1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
Dismiss Notice
Like RaceDepartment on Facebook.

car.ini

Discussion in 'Racer Physics and Technical' started by eschefo, Sep 12, 2010.

  1. I’m studying the car.ini and have many doubts.
    1. The setting [body.height] in some models is the height of the body but in others is this plus the ground clearance. I know this settings are only for collision but ¿which is good?
    2. I think there is nothing about the body stiffness, then when the body roll, the weight to the wheels change in relation with stiffness of their suspension. As much stiffness more weight receive. ¿Is ok?
    3. I had one doubt about [car.front_area], but I see it dissapear in the lastest default car.ini. I don’t know the difference with [aero.body.area].
    4. ¿What means [engine.size]? ¿Is a volume? ¿Is the displacement? ¿Is used?.
    5. Looking the pacejka magic formula, for the transversal force I need the slip angle and it depends on tyre ¿wich is the variable of tyres/wheels that control slip angle?.
    6. The settings that control the geometrical effects in suspension ¿are always linear? and ¿wich is the point 0 were the effects begin? ¿restlen? ¿the lenght when the wheel is with weight? ¿?
    7. I see changes in camber, toe... but ¿the tracks change? ¿Racer changes them without a specific setting?.

    Last, I see there are many modelers in the forum (like Mr Whippy, boomer541, AMGFan, AlexForbin, Some1..) and I have doubts about models:
    1. I keep in mind that braking is between 1’0G (tourism) and 4’0G (racing) but I see models that breaks with 13’8G (Go-kart), over 2’0G (Lancia Thema, MB w25 Avus, R-18 GTD, Renault Master T35D, Smart Brabus, Toyota LandCruiser...). ¿May be?
    2. Also, I keep in mind resonance frequency between 1’0Hz (tourism) and 4’0Hz (racing) but many, many models are below 1’0Hz (even racig cars like Aston Martin DB9, Ferrari Dino, Porsche Cayman S (Mr Whippy) ......), ¿also may be? ¿how do yo know the stiffness of the suspension? ¿where do you get the information?
    3. And the same with aerodinamics, but I haven’t any solid reference, I only know that tourism use to have a negative Cd (go up with speed), but there are models with Cd=1’69 (AC Cobra), 3’55 (Go kart), ¿is ok?

    Thank you very much for read me.
    eschefo
     
  2. Great to see you're looking into this stuff and those are some good questions as well. Have you been reading up on http://www.racer.nl/reference/carphys.htm yet? Many things are explained there and some examples are given as well.
    If you work with car.ini files, it's recommended to enable note_ini_fallback in the racer.ini.dev section. This will show more info in qlog.txt which you can use to improve your files.

    1. body.height is the parameter that defines the height of the old style collision model, which is a simple box shape. Generally, it should only take the height of the body itself, it doesn't include the ground clearance. However, nowadays we're using specific low polygon meshes for the collision detection that resemble the actual vehicle body closely. See the default Murcielago as an example. This means the body.height parameter is only used for backwards compatibility and to satisfy the note_ini_fallback call that shows up in qlog.txt.

    2. Body rigidity is currently not simulated.

    3. aero.body.area is the correct parameter, the other one is non-functional.

    4. engine.size is a parameter that's normally not used, but Racer still looks it up - perhaps Ruud once intended to link some effects to it, but never got around to it. General practise is to set size=0.

    5. See documentation on pacejka - several parameters have an influence on the peak slip angle. Assuming you work in the pacejka player (pacejka.exe in the root directory of Racer), you can play around with the values and see what they do in real time.

    6. Yes, unfortunately all the wheel kinematics are linear. Their reference point is indeed restlen, which is the theoretical length of the spring without any compression (so, outside of the vehicle with no load applied).

    7. Track width does currently not change with suspension deflection.

    The accuracy of any vehicles behavior can differ a lot, since everybody has different levels of knowledge, experience and patience to (re)create vehicle dynamics. Many older releases are no longer working properly because Racer itself changed a lot over the years. Finally, there are obviously bugs in Racer as well, so not every weird behavior is the fault of the author.

