Tyre friction/interaction with road surface

Discussion in 'General Discussion' started by green serpent, Nov 9, 2016.

  1. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,345
    Likes Received:
    6,572
    They could probably improve the output options available, namely steering torque sensitivity. Though I think it doesn't get utilised properly as it is, and should be more accessible.
     
  2. Euskotracks

    Euskotracks Registered

    Joined:
    Nov 29, 2013
    Messages:
    1,092
    Likes Received:
    191
    I would like to comment several facts that apparently seem not to be correctly modeled, which would affect FFB realism. Please correct me in case I make some kind of mistake or incorrect conclusion. In any case it would be nice to get some comments about it.
    • The convergence settings don't seem to be applying. If you check in Motec, you can see that the wheels in the same axis show the same "lateral patch velocity". Front vs rear values do show differences as one would expect. Considering convergence makes wheel point at different directions I would expect different slip angle at each wheel. I have checked as well graphically, and Formula ISI with 0,75 maximum convergence settings in both axes show parallel wheel when looked from the top of the car. The picture is included but whoever can try himself.

    • The steering trapece doesn't seem to be implemented. Seemingly both wheels rotate the same angle in rF2. At least this is the result I get when I check Formula ISI from the top at maximum wheel rotaton. The steering linkage should provide a higher rotation to the inner wheel so that all four axis converge in one single rotation centre. In reality there is no trapece geometry that provides perfect alignment, but for sure rF2 apparent rectangular geometry should be terrible for tire wear provoking excessive FFB. At the given angle there should be a 1.8 degree rotation difference, which is clearly a lot to be neglected.
    The following image shows the mentioned telemetry data and some conceptual drawings to better understand sliding angle and steering trapece concepts.

    SAT_SlipAngle.jpg
     
    Last edited: Nov 12, 2016
  3. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,345
    Likes Received:
    6,572
    Regarding effective toe, I've seen differences in FISI-derived mods in the logged toe for the front wheels. Maybe it's just not reflected graphically (though I'd rather check with an increased amount of steering lock, just to reduce error).

    I've also seen different lat patch vel values, but I'd have to double check the FISI (and with the current build) to be sure that's correct right now.
     
  4. Euskotracks

    Euskotracks Registered

    Joined:
    Nov 29, 2013
    Messages:
    1,092
    Likes Received:
    191
    I don't care about a logged toe which probably is using the same value as defined in the setup.
    I would like to see something that proves it is working.
    I have just double checked and I confirm the following:

    I used a base setup with minimum steering lock, minimum convergence, minimum camber and minimum caster.
    I took the following screenshot as an initial reference:
    base.JPG
    - Changing steering Lock is visible after maximum value is applied
    base10.JPG base14.JPG

    - Camber settings are visibly applied. Maximum values applied
    camb.JPG

    - convergence settings are not visibly applied when maximum values are applied.
    conv.JPG
    - caster settings are not visibly applied after applying maximum values. I made a different base picture to better check.
    caster5.JPG caster9.JPG

    I wonder why would they represent camber and not caster or convergence. It makes no sense to me unless that they are not being effectivly simulated. However, as I said, if someone thinks I am wrong feel free to correct me. If you do, please add some evidence supporting your statements.

    Note: The reason I took Formula ISI is simply because all the wheels are visible and becasue I wanted to pick an ISI car in order to avoid comments of people suggesting to use an ISI car if I had used some 3rd party mode. The telemetry picure from the previous post is however from ·Enduracers Flat6 v2.00
     
    Emery and peterchen like this.
  5. Euskotracks

    Euskotracks Registered

    Joined:
    Nov 29, 2013
    Messages:
    1,092
    Likes Received:
    191
    After a quick test I can confirm that convergence settings are effectively affecting lateral tire forces although not represented graphically.

    I need to later check for other parameters like lateral patch velocities which should vary significantly as well considering that this initial test has been performed at constant 106 kph speed in straight line.

    I need to figure out how to check for wheel individual rotations and see if the steering linkage is also being simulated.

    I do not unserstand however how the contact patch is determined if the wheel is not positioned in the correct position in the graphical engine where the contact with the track is calculated. Is there an additional invisible tire mesh defined in the tgm running in the background?
     
    peterchen likes this.
  6. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,345
    Likes Received:
    6,572
    Except that it's dynamic and different between the tyres, which are set to the same value in the setup.

