• This Website Is Not For Sale
  1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Active Suspension in AC?

Discussion in 'Assetto Corsa' started by Rhys Gardiner, Dec 31, 2013.

  1. I guess this post could apply to any of the next gen sims, but since AC is the only one I know of that simulates systems like movable aerodynamics (Supercars, 599XX) I felt it most appropriate to place this post here in the AC forum.

    I'm very interested in active suspension, particularly the kind used on the early 90s Williams F1 cars. No sim has managed to simulate those cars as they actually were, as sims like rFactor 1 have only allowed for "normal" suspension systems, as well as fixed aerodynamics. (Excluding plugins that added DRS and KERS, as these are driver-activated.)

    But Assetto Corsa has included one of these "advanced features" that, to my knowledge, hasn't been seen before in a PC sim: movable aero that adjusts itself as the car reaches certain speeds. And seeing these movable rear wings on the MP4-12C and 599XX in AC got me thinking: Would it be possible to go (quite a few steps) further, and simulate a Williams F1-style active suspension system in Assetto Corsa?

    Of course there is a big difference between Active Suspension and movable aero - the movable rear wings are a passive system, easy to code into a car's physics (I assume) - just set the system to activate above X speed - whereas active suspension is just that, active. Many dynamic parameters to take into account.

    We also have to consider the fact that suspension in its normal, passive form is complicated enough to simulate as it is - how much more complicated would it be to simulate suspension that dynamically changes depending on your driving inputs and where you are on the track? Even more, how much would be required to have the virtual active suspension utilise telemetry as you are driving to learn the circuit and its characteristics, as it did in real life?

    Does AC or sims like it even allow for this kind of thing?

    I want to see what other peoples' thoughts are on this...
  2. Well, the wings are not quite as simple as activating at speed - on the MP4, it's used for air-brakes, so they go up at a combination of high speed + high braking. I understand the 599xx uses them to help turn the car in faster. And so on.

    I think the hard part of simulating active suspension is gonna be picking the tuning parameters. There are less constraints on what's mechanically possible (for example, you can trivially build a damper with variable speed response, it's just another input) and even changing the length of suspension components is only a matter of editing some numbers in realtime. But in order to actively help the driver, you need a good idea of the capabilities and methods of the particular suspension system.

    Using telemetry to learn the circuit, I dunno... obviously modern CPUs are up to the challenge compared with what was available in 1990, but again, the actual parameters it learns would be tricky to find out.
  3. Ah, I forgot about the braking aids. That certainly makes the system a bit more complex.

    This is true. I was thinking that maybe it could be done by loading a track telemetry profile into some sort of plugin that manages the active suspension system as you drive. The plugin would analyse your lines and inputs and change the suspension settings to some "ideal" values as you enter those parts of the track. Not the same as in real life where the system actively learned the track, of course, but it's similar.
  4. Hmm, online info says the Williams used a hydropneumatic system with measurements of each wheel's deflection used in a feedback loop that controlled spring rate & damping to get more desirable ride characteristics.

    The hydro- would be the damping; fairly normal to use fluid displacement through variable sized inlets to control damping. Pneumatic is the spring component - an air chamber linked to the hydraulic system gives it springiness (actually with better characteristics than coil springs - no natural oscillation frequency - but also higher maintenance). I imagine what they'd be doing is getting the car to ride flat in more circumstances, and dynamically compensate for downforce. Changing the amount of hydraulic fluid in the main loop affects static height without compromising on springs. Changing inlet sizes affects the damping applied to it.

    AC has some sort of active suspension on the MP4-12C (I don't know what files the configuration lives in, but from ones I've seen there's a gap between the static suspension numbers and ingame performance)

    In numerical terms a spring is nice and simple to model - it applies force in one direction (along the axis of the spring) that depends only on the length of the spring. An active suspension, in basic terms, is just moving the end of the spring that's attached to the car - it does it through hydraulics/pneumatics for convenience but the effect is the same.

    If you want a car to ride level over time, each spring just needs to be applying, at '0' height, the average weight of the car on that corner. In a sim you can work that out directly - you know the sprung weight's distribution, you know the spring's response to force, so you just set the upper end of the spring to the exact length that provides that amount of force. In a real car (the Williams for example) you measure the deflection and if it's compressed from where you want it, you increase spring strength, if it's extended you decrease it. So really to model this the sim has less work to do - you just set how fast it can modify spring strengths & damping rates, work out whether it's too strong or too weak, and change it from there.

    I would think active damping tries to smooth out the ride as much as possible - underdamping leads to oscillation, overdamping limits the speed at which the car responds to bumps (rougher ride). So maybe that's where the 'learning the track' comes into play - it knows which sections are riding too rough, and the next time round the dampers are tuned to suit the road.
    • Like Like x 1
    • Beer Beer x 1
  5. Very interesting thoughts.

    We also have to consider the fact that the Williams system was able to stiffen the outside suspension through corners - I've also read somewhere that the system could tilt the car into a corner too. You can see that from watching onboard footage of the FW14B at Silverstone 1992- through right handers the suspension on the right side seems to compress, and through left handers decompress, which should be the opposite way round on a passive suspension system.

    There were also many other ways active suspension was played with in 92 and 93... one of them being a pseudo-push to pass system. From wikipedia:

    It certainly is possible to simulate active suspension. As you said, the McLaren has an active system... but it differs from the system that the Williams F1 cars used (Mclaren's is only semi-active, as seen here). Would AC be able to simulate the Williams System specifically? Or do you think a sim would have to be designed with the ability to set how fast parameters are modified from the ground up?

    I know that there was controversy in the iRacing DWC some time ago over drivers using button macro cheats, which basically allowed drivers to press a series of buttons over the course of a lap that changed the car's settings such that you had the optimal setup for individual parts of the track, even individual corners. It's the same kind of principle, but less dynamic. So it might be possible through an external program if the sim itself is not capable.

    Sorry if my remarks are all over the place, I'm in a bit over my head here...
  6. Hmm, I suppose at that point you could use camber change due to suspension offset to affect cornering - really you want both wheels to be cambered towards the inside of the corner afaik, so leaning into the corner gives you that without needing nonparallel A-arms or whatever. And you can be much closer to 0 camber on straights without losing the ability to camber into curves.

    I dunno if ground effect's been done properly in any sims, never really hear it talked about specifically (instead of the general wing modeling which is pretty basic). Interesting that it's a matter of such small distances that lifting the car would help.