The .tgm, and +ttool

Discussion in 'Car Modding' started by Tim Bennett, Jan 17, 2012.

  1. jtbo

    jtbo Registered

    Joined:
    Jan 12, 2012
    Messages:
    1,668
    Likes Received:
    48
    I wonder what will happen if only 3.8 litres of wheel volume is used, but that I think we see in hour or so if nothing crashes.

    13"x13.7" is 2.49... times wider than my sample wheel, maybe those wheels have also larger air spaces too because of shape so that cylinder method works better? I can't remember how such wheel looks at the moment, must look from interwebs.

    Everything has it's reason, but we can only guess why such route has been chosen until some wise wizard passes infobits down to our direction.
    I let you know what happens with this tire, it is more or less just a test to learn system better, so it will be interesting to see if there is any notable difference with it and cylinder method one that I must make some time to compare the two.


    BTW, it is perfectly possible to drive in single player at the same time while tire tool computes, not even much performance difference and my cpu is not very fast, E6750 only, sadly dev mode is not launching, that would be nice to be able to mod while computing goes on background using other core :p
     
  2. Tuttle

    Tuttle Technical Art Director - Env Lead

    Joined:
    Feb 14, 2012
    Messages:
    2,480
    Likes Received:
    775
    With a such small rim volume you'll get a huge current air volume...:)

    ..and yep, those rims are like cylinders with a flange (J bead) and almost zero humps (flat humps) after the bead profile...so the basic Cylinder formula could work for sure.

    BTW...if you've a closed (and clean) model (mesh) of the rim width profile you can try a volume calculation inside 3ds max, or Sketchup...or autocad etc...

    When I work with those TGM I feel myself like Buzz Aldrin on the moon...every little step is a huge discovery..:D

    PS: did you try to press CTRL+SHIFT (left side) in devmode*? You get a sort of pressure test on tires but I don't know what is the real purpose of this action. It's interesting as you can see how tires works under a strong stress.

    * To see the TGM display you've to change this line in the devmode PLR: TGM Display="1" // Temporary variable
     
    Last edited by a moderator: Jul 15, 2012
  3. jtbo

    jtbo Registered

    Joined:
    Jan 12, 2012
    Messages:
    1,668
    Likes Received:
    48
    I discover new all the time I read this board, thx from CTRL+SHIFT tip, I must test that too, first time I find that, maybe I have missed it sometime :D

    My car is like it would be on ice when I do CTRL+SHIFT and hold it down, slides to even slightest hill.

    Then I hit some button (accidentally) while keeping CTRL+SHIFT down and got all kind of yellow thing and arrows and text saying local, after I hit esc and hit race again I got CTD, then I felt like Kabooze in Red vs Blue, which I feel most of the time, but especially when doing tire stuff and thought that I try that again. This time without arrows and text with yellow colors all over the place.
     
  4. Tuttle

    Tuttle Technical Art Director - Env Lead

    Joined:
    Feb 14, 2012
    Messages:
    2,480
    Likes Received:
    775
    I was thinking about tyre mass as it is calculated by the ttool;

    We get a TotalMass and a RingMass.

    Now, I guess despite the parameter name (TotalMass), the real tire mass should be the Total+Ring.

    I'm asking this because I'm getting unrealistic results if I get the TotalMass as the real mass of the entire wheel...but if I add the TM to the RingMass I get a realistic result if compared to the actual mass of the real tire.

    I guess this info could be important for the suspension model as well, where we've to put Tire mass and Tire inertia for front and rears...and (maybe) we can use TGM values for mass values into the .PM file.

    If my thought is correct I've to understand how to calculate the total inertia for the .PM file...but I think we could use just the TotalInertiaStandard for the 3 entries in the .PM tire inertia.

    What do you think about this? Am I raving? :p

    PS: I've tested this procedure and I've to say the result is pretty good....(placebo?)
     