    The basis for the physics (tyre contact with the track) isn't pulled from the graphics - it's the other way around. The graphical tyre might have 25 quads across the tread (rim to rim) and 40 quads around it, compared to 50-60 nodes rim to rim in the .tgm and 120-200 sections around. The graphical tyre is shaped to roughly approximate what the physics is producing. If the graphics omits some details it doesn't reflect on the physics.
     
    Flaux, TJones, Emery and 1 other person like this.
  7. Euskotracks

    Euskotracks Registered

    Joined:
    Nov 29, 2013
    Messages:
    1,092
    Likes Received:
    191
    After some tests I can say that both camber and convergence settings seem to behave as expected for the straight line test. I have tested the 4 combinations of convergence and camber settings with minimum and maximum values. The results are shown in the following image. Formula ISI has been used.

    Increasing toe in and caster have the expected effect in lateral forces. Due to the different sign for front and rear convergence, its combined effect with increasing negative camber cancels in the front and adds in th rear.

    Slip angles correctly match half the toe in applied. Each tire has different sign as expected. Increasing negative camber reduces front slip angle and increases rear slip angle. The reason is again the different sign for front and rear toe in.

    Lateral patch velocities are significant for maximum convergence cases and negligibles for the lowest ones. Each tire has different sign as expected. The effect of camber is similar to slip angle since they are linearly related.

    I will repeat this straight line test for Flat6 and introduce caster setting as well which will double the number of tests.

    Later I will test the in-turn situation in a controlled way. I will need to create a flat surface track and modify a car to limit steering to a few degrees. I will introduce then steering torque output and see how is affected by the different variables including the so mentioned caster setting that is so much mentioned when talking about FFB. Toe_In_Out.jpg
     
  8. Emery

    Emery Registered

    Joined:
    Oct 24, 2010
    Messages:
    3,035
    Likes Received:
    1,654
    Euskotracks, you are a testing monster these days and it is appreciated!
     
    peterchen, Flaux and Euskotracks like this.
  9. Jokeri

    Jokeri Registered

    Joined:
    May 17, 2014
    Messages:
    204
    Likes Received:
    32
    Euskotracks, test ISI nascar. You can adjust ackerman in that car. See how it affect things..
     
    Euskotracks likes this.
  10. lokidan

    lokidan Registered

    Joined:
    Nov 10, 2016
    Messages:
    35
    Likes Received:
    24
    this is all informative and all. But I'm a bit lost here, what exactly are we testing for...rf2 lack of forces on the tyre as they turn at speed? And why we have to increase caster to get the desired ffb.
     
  11. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,345
    Likes Received:
    6,572
    It's probably not really true to say we need to do that as a rule - someone with a very strong wheel wouldn't be doing that, because they don't want to be fighting the wheel that hard in every corner. And often people do that to 'boost up the forces' and completely lose the feel of the front end (I tried a decent direct drive wheel at a friend's place and couldn't work out why I was getting less feedback from the front tyres - mainly the force dropoff when entering understeer - than with the my G27 I had at the time, and finally, after leaving, realised the setup had too much caster)

    So, while we could argue about how rF2 feels and whether the caster for example is too large a component, or whether we think this or that is having the proper effect, plus throw in some wheel-specific issues that are only sometimes acknowledged, it seems to me Euskotracks is taking the opportunity to objectively test whether various setup parameters are having a realistic effect on the FFB, and would therefore be increasing the information needed for people to make a reasonable assessment of the rF2 FFB.

    Bearing in mind, as always, after rF1 there was a push from a portion of the community to have 'raw' FFB, and yet every second person you hear talking about racing in real life will say the wheel feedback, and for that matter the feedback from what they can see, are well behind seat-of-the-pants feel for working out what the car is doing and reacting to it. Even if we had perfect FFB (in combination with a wheel that could faithfully reproduce it), if most of us don't have some sort of motion rigs to convey the movement of the car itself we'll never really get true feel or have the same challenge driving fast in the game as we would in real life (physical requirements also aside).
     
  12. Euskotracks

    Euskotracks Registered

    Joined:
    Nov 29, 2013
    Messages:
    1,092
    Likes Received:
    191
    In order to avoid replying individually I will try to make a general statement.

    The idea of my tests is to analyse the influence of main tire positioning parameters which should have a significant effect on tire contact patch forces and on contact patch itself. This forces are responsible for the torque in the steering column: FFB.