    As to where you can find information - Racer's official documentation is a must read. Then, the forums are open for questions and we try to post useful information here as well. The biggest source for actual vehicle information is everything else - books, people you know, your own measurements and of course the internet has huge amounts of data.

    I hope this answers your questions to some extent :)
     
  3. Thank you very much, I need some time to study your answer.
    In 5. my doubt is if there are some settings to define the tyre/wheel stiffness over transversal slip (grip) because we need it to calculate the max transversal force.
    About the pacejka, I'm studying the pacejka89, somebody are using the new pacejka formula?
    All others ok.
     
  4. Vertical tyre stiffness is a fixed setting at the moment, car.ini.wheel.tire_rate (and tire_damping, also fixed). There's no dynamic adjusting going on as far as I know and it's not specified in the old pacejka code directly either.

    Some of us have been experimenting with MF5.2 Pacejka from the start, but personally I'm not commiting to it myself yet because of other physics bugs that I think would make it difficult to judge MF5.2 tests properly. In theory, it gives us greater control over some aspects and also includes combined slip calculations that aren't supported in the 89/96 versions.
    The good thing is that the principle doesn't change from old to new pacejka, so you're looking for the same characteristics and "curve layouts", with the added bonus that MF5.2 is one of a few current industry standards, which means that when you happen to find real life data, it's easier to implement it directly compared to before.
     
  5. Cosmo,
    sorry, my english not good.
    My doubt is how to setting the "grip stiffness" in Racer, I say the stiffness when the tyre receive a torsion moment not a vertical load. When wheel and patch have differents angles. I think this angle is the SA (slip angle, transversal) in Racer, no?
     
  6. And
    can I ask you about the car.ini you did? I see you do the car.ini of Nissan Skyline, Toyota sprinter, Alfa Romeo GTV6, Toyota Supra, BMW Alpina and M3, Chevrolet Camaro, Honda NSX.... and help in others.
     
  7. If I got you right, you want to look at pacejka parameters a3 and a4 mostly - they control the inclination of the linear section of the lateral force curve against slip angle changes.

    Sure, you can ask about vehicles I worked on, go ahead.
     
  8. Yes, I see that the transversal tyre stiffness depends of a3 and a4, also a bit of a5.... is difficult to understand. Also I see in the source code that slip angle is only the difference between the line velocity of the car and the line of the wheels without any relation with the transversal tyre stiffness.

    More doubts:
    - I don’t know where is the nullpoint and how can I know it, looking the car.ini. If I suposse is in the center of the body (half heigth, half width and half length) the results are no always logic.
    - The rotational inertias. The Iy (yaw) is easy, knowing weight distribution and wheelbase:
    - Iy = Weight * %front * (Wheelbase * %rear)^2 + Weight * %rear * (Wheelbase * %front)^2
    - The Ix (pitch) I think must be equal or greater than Iy because the horizontal distance is the same but there are alse heigth (difference between CoG and CoG of front and rear). The Iz (roll) also difficult because I don’t use to know where is the weight of each side (at width/4 from CoG?).
    Alfa Romeo GTV6 2,5
    - If I suposse null point is just in the center of the body is ok, ground clearance is 14’4cm.
    - Ix = Iy = 1.860 kg.m2 if the heigth are equal.
    - Iz = 111 kg.m2 if lateral mass are at width/4 from the center.
    - Of course, the fuel is not implemented.
    - If Racer don’t find a setting in the car.ini it looks in the default.ini, no? then Racer uses in this vehicle default fuel and default bumpstop.
    - About wheels, where you find values for pacejka? is very hard to understand and know if it correspond with a real tyre.
    Toyota Supra RZ
    - If I suposse null point is just in the center of the body, ground clearance is 29cm that is not normal, where is the null point?.
    - Ix = Iy = 2.618 kg.m2 if the heigth are equal.
    - Iz = 204 kg.m2 if lateral mass are at width/4 from the center.
    - Idem about pacejka.
    - You don’t set the optimal SA (slip angle) and SR (slip ratio) and I suposse Racer use ther default.ini. Racer can’t calculate them knowing the grip curves?
    Nissan Skyline 2000 RS Turbo
    - If I suposse null point is just in the center of the body, ground clearance is 12’4cm that is ok.
    - Ix = Iy = 2.083 kg.m2 if the heigth are equal.
    - Iz = 108 kg.m2 if lateral mass are at width/4 from the center. There is a big difference. ¿?
    - The turning diameter I think is not ok, but is easy to calculate: Turning diameter = 2 * Wheelbase / tan (wheel_front.lock/2), no?
    - Which is the formula by Justin Martin?
    - Idem about pacejka. I think the optimal SR and SA don’t correspond with the curve max values.
    - In the torque.crv, points=19, but there are more points (up to 25), many of them are over rpmmax, why?. Racer read them?