  5. Marek Lesniak

    Marek Lesniak Car Team Staff Member

    Joined:
    Oct 6, 2010
    Messages:
    1,585
    Likes Received:
    101
    Wheel inertia in PM is not only about tires but for the whole package: tire, rim and brake disc + disc bell, if present (and also half of a half-axle, for the rear wheels inertia).
    And for tires, I hope we can use values from ttool, because for rF 1, I had to calculate that stuff by myself (using simple and general methods) :)
     
  6. Tuttle

    Tuttle Technical Art Director - Env Lead

    Joined:
    Feb 14, 2012
    Messages:
    2,480
    Likes Received:
    775
    I see the ttool calculations for the total inertia get me a result very similar to my previous calculations for the PM...

    BTW, I know .PM body values are for groups but usually they're average or underestimated...
     
  7. jtbo

    jtbo Registered

    Joined:
    Jan 12, 2012
    Messages:
    1,668
    Likes Received:
    48
    This is what reads in quickstart manual:
    Then there is this one too:
    Could it be possible, that ring is everything apart from those that are part of rim and total includes parts that are as part of rim too?

    Wheels themselves are different weight and there is no possibility of ttool to know how much inertia they have as wheel weight is not given to ttool?

    Wheel + tire combination can be easily be 25kg, add brake and hub to that and it is quite lot of mass, but I guess most of the inertia is at tire as sidewalls and beads are not that much compared to thread that is most of the weight, imo. Also at it is outermost section that weight has more to say to inertia sense than other parts.

    I have impression that thread area can weight twice what rest of the tire weights, but that is just my estimation when I did cut one tire, just hand scale, not really weighted sections.

    It really is guessing in the dark, but I'm rather skeptical total wheel mass or inertia being there, ttool in my opionion is not aware of wheel, only edge of wheel where tire connects to wheel and thus not including really total wheel mass or inertia.

    For testing, here are wheel weights of normal wheels, also there is link at top of the page to weight of tires, might be interesting to do some wizard stuff and compare if total + ring still would be close, which I doubt, but testing is bit above my skills, but there is real information to base calculations.
    http://www.miata.net/faq/wheel_weights.html


    edit: I got load of 3066N now with ISI 60's formula tyre that I just changed size, now I put machine to calculate lookup table and then I try to figure out how to use that 3066N to get temporary bristle spring and damper values, I guess that default 16000 might be good starting point.

    edit2: After evening nap machine did manage to compute lookuptables and I did grip adjustment to realtime section that I copied from ISI tire, then loaded file again to ttool and found my peak radius, put that to ground plane level and did testing, which now resulted Load of 4626N, which I believe means that I should try to find lower tempbristlespring setting than originally there was and perhaps adjust damper accordingly to maintain hysteris at similar level. But then again it is area of more questions that answers, quide states this:
    I don't have quasi static analysis section, I have quasi static testing and quasi static model, but no analysis, in QSM section I have load, in QST section I have ring test force (Y) which I believe is one I should use to try to determine deflection without using ground plane?
    Realtime test parameters section has deflection, but it is only one where I can see deflection, imo and I can set that to be what I wan't.
    It is bit unclear in manual what is QSA referred in this case (sometimes it seem to refer to whole model, sometimes program interface) to and when program interface is referred to and when realtime section in TGM file. But this can always be my imperfect english skill too, however thought to bring this up so someone might know to push me towards right direction (again) :)
     
    Last edited by a moderator: Jul 16, 2012
  8. jtbo

    jtbo Registered

    Joined:
    Jan 12, 2012
    Messages:
    1,668
    Likes Received:
    48
    I had bit difficulties to understand QSA and vertical components, so I started writing it to my terms to understand it better, but there can be errors and misinformation, at least spelling errors as this is just raw writeup of thoughts that needs to be modified quite bit to be useful for others, but if someone spots errors, could you please post about it so I can refine this document, that might help then someone in future?

    Edit: No need to check spelling, that I should be able to take care of, but I'm not 100% certain if I have understood method right way around.

    Code:
    Things to do after lookup table is generated.
    
    First one needs to go to bottom of menu, Quasi static testing section, from there iterations per frame box enter
    number between 10-30, then wait several minutes until percent done (bit lower in menu) is 100.
    
    After that push perform one iteration and computing stops. 
    
    Now from Quasi static model section you can find peak radius of your tire, write this down as it is needed next.
    