    For sure the caster problem will specifically be addressed. So far I have started with straight line tests. Due to its simplicity, it is possible to extract many conclusions regarding tire forces before jumping to in turn tests where FFB will be possible to be measured as well.

    I have already done some initial in-turn trials in order to define how the definitive tests will be perfomed. For those tests, I have modified one of my tracks in order to have a huge flat surface where I can maintain turns at constant steering without caring about track limits.

    I have also modified Formula ISI to have several possibilities of low value steering lock angle to in order to simulate a constant steering position which should also provide a constant turning radius for a given speed.

    The first results are very interesting despite I need to polish several things before final tests:
    - I need to define a specific motec worksheet in order to account for main outputs.
    - I am thinking of a test which consists on two parts: a full straight initial part followed by a controlled turning part. At both parts the situation is maintained at each gear limit speed for 3 seconds before shifting up or down. The idea is that the overlay of different runs is easily achieved with motec.
    - I also need to check the influence of using rev limiter to simulate constant speeds. So far the different testing speeds are the maximum achievable speeds for each gear as restricted by the rev limiter. I am not sure if the power is constantly delivered at that rev limiting situation. I will compare those situations to the ones where the car is gaining speed with the uncapped power of the engine.

    Hopefully, I will post initial results soon.
    All the initial tests will be done with Formula ISI before using any other vehicle since the work is quite intensive and I am still trying to understand the best way to test things.

    Thanks for the interest shown.
     
  13. green serpent

    green serpent Registered

    Joined:
    Feb 2, 2016
    Messages:
    704
    Likes Received:
    713
    Reading the information in this thread and also doing some research, it's obvious to me that caster has a HUGE effect on steering feel.
    For example IRL if caster is 0°, there is basically no centering torque trying to straighten the wheel, you physically have to turn the wheel yourself on corner exist. If caster is really high, there will be a tonne of aligning torque, and the steering will be very heavy and stable.

    I had a 'spirited' drive today IRL , and I noticed that although it required some force to get the wheel turned (no power steering), once I had it steady through a turn it kind of wanted to keep going along the same path. There wasn't a tonne of force trying to straighten the wheel mid corner.
    I guess I've been comparing how sim ffb feels compared to my real life car, but if my car had a more race oriented caster setting (i.e 8°) it would feel totally different, and thus my perception of the sim ffb would be totally different. Most road cars have around 3-4° of caster (my car has 4°). Through a bit of testing, I personally think a lower caster setting in rf2 gives you more 'feel', possibly because you're getting more lateral tire forces acting on the ffb and less mechanical trail forces (aligning torque). But then things feel a bit odd when you push the front end...

    At least with a g27, when caster is set to around 3-4°, the ffb goes completely limp and loses all centering forces when you get even a small amount of front slip angle. For example in the Cobra, past a certain point of front axle slip, the ffb goes completely light (phenmatic trail decreases at the higher slip angle, causing less SAT?). It makes sense, but this to me is what 'felt' unrealistic. I've never pushed the front end grip that hard IRL to get the feedback through the wheel to go light. It seems to load up the more that I push, but maybe I have never really pushed it THAT hard to get proper understeer.

    It would be very interesting to do a track day and see if caster settings IRL somewhat relate to rf2. I.e with low caster the steering goes light with understeer, but with high caster setting the steering stays heavy due to more mechanical trail/jacking effect.
     
    lokidan likes this.
  14. lokidan

    lokidan Registered

    Joined:
    Nov 10, 2016
    Messages:
    35
    Likes Received:
    24
    There wasn't a tonne of force trying to straighten the wheel mid corner.

    this is my issue with enduracers mods, there just too much of it pushing the wheel. Whereas ISI cars theres not enough.
     
  15. lokidan

    lokidan Registered

    Joined:
    Nov 10, 2016
    Messages:
    35
    Likes Received:
    24
    I think the f2 is.prime.example that rf2 ffb is flawed in that car. It's not.just about understeering although understeer does cause lightness of wheel. It's not understeer that I'm.experiencing when doing 100mph Into a gradual curve. As I said I tried the skip and it doesn need caster its just some old cars. Maybe its an old physic's issue. And as said s397 are updating the f2

    ive driven a similar car to the f2 in real life and it never felt light like in the sim. If anything it was very hard work to work the wheel. Feels more like the caster set at full with full ffb. I know my t300 can't.handle that feeling. But it should feel better than it does currently.
     
