. my problem the most time when i try to tune/fix a car is my bad englisch

 

[mclarenf1papa]

I suspect not many people out there is interested in tuning suspensions for the downloaded mods to behave more "believable". So this is more of a journal or reflection than a how to tutorial.

I do not have much credential for modding cars in Assetto Corsa. The extend of my messing around is limited to suspension.ini, car.ini, setup.ini for about a handful of cars posted in RaceDepartment. Though I teach science in a K12 school and have an BS in engineering.

The recent obsession is on a Mercedes Benz 190E 2.3-16 model posted in RD. Though there is a 190E EVO 2 model from Kuno, there has not been any stock spec. W201 models out there until this one in October, 2023. Digging a bit deeper, I found the same suspension and aero as the Kuno’s 190E EVO 2 car. So essentially someone, a group of people, created or rip a 3-D model, did a bit of tuning on engine, drivetrain, suspension rates and got released, by multiple people. The mod itself, came with little suspension tuning option, like a proper stock car. Although the wheelrates are reasonable at around 1.4/1.5 Hz, the handling is ..edgy. The first obvious problem was the suspension travel of course, sitting in bumpstop or almost. Even after taking cared of that, the car looses grip near the limit. Using camber cannot dial that out. I then put in the proper front Strut and edited the rear to be close to the multi-link suspension. With much tuning it got better. I can see it in the suspension app, and how the car break away. At my better tune, the rear break away almost at the same time as the front and the car makes a slow spin toward the inside of the turn. So I struggled to give the rear more grip. The car was supposed to understeer, according to multiple reports, and can tighten with throttle, my goal.

I tweaked the aero.ini. It is an estimate, but I give it more lift in front so that the front grip is reduced at speed. It was bettered. And weeks after weeks, I was still stifled by the handling. The car would be fishtailing going through chicanes. Finally I was so desperate and about to slap on a rear DBW from a E36 M3.. but I did something else first, small thing, I changed ride height balance. Going out of “reasonable measure” I raised the front rideheight by 30mm and lowered the rear by about 20 mm. Just like that the rear was tamed.

Why? Allow me to think out loud for a bit. I think this CG height/roll center front/rear balance is probably something more fundamental than the camber/toe/wheelrate/ARB. All those forces work through CG, ALL forces. The most basics are mass, CG front/rear location, inertia and then CG height. Everything is locked in the chassis/body, but CG height can be adjusted by rideheight or suspension geometry. No, they are not the same thing CG height and roll center but in my understanding they are connected. So basically the soul of the car lies in these parameters, mass, CG location, inertia and f/r roll center. And then you can make the car stiff or soft, quick to respond or more forgiving with the suspension and tires.

Anyway, I know tuning the rideheight/roll center (rake) is probably tampering the data or reality. But creating a believable driving model is ultimately an art, so I am taking the liberty to be creative to achieve the my version of “reality”.

Thoughts?

Thoughts are that what decides mechanical balance is (mostly) something called LLTD (lateral load transfer distribution). Fxn of cg height, roll center height, and roll rate (per degree, so includes track width as well) for both axles.

You said the suspension frequencies were 1.4 and 1.5Hz respectively. All else equal, this will cause an oversteer tendency. Need either stiffer front roll rate via ARBs, track width differences, or a lower rear roll center (assuming same tires f/r) to make the car even be neutral mid corner (and for RWD you want an understeer steady state tendency, so even further in that direction).

In performance engineering, you generally look at the car's characteristics in a range of different situations to determine which part of the car/setup is the issue. Corner entry/mid/exit at low/med/high speeds are where you start. Understeer or oversteer in those specific phases informs the change that needs to be made. Takes a pretty specific and in-depth understanding of vehicle dynamics and the car itself to make good decisions based on that input (why some teams have fast cars in real life and others don't...). e.g. if you have oversteer on entry but the car handles well mid corner, you wouldn't want to do something that significantly changes the mid corner balance. Sounds like the problem you have with the car is multifaceted. Chicane issues would indicate problems with transient balance (ie dampers and bumpstops), general oversteer tendency in a car without significant aero would point to LLTD, etc.

There are many books on these topics (vehicle dynamics are quite well documented at this point), and I would say you're generally trying to skip a few steps with your approach.