    
    Next test is for finding out load value (it is in Quasi static model section) for getting proper load
    we need to adjust ground plane so that tire would deflect same amount that it deflects in real world.
    So jack up the car, measure sidewall height, lower the car and measure sidewall height, difference is the deflectin amount.
    Now take peak radius that you wrote down and substract difference of sidewall height measurement from it.
    
    You got new number that goes to Quasi static testing section, there is box labeled Surface height, input number there and remember
    to put minus '-' to front of number, then we are ready to start calculating load.
    
    Type again number between 10-30 to iterations per frame box (Quasi static testing), wait few minutes and when percent done is 
    100% hit perform one iteration to stop computing.
    
    Now we can go up to quasi static model section and find box labeled Load (N) that is tire load required to get that deflection.
    Write that load down.
    
    
    Next becomes part where we adjust temporary bristle spring value in TGM file, it resides in file, section named realtime, between nodes and lookup table.
    
    First it is needed to get some idea what way to adjust bristle spring, in ttool go to realtime test parameters section.
    type deflection you did measure earlier from car to box labeled deflection, adjust nominal radius and width to correct numbers.
    Remember peak radius that we got earlier, put that to nominal radius. Make velocity and rotation zero
    Now from very top of realtime test parapmeters section switch 'Run Physics' to on, wait until 'vertical force' in Realtime test results section
    stabilizes, after that you can switch physics off. 
    
    Now compare vertical force from realtime section to Load(N) that you did wrote down, these should be same, if not then from TGM file you need 
    manually edit temporary bristle spring 2nd parameter, middle one, labeled as X here (0, X, 0). 
    Increase bristle spring value to increase the load at realtime testing section, edit TGM file, save, load tire to tool, do adjustments as previously (mainly ground plane)
    and test again. 
    
    Remember that when you load tire, it resets all Quasi static testing and Quasi static model adjustments and results.
    That is why everything you might need later is good to write down as to get values again it is several minutes of computing.
    
    Also remember to adjust pressures! Tire pressure should be adjusted from realtime testing and quasi static testing sections to be 
    correct pressure as in your car. 
    edit continues:
    Ok, so I made tire from ISI 60's formula front tire, changed size, made new lookup table, adjusted bristles according to writeup above.
    What I notice is that there is ton of grip, 325bhp with 3.90 rear end can't really make a burnout with default grip parameters, at skidpad I can get around 1G with 1960's american car whom suspension I have made with car factory and factory data as much as possible, it has positive camber and zero caster!

    I would need to go below 2.0 and 1.0 to get enough low numbers at skidpad and car becomes ice skater then and also what I have understood such low values should not be used.
    But tire is 60's bias ply tire that ISI has made, it should have 60's materials in it and from my understanding it should not give that much grip, but maybe I have mistaken once again, it happens.

    I did test crazy values to bristle spring and I did manage to get somewhat nice grip at acceleration, slalom and skidpad, but then car did floated in rather weird way when racing around the track.

    With Brabham at Mills I did corner easily 1.3G, peaked at 1.5G according to Motec, that seems bit high to me for threaded 60's bias-ply tires and no wings?
     
    Last edited by a moderator: Jul 17, 2012
  9. jtbo

    jtbo Registered

    Joined:
    Jan 12, 2012
    Messages:
    1,668
    Likes Received:
    48
    Sorry, this is triple posting and beware this is really long :)

    I thought about that tire grip and have been now looking different tires from ISI, comparing different materials I try to form understanding of how material affects to grip of tire and if it does.

    I take TGM files of 60's formula and Megane and compare material of tire thread between two of them, please comment if I make error in conclusions again. Mostly this is just writeup from my current understanding and attempt to write it down so I can explain things to myself when I forget things next week, but this stuff might help someone and might be something that can be improved so it is more useful, this will be very long post.

    Think about plastic, plastic at your keyboard is hard, it is not flexible, but other plastics can flex quite a bit at same thickness still they are plastic, that plastic is not really definition of material, but more of class of materials, now PP (PolyPropene?) plastic has already more narrow window of parameters for it to be PP plastic. Rubber used in tire contains several substances and without knowing those it is difficult to make exact material, of course that is the big secret in tire industry, but we should be at least be able to form some window where to put our values to avoid totally out of the whack tires.

