Car physics - how are the values in the files related?

Discussion in 'Car Modding' started by Rocksor, Aug 28, 2013.

  1. Rocksor

    Rocksor Registered

    Joined:
    Oct 20, 2010
    Messages:
    99
    Likes Received:
    1
    Hi,

    when doing some experiment with the physics and modelling a suspension, some questions came up.
    I realised, that there are some values, that you can set to different values in different places. E.g. you have a caster value you can set in the .hdv, whereas the caster basically already is defined by A-Arm connection points at the upright, that is set in the .pm file...

    So the question is: What is calculated in the sim, when I have the caster that derives from the .pm set to about 13° and in the .hdv to 4°? Is it 13°(irrelevant), 17°(stacking) or 4°(overwriting)? As I feel differences, when altering the .hdv caster value I think its not irrellevant, but what do I have to set the value to, to really get the .pm caster value as base setting, but still be able to change it in the car setup?

    Based on this example, I'd like to know, how those values that you can find in more than one places are related with each other...

    Is there any documentation on this?

    Thanks
     
  2. lordpantsington

    lordpantsington Registered

    Joined:
    Oct 5, 2010
    Messages:
    849
    Likes Received:
    79
    Caster
    Is measured then adjusted for. Build caster into the pm. If the caster in the pm does not match the garage setting the upper and lower ball joints are shifted fore and aft an equal amount until the correct caster is reached.

    Camber
    Is assumed to be 0 in the pm. Do not build it into the pm. Camber is adjusted by rotating the wheel about the center, so if your upper and lower ball joints are not equally spaced from the wheel center, they will not shift an equal distance.

    Toe
    Is assumed to be 0 in the pm. Do not build it into the pm. Toe is adjusted by moving the spindle side straight link pick up point outwards from the car's centerline.
     
    Last edited by a moderator: Aug 29, 2013
  3. Rocksor

    Rocksor Registered

    Joined:
    Oct 20, 2010
    Messages:
    99
    Likes Received:
    1
    Thank you! Now I got my caster adjustment right! But it seems to be -13° to get the right feeling as +13° pulls the steering wheel to the ouside...

    But I have another question on the documentation. I'm not quite sure if this is the right place to ask for this, or if this would rather be a new topic: Is there any docometation of how all the values in all the files are related to each other, or is there at least a documentation wich instances can be generated, how the have to be named, which physical values/ names can be adressed, what dimension they have, and so on?

    I started building the physic I'm working on at the moment based on the rtrainer files. But how can I be sure, there are all possible values in there including the not needed remaining as comment?

    Maybe there are some more possible parameters to address, that I don't even know...
     
  4. Rocksor

    Rocksor Registered

    Joined:
    Oct 20, 2010
    Messages:
    99
    Likes Received:
    1
    The next question to relations is this:
    How are the different height parameters related? I made a picture to show I do understand this. Is this right?
    View attachment 9567

    They height of the y=0 point is defined by the wheel radius. From this point I can define the lowest point in my suspension geometry, which defines the relative plane do define the CG height. And all of these parameters are defined in different files.

    If this is not completely wrong then how does the ride height I can define in the .hdv for each wheel fit in this picture and what do those undertray points do at the y=0 plane?
    I am a bit confused:confused:
     
  5. lordpantsington

    lordpantsington Registered

    Joined:
    Oct 5, 2010
    Messages:
    849
    Likes Received:
    79
    If you haven't already, the pm should also be built as if the car was up on jacks (wheels not loaded/wheels in full droop).
    I set the reference plane to the bottom of the tub, so that design ride height=0. This way I drop a term when calculating the value of CorrectedInnerSuspensionHeight.
    I suppose that setting your reference plane to the lowest suspension pick up is also okay. (Or is it that you set the reference plane to the center of the axles?) As long as you remain consistent in your approach it won't be an issue. I've swapped from PM origin @ CG, Wheel Center, Reference Plane and haven't felt a difference as long as I set CISH correctly.