  16. lokidan

    lokidan Registered

    Joined:
    Nov 10, 2016
    Messages:
    35
    Likes Received:
    24
    that feeling u describe about the f2 is my exact issue with it. I cant. imagine any car actually being like that, it would feel very unsafe and unassuring to drive.

    I had miles more confidence in the real life open wheeler, than I ever had driving the poorer ffb cars in rf2.

    let's hope they improve these for mid range wheels ☺
     
  17. green serpent

    green serpent Registered

    Joined:
    Feb 2, 2016
    Messages:
    704
    Likes Received:
    713
    Paul, in the sim do you feel the steering go light during under steer even with a higher caster setting? Personally I don't, but again... really need a DD wheel!

    I think what makes rf2 quite a challenge for your average gamer, is that if you apply too much steering input, the back of the car tends to step out (even without excessive throttle), and it can be quite hard to control. Essentially all other seems go into a lazy easy to control understeer, wheres in rf2, understeer can very easily turn into oversteer.

    I have come to expect this behavior and I'm getting better at controlling it when it happens and trying to prevent it in the first place, but for friends who jump on, it completely throws them off. Too much steering input leads to a spin almost every lap, and they label the sim 'too hard'. I'm constantly saying 'turn the wheel less', but they can't seem to get a feel for how much steering to dial in. I'm thinking I might lower the caster, and tell them to ease off the steering/throttle when the ffb goes light. I just wish that there was some ffb after the front end slides, but as Paul suggests, there is in fact still feeling with a dd wheel.

    Not that I care what others think of rf2, but it is quite pleasant drinking and playing racing games with your mates. Currently rf2 falls into the 'too hard basket' for them, and I'm stuck playing lesser titles. I think if you can wrap your head around the feeling of rf2, it can actually be easier to drive than other sims, however for a newbie facing a steep learning curve, other sims feel much easier and their perception is that they are more realistic.
     
  18. Euskotracks

    Euskotracks Registered

    Joined:
    Nov 29, 2013
    Messages:
    1,092
    Likes Received:
    191
    After completing the first tests I thought it would be nice to summarize them all in a brief report.
    I think the results are quite interesting. Not only from a FFB point of view but from the car behaviour as well.
    Hopefully someone can provide some info regarding the strange behaviour change when reaching 200 kph.
    I hope you find it interesting...

    https://www.dropbox.com/s/sx5b6gd792tiizd/Tires and FFB.odt?dl=0
     
  19. Korva7

    Korva7 Registered

    Joined:
    May 22, 2014
    Messages:
    700
    Likes Received:
    206
    Nice work.
    Are you sure that the steering lock(angle of front tyres in relation to car) stays the same during the tests? Do you have Speed sensitivity set to 0? I ask this because your slipangle also increases when you drop to 3 and 2 gear which i find weird. Like more steering lock was applied as speed drops.
     
  20. Euskotracks

    Euskotracks Registered

    Joined:
    Nov 29, 2013
    Messages:
    1,092
    Likes Received:
    191
    Thanks for your comments. Indeed I completely forgot about speed sensitivity. I was using keyboard default settings with speed sensitivity at 90%!!! Hence while steering input says 100% we don't know the real value.

    At least for all cases the rule is the same so the comparison between them is still valid. I did a quick test with speed sensitivity at 0% and the results showed higher values confirming the reduction of wheel rotation. I will run all cases again and update the document with the corrected graphs.

    Considering the speed sensitivity issue, I need to figure out a reliable way to monitor the real steering that is being simulated in terms of tire rotation. Maybe @Lazza can help with this since he knows the best what rF2 logs.

    I will also check if speed sensitivity is taken into account when representing graphically the steering at its locking positions.

    For what I have seen, the simulation of tires themselves seems to be correctly done. However, it would seem that there is no proper steering mechanism linkage simulation. Apparently rF2 calculates tire positions based on reduction of the steering system and controller.json modifiers.

    The fact that suggests me this, is that cars with no convergence trend to maintain similar slip angles for both tires even at in turn situation. This would imply a "perfect"steering system where both wheels would be permanently pointing at the right direction.

    It would be nice if a vehicle modder could confirm whether the real steering geometry is modeled or not. I would expect it necessary for both calculating real tire reactions and to properly transform those into torque at the steering shaft.
     

Share This Page