    For both tires I chose Node 10 and I put Legend from tire manual between them to make it easier to read which value is which.
    [Node] // 10
    TreadMaterial=(273.15,837,8000000,0.47,-1,1900,0.188) //60's formula
    TreadMaterial=(373.15,814,5900000,0.47,-1,2200,0.18) //60s formula
    Temperature (Kelvins), Density, Young's Modulus, Poisson's Ratio, Damping Factor (leave at -1) , Specific Heat, Thermal Conductivity
    TreadMaterial=(273.15,950,8150000,0.47,-1,1950,0.172) //Megane
    TreadMaterial=(373.15,925,6030000,0.47,-1,2200,0.165) //Megane

    From my understanding these are along lines I describe below, every material has it's own real world parameters, but many materials are mix of more than one substance, what kind of mix that material is affects to it's parameters in certain range, values outside that range are unrealistic, but currently I don't have very good understanding of material engineering so that I could tell what kind of ranges these materials can have.

    So let's get to business.

    Temperature has been same for all tires I have seen, so I assume it is safe to use those values for now.

    Density affects how much weight that material adds to tire and also temperature, how fast tire gains heat, how fast it cools etc.
    We can see that thread in Megane tires is made from denser material than thread in formula tires, this means that if tire would be same size and same depth of thread, megane tires would have higher inertia and weight, also they could 'store' more heat.

    Young's modulus, there are small differences in these, Megane tire thread being made from bit stiffer material.
    Wiki says:
    Link
    Engineering toolbox
    Higher value tends to be stiffer and lower value tends to be less stiff, I imagine this being how well it bends, imagining having 1m long rod of material and attempting to bend, lower values bend better, but can this be related to rubber shore value which is softness of rubber? edit: Answering to myself, Yes it can, scroll down to edit where I represent how it relates to hardness of rubber value.

    Poisson's ratio, both tires have 0.47, it probably is not much different from that between tires.
    When you have brick shaped piece of fudge (soft toffee for some parts of world), you heat it in hand and then start pull or push it, you can notice that it is getting thinner or thicker from sides depending from if you push or pull and Poisson's ratio describes how much of this effect does happen. According to Wiki 0.5 is maximum value, negative values being possible, but for rubber 0.5 is mentioned. 0 means there is no expansion to transverse direction when compressed and indeed 0.5 ratio being the maximum. How exactly this affects to tire is bit unclear to me, but then again we probably don't need to mess with this value much if at all (naturally for plies and different material plies this needs to be adjusted, but for simplicity let's focus only to thread material for now).
    Wiki

    Damping Factor is not implemented yet, it is told to be left at -1, so until tire model is finalized this does not need attention. I would believe that this replaces Temporary bristle damper setting we currently use.

    Specific Heat This seems to be same for both tires, here is link to thread about specific heat, all materials should have own range for it, but this does require more research, anyway one should choose value for material he uses for tire, but when it's nearly impossible to know material tire makers use, we can only guess that maybe ISI had some insight when chose to this value for thread material.
    http://www.eng-tips.com/viewthread.cfm?qid=131840

    Thermal ConductivityThere are small differences, Megane tire thread seem to conduct less heat than 60's formula tire. Smaller value means material conducts less heat. If you heat metal bar from one end and hold from other end your hand soon gets burn, but you can heat brick for long time until it becomes hot from other end, so metal conducts heat better than brick. This affect how fast tires will heat and cool, together with density this parameter affects tire heating and cooling. All material have also their own heat conductivity which should be respected when making tires. From Wiki page you can find link to page that has table of different materials and their heat conductivity, which helps to learn more from possible ranges that might be possible to use.
    Wiki


    How to make tires grip less then? I would imagine that increasing Young's modulus for thread material should give tire that slides more easily as rubber mix would then be harder, but with this new understanding of tires, I must do experiments to learn if that would really work.