    In rf1 it helped me to think of things as this step by step process:
    1) Geometry of each corner moved from defined wheel center in PM to mesh center of individual corner if it didn't match.
    2) Outer pick up points raised so they are the correct radius from the ground.
    3) Inner points raised by CISH.
    4) Upon loading the track the car body is then set to garage values (inner points raised by ride height).
    5) Car is dropped onto wheels upon entering the track which determines droop.

    CG height is measured from the reference plane.
    The under tray points also are measured relative to the reference plane.

    I'm not sure what is going on with your caster.
     
    Last edited by a moderator: Sep 3, 2013
  6. Rocksor

    Rocksor Registered

    Joined:
    Oct 20, 2010
    Messages:
    99
    Likes Received:
    1
    Hi,
    thank you very much for your help. I've been thinking a bit about your explanation and checking some ISI Car physics files and some more thinking, and I think I understood it better now, but some new questions came up. But I need some time and a quiet mintue to think of it again and write it down. But I'm afraid I wont have that time in the next days.

    Concerning the caster: The physics is based on the rtainer model, that ISI released with the moddev. This also has a positive caster setting and also pulls the steering wheel to the outside whereas e.g. the skippy centers the wheel with its positive caster.
    I'm not sure whats going on here as well, but maybe someone can find this out by looking at the rtrainer files. Or at least confirm this behaviour of the rtrainer on track. So maybe something is messed up in my installation.

    Thank you very much.
     
  7. Rocksor

    Rocksor Registered

    Joined:
    Oct 20, 2010
    Messages:
    99
    Likes Received:
    1
    Allright then. I've had some time to think about what you explained to me and I painted another picture to show you how I did understand this:
    View attachment 9620

    So basically the reference plane is lowered to the lowest point of the chassis. based on this reference plane I define all the points of the suspension, in the state of onloaded wheels.

    To define the loaded state when loading the car you can set this to the design ride height via defining the CISH.

    Did I get this so far?

    So here comes the first thing I'm not sure about: In my opinion this should mean that the rebound travel that is left at this state (fully unloaded) is "0" which also would mean that the bump travel value is equal to the full amount of supsension travel.

    In other words, the CISH is equal to the rebound travel, or did I mix something up here?
    But when looking at the ISI files this doesn't match.

    In other words: The way I do understand this I would set the pickup points to the loaded value (which is the basic ride height the cars is constructed) and leave the CISH at 0. This would be easier for me, as I have the complete CAD model of the car here, and to get the unloaded values I need to do some calculations. Could this work also?

    The next thing I still don't completely understand is the undertray points. I thought this would define the height position of the underfloor, so the car scratches the ground, when set too low. But the skippi has its undertray height values set to "0" and the reference plane at the height of the wheel radius. So this doesn't make sense at all to me...

    I hope my thinking is not too far away from the rF2 reality...
     
  8. lordpantsington

    lordpantsington Registered

    Joined:
    Oct 5, 2010
    Messages:
    849
    Likes Received:
    79
    Sounds like you've got it.
    Generally you are correct about no rebound travel. Specifically the way rf handles this is: you are at the point at which the bump rubber is affecting the spring rate in a rebound direction. Remember there are no hard limits to the suspension travel in rf.
    Mixed up. When you first get into the car in the garage the car is 'dropped' onto the wheels, this will set rebound travel (by load vs spring rate). While the spindle side points are adjusted to radius, there is nothing other than ride height to adjust the body side points. CISH adjusts the body side points up so that they have a correct starting point for the kinematics.

    I understand trying to shortcut it. I ended up building a solid model and jacking the body since trying to calculate how 5 bars are moving in relation to each other in 3D made my head spin.

    The undertray is measured from the reference plane. Most of rf's stuff is built with the tub sitting on the ground (ride height=0). From this we can say the undertray is where the reference plane is (and not @ the axle height). The PM locations are basically relative to the wheel centers, and get moved around as necessary for the vehicle. This is to allow modders to use pm files from other mods since suspension design is a very heady subject.

    What you are doing to understand is almost exactly the route I took. Unfortunately even when given "Static Loaded Radius - Wheel Center Height(.PM) - Design Ride height" while examining the FR3.5 I could not make the numbers work with the data I had and what was in the mod. There is going to be a point at which we are both equally stumped.
     