    I’m doing an aplication in excel@ for test the car.ini, but I’m not finished the rotational inertias, the bumstop, the antidive... and I’m testing the results of transversal dynamics with many models. Show you the test of yor models, in jpg looks very narrow.

    Thanks again
     
  9. I can't upload images ¿?
     
  10. I think in pdf is possible. Perhaps the size in jpg is too much.
     

    Attached Files:

  11. The nullpoint is actually an arbitrary point that the creator of the body mesh chose and we used the x/y/z offsets to define the CoG position relative to this nullpoint. This is bad practise and should be avoided however, because of a bug in Racer. What happens is that if you use large offsets, the vehicle will start to shake side-to-side - you can see it clearly from ~0.5m offsets if you want to try. Most cars use offsets of 0~0.25m, so the effects are harder to observe, but still present.
    The recommendation is to position the body mesh so that it's origin (nullpoint) and the CoG are identical, which means you can leave all CoG offsets at zero.

    Like you say, if Racer can't find a parameter, it looks it up in the default car.ini file, which is why I recommend turning on the option to "note_ini_fallback" in racer.ini.



    The three cars you sampled are from three different eras, but they're all more than two years old. At that time, we had a significant update to the suspension system in Racer and a lot of things changed around it as well. Also, my own work in the old days certainly looks and behaves oddly if I check it out now - I don't know if I should be happy to see people still driving them, 3, 4, 5 years later, or embarrased because of all the mistakes :)

    As for inertia numbers, I use a tool that Justin Martin wrote a long time ago. The results usually make sense if compared to real life data and it's very convenient. The tool has different "multipliers" for the results, depending on certain vehicle classes, which can explain some of the cars using odd values (if a non-ideal factor has been used). Link

    Justin also came up with a handy rule of thumb to estimate CoG height. He found that as a general guideline, taking 37% of the vehicle's roof height will get you very close to a realistic CoG height value. We generally use this approximation on vehicles we don't have more precise data for. This is the formula I credited him for in my car.ini files.

    Useful tyre data is obviously hard to come by - you can google around and find many graphs and such easily, but not all of them are real, make sense or use a format we can work with. In the early days of Racer, somebody named Alpine provided a few sets of road car tyre data which are our main source for basic range of values we use. In actual setup work, you don't change all the coefficients, all the time - you only work with a few of them to set the absolute grip and the compound characteristics. For the rest, we leave the basic numbers that Alpine provided because from experience, they work best and are "real". I attached them at the end of the post.



    Now let's look at the individual cars:

    The Nissan is the oldest here and it was a quick project with little data available - a lot had to be guessed. If you load it up in a modern Racer version, a lot of parameters will be missing from the car.ini file, so I don't if it's even drivable at the moment. For example, ackerman was defined differently back in those beta versions - the resulting steering angles are vastly different left and right as a result of this change in Racer.
    Optimal SR/SA were a topic of discussion between Outlaw, Mr Whippy, myself and others for some time, I'll get back to this on the Supra.
    The odd points on the engine curve files are something I never noticed, thanks for pointing that out. Since I use curved.exe to create the files (and the extra points are not visible there), I never saw them. My explanation is that at some time I used an existing engine curve to start my own torque graph, and when I deleted the excess points, curved.exe removed them from the screen, but not from the code? Going through my engine curve library, I can see that the graphs I started from scratch are fine (ie, used a new file), but out of convenience, you normally don't do that because you don't want to type the engine speed, flywheel torque etc... I'll keep it in mind from now on.
    In general, we define a bit more than what maxrpm is set to, so that the car still responds well when you downshift early etc, but not like it is in these cases.