    Of course there are these two in Realtime section of TGM file:
    StaticBaseCoefficient=2.050 //60's formula
    SlidingBaseCoefficient=1.5375 //60's formula

    Those I believe are used to fine tune grip level? My experience is that these alone can't be used to get tire grip to 60's street tire levels, more need to be done and my current bet is adjusting Young's modulus to represent different rubber mix that was in street tires of 60's, but this is all very much of guessing and feedback is appreciated from this :)

    edit: I found something interesting, so I just put it down here.

    Here winter tires are often compared with Shore A rating, hardness of thread rubber, 60 is now typical for winter tires, 70 for summer tires.

    There is also other scale used for measuring hardness of rubber, IRHD, International rubber hardness. while searching if units are comparable at all, I found information that it is impossible to compare, but then I found scientific paper which from I understand that they match very closely 1:1 each other, especially for our purposes.
    Link to paper

    Then I started looking into over 700 pages long document that Gijon posted at early days of beta, Pneumatictire HS-810-561.pdf and at page 33 there is table with IRHD values of thread and their respective Young's modulus values. Page 30-33 are good part in there also that there are told methods to get modulus for different temperatures, as temperature increases rubber stiffens, which is interesting actually.

    Because I'm not able to copy/paste tables on forums, I just represent value for hardness values near 70 that is for street tires with summer thread.
    IRHD 60 = Young's modulus (E)Mpa 3.6
    IRHD 70 = Young's modulus (E)Mpa 5.5
    IRHD 80 = Young's modulus (E)Mpa 9.5

    ISI tires are around (E)Mpa 8 and their modulus decreases with temperature increase which confuses me, maybe something I have understood wrong.

    Also there is Tensile modulus and Elastic modulus mentioned, but both are labeled (E) and also mentioned to be Young's modulus, I have not studied enough yet to understand if there are differences that should be taken account into.

    Sorry that I have no link at hand to that over 700 page pdf, but I believe such is edited to first post of this thread, I might edit it here back later.

    Until next time...

    edit2: Ok, I thought this bit from different angles and this is my theory of why I have had difficulties getting enough low grip to feel good.
    Rubber is too hard.

    Sounds opposite of what it should be?

    I think it this way, softer rubber has more bite, so lowering BaseCoefficient should not make things so easily ice skating as with harder rubber, rubber does not bite so much and fails to keep contact patch at static and goes to sliding friction very easily, while softer rubber might maintain static state longer for larger part of contact patch and thus allowing lower coefficient to be used.
    But this is just a theory, I need to test this in practice next, by changing material settings of thread to shore 70 level and compute tire's lookuptable and do all that routine, after that I should know what will happen when I adjust BaseCoefficient to lower values in attempt to reach 0.66g with Firebird. Not exactly best scientific method, but at least experiment should provide some new thoughts or lead to dead end, we will see that later.

    edit3: 2006 paper of then new Carbon black material with some good number data:
    http://www.sidrich.com/PDF/Papers/Paper200609.pdf

    edit4: Here is some paper where one can find high performance tire properties, there are inner liners, sidewalls and treads of tire, some different compounds and generally lot of good stuff:
    http://www.thefreelibrary.com/A+review+of+isobutylene-based+elastomers+used+in+automotive...-a054117641

    Never had chance when I tried to find such, but as I tried search relation to shore rating and wear index (which there is not) with terms tire tread wear index shore a table it seems that I found a lot of other good tire stuff.

    edit5: Micro and macro textures of pavement explained relatively well, I did browse from page 40 to 47 of pdf and it had good pictures to illustrate those things:
    http://onlinepubs.trb.org/onlinepubs/nchrp/nchrp_w108.pdf

    We have these at TGM file and above pdf might become useful for adjusting these, if there would just be understanding of what each number represents and why some are negative and some are not.
    SlidingAdhesionCurve=(-9.2, 0.4, -4.7, 1.7, -0.2, 0.2)
    SlidingMicroDeformationCurve=(-5.2, 0.3, -0.7, 1.8, +2.8, 0.3)
    SlidingMacroDeformationCurve=(-1.2, 0.2, +2.3, 2.0, +6.3, 0.35)
     