  9. Rocksor

    Rocksor Registered

    Joined:
    Oct 20, 2010
    Messages:
    99
    Likes Received:
    1
    Praise the Lord!

    I think now I understood most of the points. Still I'm not quite sure about the CISH although I think I know what it does. It corrects the body suspension height to a value that sets the bottom tub to the ground in relation to the outer suspension points. From that point the design ride height raises the car to the level it should have.

    But does this correction affect the bump and rebound travel values? I guess not...

    And what effect does "dropping the car on the wheels" when entering the track have on the bump and rebound travel? I guess none...

    If my guesses are right, then I think the following steps should define the suspension the way it was designed in reality:

    1. I set the reference plane to bottom tub.
    2. From that point I define my suspension points and undertray points and suspension travel values according to the CAD model (fully loaded - weight vs. spring rate)
    3. Using the CISH to correct the suspension height until the reference plane and the road are at the same height.
    4. Raising the suspension by the design ride height (which in this case happens to be the same value as the CISH, as the suspension was built in design ride height)

    Could you please confirm or deny this theory?


    That's just the reason, why I would like to use the points I can take straight out of CAD. I think it can't get more accurate than this...

    Good to hear I'm not alone;). But maybe someone of the ISI physics designers can bring some light into the dark. Especially the untertray point at the height of the wheel center is a thing I can't understand at all...:confused:
     
  10. lordpantsington

    lordpantsington Registered

    Joined:
    Oct 5, 2010
    Messages:
    849
    Likes Received:
    79
    I've approached a few ISI guys for answers, and am always impressed that they not only take the time to respond, but also how they respond. I can only imagine how busy they are with development so I hope people don't just go straight to pm and slam their in-boxes. Most of what I've replied to you is what was given to me, I can't take any credit. It is only on the back of the community that I've gotten this far.

    I suppose the easy option is to inflate CISH to move the body side pick up points an additional amount to compensate for design ride height. It might not be exact because kinematic motion of wheels are curves, but it is probably going to be close enough for most.

    The reference plane is set to ride height, and thus the undertray is too if the vertical dimension in the hdv is 0.

    The locations in the PM don't matter as long as they are set correctly relative to the wheel center (defined in the pm). The locations of the pick up points in the PM move to suit the car. If you were to offset all of your points in the PM 1 meter up, the car will behave the same. Move all points 1m forward, the car behaves the same. Each individual corner is relocated to its correct place. Where is the "correct place"? I believe the pivot of each individual wheel mesh.
     
  11. Rocksor

    Rocksor Registered

    Joined:
    Oct 20, 2010
    Messages:
    99
    Likes Received:
    1
    Does this mean CISH does only affect the y-position of the suspension pickup points of the body without using the given kinematics?
    In that case using the CISH will always produce some inaccurracies.
     
  12. lordpantsington

    lordpantsington Registered

    Joined:
    Oct 5, 2010
    Messages:
    849
    Likes Received:
    79
    It is an educated guess, but I'd say Yes, CISH is only moving the body in the vertical. I'd be surprised with all of the adjustments made to the suspension if the length of the bars are fixed as defined in the pm. I mean if your caster is not correct, the upper and lower points are moved fore and aft an equal distance from the center of the axle regardless of their vertical distance from center, this is not possible while maintaining the same length bars. Likewise, camber adjusts the upper and lower and equal amount. Whereas in real life, for the car I'm modding, shims are only added to the lower control arm.

    My understanding is that the entire suspension geometry is moved to the correct corner, and zeroed (moved the center of the tire to its respective place on the ground plane, a vertical movement). Where then the radius moves the outer points vertically, the inner (body side) points need some distance to move (CISH). Basically CISH offsets the funky method used to locate the suspension to the corner. If done as ISI does it, this locates things back to where they belong. I'm not sure what happens if you move additional. I suspect that until the car is set on the wheels all movements are translation. rF Pro was marketed with a better (full kinematic?) suspension system. Which leads me to believe what is in rf and 2 is simplified.