    The Supra was a joint venture with Outlaw and during that time, we tried to find out more about what's going on in Racer, why some things happen and so on.
    You picked up one such experiment, the missing oprimal SR/SA values. Back then, we asked ourselves what Racer is actually doing with them and (incorrectly) thought that because pacejka.exe automatically calculates these values, racer.exe would do the same... so we left them out for one or two vehicles I believe. Similarly, on the Nissan, I wanted a certain optimal slip characteristic, so I simply "forced" it instead of using the calculated values.
    With Pacejka MF5.2, we should get better, combined (dynamic?) SR/SA behavior eventually.

    The Alfa Romeo and the BMW from the same time period, I'm not sure about anymore - they felt food back then, but I don't like their handling anymore when I try them now. The tyres are, like on the Nissan, based on Alpine's 195 R15 data, I believe. They're the last from the "old breed", I changed my setup work a lot afterwards, with the introduction of improved spring length definitions (to allow pre-compression), damper curves, bumpstops and so on.



    Your spreadsheet seems like a great tool, reminds of TRex by Niels (rFactor).

    Code:
    225/45 ZR 17 on 7" rim, 2.4 bar
    
    ; Lateral force
    a0=1.6
    a1=-38
    a2=1201
    a3=1914
    a4=8.7
    a5=0.014
    a6=-0.24
    a7=1.0
    a8=-0.03
    a9=-0.0013
    a10=-0.15
    a111=-8.5
    a112=-0.29
    a12=17.8
    a13=-2.4
    ; Longitudinal force
    b0=1.7
    b1=-80
    b2=1571
    b3=23.3
    b4=300
    b5=0
    b6=0.0068
    b7=0.055
    b8=-0.024
    b9=0.014
    b10=0.26
    b11=-86
    b12=350
    ; Aligning moment
    c0=2.3
    c1=-3.8
    c2=-3.14
    c3=-1.16
    c4=-7.2
    c5=0.0
    c6=0.0
    c7=0.044
    c8=-0.58
    c9=0.18
    c10=0.043
    c11=0.048
    c12=-0.0035
    c13=-0.18
    c14=0.14
    c15=-1.029
    c16=0.27
    c17=-1.1
    
    195/60 HR 15 on 6"1/2 rim at 2.1 bar
    
    ; Lateral force
    a0=1.3
    a1=-49
    a2=1216
    a3=1632
    a4=11
    a5=0.006
    a6=-0.04
    a7=-0.4
    a8=0.003
    a9=-0.002
    a10=0.16
    a111=-11
    a112=0.045
    a12=0.17
    a13=-23.5
    ; Longitudinal force
    b0=1.57
    b1=-48
    b2=1338
    b3=6.8
    b4=444
    b5=0
    b6=0.0034
    b7=-0.008
    b8=0.66
    b9=0
    b10=0
    b11=0
    b12=0
    ; Aligning moment
    c0=2.46
    c1=-2.77
    c2=-2.9
    c3=-0
    c4=-3.6
    c5=-0.1
    c6=0.0004
    c7=0.22
    c8=-2.31
    c9=3.87
    c10=0.0007
    c11=-0.05
    c12=-0.006
    c13=0.33
    c14=-0.04
    c15=-0.4
    c16=0.092
    c17=0.0114
    