    Last edited by a moderator: Jul 19, 2012
  10. Tuttle

    Tuttle Technical Art Director - Env Lead

    Joined:
    Feb 14, 2012
    Messages:
    2,480
    Likes Received:
    775
    I think this paper could be interesting and useful for a good tire model...:)
     
  11. Diablo

    Diablo Registered

    Joined:
    Jan 20, 2013
    Messages:
    404
    Likes Received:
    0
    Broken links corrected

    I just stumbled in here, since the wiki links to this thread. Those two links are broken, which I guess has to do with the forum shortening them with the "...", so I put the correct link as plain text:
    http://www.nhtsa.gov/staticfiles/safercar/pdf/PneumaticTire_HS-810-561.pdf

    BTW, seeing as both those links end in "HS-810-561.pdf", I think they are actually the same, so why the distinction?

    Edit:
    is dead.

    Best regards,
    Marcus
     
    Last edited by a moderator: Feb 3, 2013
  12. fuin

    fuin Registered

    Joined:
    Aug 21, 2012
    Messages:
    5
    Likes Received:
    0
    Has anyone tried making small tires? Because I've been trying to make 62 mm diameter tire, the tire tool gets stuck on 4 of the automated test index. i have tried lowering the DT multiplier all the way down to 0.2 but its still stuck.
     
  13. mschreiner

    mschreiner Registered

    Joined:
    Oct 5, 2010
    Messages:
    390
    Likes Received:
    262
    ok I have read this and a lot of stuff is not very clear.
    How do I set the size of the tire? Let's say it is 88 inches for the radius of the tire I am working with. Were do we actually put the radius to have the proper size?
    Also, let's say the tire is 11 inches wide. Where do we enter this to get the proper width?

    Next,
    I want the tires optimal pressure to be around 45 psi. Were do we put this entry as well?

    Last,
    I want the optimal temp to be around 220 degrees. Where doe the optimal actually go?

    I understand I will need to convert the #'s but that is not hard. I just need to be clear on where all of these go and then what to do with ttool to make these #'s work.

    Would be nice if someone from ISI would post a better step by step procedure that is easier to understand than the pdf provided.

    Something simple without all the math explainations for us to follow.
    Example: (I know this is not proper order or anything)

    Step1:
    Open ttool
    Step2:
    Load .tgm file located in ptool folder
    Step3
    Tell where to enter tire size and width
    Step4:
    Tell where to enter optimal temps
    Step5:
    Run QSM tests or realtime

    Something a little more simple would be very helpful to modders.
    We really do not need to know all the math we just need to know what variables need plugged in to the right places and what they do.

    Maybe even go Line by line on the ttool menu and explain what it does or affects and when it needs to be run or used.

    I have figured out a bunch of stuff by guessing at things but dang you guys made it a lot more difficult to do things.
    It would not be so bad but it takes about 3-4 hours to run tests and that slows me up a ton.

    Well if anyone can answer my questions I would appreciate it.
    And if ISI could put together a more simple to understand tutorial so we can better understand this tool it would be appreciated by all of us so we do not have to guess at what does what to the tire as much.

    Regards,
    Mike
     
  14. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,346
    Likes Received:
    6,572
    I believe radius and width are part of the geometry, so you can't just edit some numbers to adjust those (well, actually the geometry is defined with numbers, but it's a lot of numbers and would be easy to break the tyre by changing them, so you'd need to work at it)

    The [Realtime] section is where you'll find numbers you can edit more easily, but you need to look at pretty much all of them and what they say to work out what they refer to and how they work.

    For your optimum temperature, the main line is this one:

    StaticCurve=

    Each pair of values in the () is a temperature and a grip coefficient. So this:

    StaticCurve=(173.15, 0.5, 373.15, 1.0, 473.15, 0.5)

    Would indicate full grip (1.0) at 373.15K (100°C), and half grip (0.5) at 173.15K (-100°C) and 473.15K (200°C), for example. You'd obviously want to convert your °F to K first.

    The whole process (from scratch) is pretty much like designing a real tyre I'd imagine, but being able to set the properties of the compound(s) rather than literally having to create them. It's not as easy as it was in rF1 to 'fake' a tyre with a few target figures in mind - but a lot of rF2 is heading that way, which is great for realism but a pain for the less knowledgeable modders. We probably need a fair few ISI examples covering various sizes in order for most people to have a base to work from.