    To further illustrate CISH:
    If in your CAD model @ 5cm Ride height: the distance the lowest part of the frame is to the lowest body suspension pickup is 5 cm, the distance from the axle center in the pm to the pickup is 15 cm, your tbc wheel radius is 25cm, and your pm coordinates are defined with the center of axle @ 0. Your CISH ends up @ 20 cm. The only logical reason I can think of why we would be moving the body points up 20 cm, is to have the tub sitting on the ground (so that when moved up the ride height it actually is in the correct place).

    If your PM defines the center of axle @ 25, your CISH is -5. Once again moves the tub to the ground.

    While I took the time to model the system in solids to look closer at the kinematics, I'm not so sure the time spent amounted to much. Certainly it helped me locate where the pick up points likely are. But if I had CAD documents already, I'm not sure spending time was worth it. Knowing just how much the geometries are manipulated internally I'm not sure a vertically translated distance considered into CISH is going to have enough difference over a full kinematic model to warrant the time spent. But there is a piece of me that also says the closer I can feed rf data versus the real deal, the more the character of the car is retained. I suppose this is the art behind modding physics.
     
  13. Rocksor

    Rocksor Registered

    Joined:
    Oct 20, 2010
    Messages:
    99
    Likes Received:
    1
    You've got a point here. So another educated guess from me is, that ride height setup also doesn't follow kinematic rules. In this case I should get the best possible kinematics starting point by lowering the ride height via CISH an raising iot again via design ride height.

    But maybe all of this doesn't matter anymore with the new constraint system in the new build. So for now I'm off my computer for the next 2 weeks, as I'll go on holiday and when I'm back I'll have a look at what the new system does...

    Thank you very much for sharing your knowledge, no matter how you acciured it. You helped me a lot in understanding the suspension design in rF (even if there still are some questionmarks)

    I hope some day in the future I can release the product of all this. I think this is going to be a really fun car to drive!
     
  14. smbrm

    smbrm Registered

    Joined:
    Nov 11, 2010
    Messages:
    440
    Likes Received:
    50
    Lordpantsington & Rocksor

    Thanks for posting this discussion. I appreciate that this is an old thread.
    I think I follow most of the logic, but would be helped by some additional clarification.

    For a car I am reviewing:

    1) I can find the CISH or its equivalent in the HDV. In an example I am looking at, the CISH is -1 which is defined as the "original behaviour".
    2) I can find the tire radius in the tbc Example 0.310 front, 0.320 rear
    3) I cannot find a tire centre distance in the PM other than it is located at zero in the particular car pm I am looking at.
    4) I am also trying to locate/determine the design ride height.

    do some of these parameters have different names comparing rfactor 2 to rfactor 1?


    I am trying to determine the CofG height relative to the ground for this example, so need the height of the reference plane above the ground.

    Appreciate any further clarification that can be provided in how to figure out the position of the reference plane for this example?

    thanks
     
  15. lordpantsington

    lordpantsington Registered

    Joined:
    Oct 5, 2010
    Messages:
    849
    Likes Received:
    79
    1) As I understand it, -1 is equivalent to holding the inner points in relation to the outer. If your pm is defined with the kinematics @ optimum, the inner points will move in relation to the body and stay at optimum regardless of ride height.
    2/3) The pm locations are defined with respect to wheel center, so that you can interchange suspension geometries easily. The outer positions get moved with respect to the tbc radius. If they are defined in the pm with wheel center @ 0,0,0 that is fine, as long as all other suspension pick-up points are also correctly located as if the wheel center is at 0,0,0.
    4) The design ride height is what you make it. If you measured your inner pickup points' vertical distance from the ground @ a ride height of X, then that X is the DRH. If the tub was on the ground then DRH is 0.

    Yes some of the parameters have changed names, but you should identify them easily enough. Check the patch notes. Also note that the tire radius in rf2 is a bit more complex than just being read directly from the TBC.

    CofG height as defined in the HDV is measured from the reference plane IIRC.