    225/60 HR 16
    
    ; Lateral force
    a0=1.9
    a1=-41
    a2=1210
    a3=2180
    a4=10
    a5=0.014
    a6=-0.023
    a7=0.67
    a8=-0.051
    a9=-0.018
    a10=-0.06
    a111=-2.35
    a112=-0.37
    a12=4.3
    a13=-14.9
    ; Longitudinal force
    b0=1.5
    b1=-5.2
    b2=1190
    b3=26.7
    b4=255
    b5=0
    b6=-0.000093
    b7=0.05
    b8=0.49
    b9=-0.007
    b10=-0.23
    b11=141
    ; Aligning moment
    c0=2.53
    c1=-3.
    c2=-6.6
    c3=-0.56
    c4=-8.9
    c5=0
    c6=0
    c7=0.016
    c8=-0.39
    c9=0.35
    c10=0.014
    c11=-0.014
    c12=-0.006
    c13=-0.15
    c14=0.023
    c15=-0.89
    c16=0.025
    c17=-0.4
     
  12. A lot of thanks,
    I'll try to digest all the information. There are many threads in the forum that I don't know.
    The link for the aplications is over, but I have both the inertiawheels.xls and carInertia.exe (sorry, I don't know the author are Justin Martin).
    I begin to studying the car.ini wiyh the intention to do a new car, I found there is no applications to help (RacerEng.exe is only to write the values but dont' teach).
    My future intention is to do an aplication to create the car.ini from the usual car knowledge, but the first problem is to know how works Racer and the car.ini.

    A lot of help in your mail and thanks by your honesty, this help me to understand many things. I see your new Hummer and I'll study the car.ini, to see a new car.ini version.
     
  13. Niels' isn't done with the his tool and it's not available anywhere yet, but as you said, carfactory has been out there for a while now. I think that some of us in the Racer scene have at least partially created something like this, too, but not for "everything" in one place. Like the inertia spreadsheet you got, it's mostly scattered around.

    Would be interesting to see where you can get with your spreadsheet, it looks comprehensive already.
     
  14. Hey, great to see somebody working on the car.ini. I remember that Lance Truong did a new pacejka from real test (his Lancer Evo VIII). I've used these settings in my releases (a long time ago), the lastest one was an Evo IV (stock) and felt pretty nice. Maybe you can see it to compare agaisnt the Apline's pacejka and the others "good" ones.
    Also, BMWSpain helped me in the later era for the inis.
     
  15. I'll continue, slowly but constant. I suposse you can help me with my doubts (thanks again, Cosmo) and telling me what are the most interesting for a modeler, I really never did a vehicle in Racer.
    I see "El Tate" did the car.ini of many interestings models: Nissan Skyline (other that Cosmo), Alfa Romeo 155, Chevrolet Camaro, Dodge Viper and Mitsubishi Lancer. And helps in other like De Tomaso Mangusta, BMW 325i, Fiat 125, Porsche 911............. and more !!.
    First, I send you my excel aplication (with a little of Visual Basic) translated to English (I do it in spanish) and a resume of all the physics I'm using to get the values. Is not finished but yo can see if is useful and what to change.
    I hope there are not an equal application for Racer and that I'm working for nothing, if sombebody did it.

    I send you a worksheet with the models I download to work and play. Is arranged by authors of the car.ini and by manufacturer and year.
     

    Attached Files:

  16. Just a quick note.. the Plymouth Barracuda 440-6pack uses RyanT's mesh, other than that the rest is changed (mod by Luthobu, with help on the ini from AMGFan and Splashonda). Ryan's original Plymouth Barracuda 340-cid is however missing from the list. The Fury III was done by Splashonda.. with a little help from his friends.

    Just to make it clear :)


    EDIT: Actually, I don't understand what this list is.. so disregard that post..
     
  17. Nice list, but i'm afraid you only take in count only the car.inis which came with the car, but not the ones from the "updated car.ini" thread. There i've posted some inis for the cars who did not have good ones.

    I think that is lost that, when moved to this new forum.
     
  18. I download almost all the cars from "tracciontrasera", some from "xtremeracers" and a few from this forum, and only show in the list the models I can download, other I couldn't are not showed.
    I can improve the list but I do it only to have a reference of the vehicles available and their authors. Is really useful? it worth to improve?
     
  19. Well... it's useful if you want to have an detailed (by schedule) car list, and car.ini specifically; So you can know which ini to view in order to get some values and know which is "good" or which is "old", "bad" or crap directly.

    And i don't know why we're talkin in english since both we talk spanish jajajaja.

    Cheers