    *I should say there is a SizeMultiplier=(1.0,1.0) line, which can be used to resize a tyre in either dimension, but I'm not sure what adverse effects that could have if you overdo it.
     
  15. mschreiner

    mschreiner Registered

    Joined:
    Oct 5, 2010
    Messages:
    390
    Likes Received:
    262
    Thanks Lazza!
    Yeah this is way different than rf1. I have been modding since 2005 and ISI did not make it easy to convert our work over....LOL
    It would just be nice to have some better examples and explainations out there for the modders. I think after playing with the ttool for the past 2 weeks I have a better understanding and should be able to make a proper tire size by moving my nodes around. If I figure some stuff out I'll post my findings to help the next guy.

    Thanks,
    Mike
     
  16. dv8r

    dv8r Registered

    Joined:
    Oct 13, 2010
    Messages:
    55
    Likes Received:
    0
    I have been testing this tyre for some time now. Over time ive created some likable tyres which seem to respond in a predictable manner.

    As i narrow in on the right level of grip i need to tune specific characteristics. I would like to increase braking performance without lockups, keep lateral G loads as they are, but create more slip under acceleration. I know there are other factors that impact this with vehicle physics, but the values i have in those areas are where i need them. I have a few ways to try and fix this, but how would others go about addressing this scenario. Im interested to know which variables you all turn to as a starting point, because ive found theres many ways.
     
  17. Rocksor

    Rocksor Registered

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

    I think this is the right thread for general ttool and TGM questions. So here for me it doesn't work out at all. I wanted to start by resizing ISI Tires to the size I need by adjusting the node geometry. In Addition I adjusted the size in the tbc. But when loading the car to track in developer mode the car is still floating in the air. But the strange thing is, that it doesn't change no matter if I switch between the tire models. I'm sure the tire files are referenced. Checked that quite some times.

    Could it be that I have to generate a new lookup table in order to make the tgm work with the new size? But that still doesn't explain, why the old model tires size doesn't change.

    And the next problem is generating the lookup table. The way I tried this is choosing the automated test in the ttool and start it. Afterward it starts the first test with which takes about 5s. The same for the next 2 tests. But then there comes one test that counts up to abou 5% done and then counts back to 0.005% and stay there. When I increase the Test index in this state it starts a new test. That can behave just like either of both cases I just mentioned.

    Does anyone know what I'm doin wrong?

    Thanks and greetings
     
  18. Kev

    Kev Registered

    Joined:
    Oct 28, 2010
    Messages:
    156
    Likes Received:
    0
    If you change geometry of the tire you must recalculate all loopup tables for it. There is, however, simple size multiplier somewhere to use small amounts of scaling without recalculation.

    If some tests getting stuck you should try to play with DT mult. I believe you should lower it if your tire explode during test and raise it if test doesn't progressing. Skipping tests is not the right way.
    By the way, was you patient enough? Some tests can be pretty long, 10 to 20 minutes on my CPU.
     
  19. Rocksor

    Rocksor Registered

    Joined:
    Oct 20, 2010
    Messages:
    99
    Likes Received:
    1
    Thank you for your reply. Then I'll have to play with DTmult. I'm pretty sure I have been patient enough. Because for 0.005% 5h is enough wait I think...:p
    I already lowered the DT, but now I tried lowering it even more and with 0.9 it seems to work. I will let that run, when I go to bed.

    By the way: What does the DT multiplier do altogether?
     
  20. Kev

    Kev Registered

    Joined:
    Oct 28, 2010
    Messages:
    156
    Likes Received:
    0
    Delta-T, i.e. timestep for calculation. Lower DT - higher precision each itteration but more itterations needed. So it's kind of trade-off between stability of test and time this test taking.
    I believe when tire explode means that there is divergence through itterations at some nodes and tire became unstable. This can happen when timestep is higher than period of natural frequency of any node. So lowering multiplier prevents this.
    Sorry for so much science. :)
     

Share This Page