    The reference plane is where you put it. Think of it as a level laser plane projected under the car that if you were measuring pick up point locations, plumb bobs would be hung down to for the vertical distances/positions. Just make sure it doesn't move so all positions are correct in relation to each other.
     
  16. smbrm

    smbrm Registered

    Joined:
    Nov 11, 2010
    Messages:
    440
    Likes Received:
    50
    Thanks for the reply Lordpantsington.

    You clarified a couple of things. Can you comment further on:

    1) I understand the design ride height is what you make it. Is there a place where the design ride height for an existing car is specified/recorded in the files?

    2) where would I find the patch notes to which you refer? Are these the Build notes, or something else?

    3) I understand about measuring the reference plane, but am wondering where that info is located in the files or derived from the files?

    Cheers
     
  17. lordpantsington

    lordpantsington Registered

    Joined:
    Oct 5, 2010
    Messages:
    849
    Likes Received:
    79
    1) No
    2) Yes, with each build there is a readme usually posted in the forum. Pay attention to the modding section.
    3) It isn't but with some logic you can probablylocate it. Commonly it is placed on the bottom of the chassis.

    Sounds like what you are looking at exists already, what are you referencing?
     
  18. Marek Lesniak

    Marek Lesniak Car Team Staff Member

    Joined:
    Oct 6, 2010
    Messages:
    1,585
    Likes Received:
    101
    Just to confirm, everything what lordpantsington said is correct.
    Reference plane is what you can call a flat underbody of a race car. It's also used for aero calculations and ride heights.

    As for suspension geometry - whatever you set in PM file, rFactor takes the "outer" stuff of the susp up by static tyre radius taken from TBC.
    If CISH is "-1" it will also be taken up by the same value. Whatever ride height change you make in garage, it won't affect susp geometry.
    If CISH value is "0" or higher, then that value will be used to move up "inner" stuff of the susp. Then, another correction is made using ride height values from HDV.
    In other words, If CISH is "0" or higher, once rF rises inner susp points using that CISH value, those inner points are "welded" with vehicle's body (which is what we want) and from now on, every ride height change will affect both body position and suspension geometry.
     
  19. fireeyes

    fireeyes Registered

    Joined:
    Oct 5, 2010
    Messages:
    25
    Likes Received:
    2
    Thanks for a great thread. Very useful. I have two questions... and some assumptions.

    Assumptions:
    The PM file needs the defined pickup points for the wishbones (bars) with the car in full droop (needs starting point to apply springs rates etc.).
    The upright/wheel assembly is defined in the PM with no camber, caster or toe (same deal).
    Internally, after the HDV file has applied camber, caster, toe to the wheel assembly, rFactor has to mate all these points up and load the car (HDV may change track and wheelbase as well if defined).
    Bristow, this thread, and others have made it very clear that the wishbones will be tweaked in both length and angle to resolve points that don't align at the end of this process.

    First Question:
    Assuming all that is accurate... Do we know if the suspension is "dynamic" after the car is loaded and the pickups connected? Meaning the "wishbone map" or arc will accurately cause slight changes to the track, camber, caster etc. as the suspension moves? Or, is the wheel pivot point always constrained to vertical (y axis) movement in the physics model?

    Second question:
    Despite reading everything I can find on the subject, I have managed the miss where I'm supposed to place the pivot point for the graphical wishbones (chassis pivot, wheel pivot or other). I also assume they are positioned graphically where they will be with the car loaded using default HDV settings (as opposed to full-droop)?

    Thanks in advance.
     
  20. lordpantsington

    lordpantsington Registered

    Joined:
    Oct 5, 2010
    Messages:
    849
    Likes Received:
    79
    Assumptions seem correct.

    Answer 1: Yes, the suspension is dynamic after load in and will follow the kinematic curves.

    A2: I've not found any tutelage on the graphical suspension. As far as the approach I'd take, I'd either put the local pivot on the body side uni-ball joint, the upright side ball joint, or the Center of Gravity of the link. I'd guess they are positioned @ full droop, but I've not gotten to this point yet so it is a pure guess.
     

Share This Page