EuskoTestTrack

Discussion in 'Locations' started by SPASKIS, Nov 26, 2017.

  1. Skan

    Skan Registered

    Joined:
    Dec 14, 2014
    Messages:
    333
    Likes Received:
    135
    Now I just need a 3D printer big enough to output 4 lanes at a time. Imagine simulating rF2 in real life instead of the other way around.
     
    Rastas likes this.
  2. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    It is just a teaser of what can be done with the track design tool. I want to test it to see what's possible and what is not in rF2. It will help me decide what type of elements I may use in the final version. The weird behaviour when the car is in vertical position worries me with respect to the twisted loops after which the F1 should be able to drive backwards
     
  3. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    Unfortunately physics don't work OK when the car is in vertical position.
    Besides due to how BTB imports XYZ data precisely when track is vertical its difficult to get a smooth surface there.
     
  4. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    I have been using Formula ISI for testing in order to have a reliably adjusted vehicle.

    As you say the way aero forces are programmed are probably not using the correct local reference and this is causing incorrect result when the car is not horizontally placed. I have no idea what those coefficients mean.

    I did a more controlled test two days ago and the result was the same. When the car starts pointing up aero forces start doing weird things. What is curious is that it seems to have a hard value upon which things get wrong. The initial buildup for the normal reaction seems correct but from a given point it stops increasing as my worksheet suggests and drops to near zero values. In the same way some meters after the car surpasses vertical position, normal force restores for the given velocity. However the speed loss in the around vertical position region is too big and the car does not get enough speed to be able to drive downwards in a flat surface.

    I am very reluctant that these problems can be solved by adjusting values in the hdv. It seems to be a core problem linked to the fact that something has been simplified in the physics considering that the car drives near the horizontal. I wonder which side effects could have this physics bugs in heavily banked tracks.

    The best way to avoid side effects is to properly simulate things. Furthermore, considering that this issue only implies a correction of the reference system for aero forces I would suggest to be looked at and corrected. You might discover that the benefit not only applies to Spaskis crazy tracks. @Michael Borda ?
     
  5. Raintyre

    Raintyre Registered

    Joined:
    Jul 28, 2013
    Messages:
    226
    Likes Received:
    187
    Ok, but are you sure that the problem is not caused by some issue on the track surface construction?
    Could you build a track with a long vertical straight, so that we can check whether the vehicle is capable to recover natural tyre loads after some meters going on pure vertical straight?
    We just need enough downforce and also a horizontal straight to accelerate and earn enough speed before the vertical straight. The vehicle should be able to stick on the vertical road for a long time so that we can see what really happens.
    Of course the only source of load should be aerodynamic downforce.
     
  6. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    I am 99% sure that the track construction does not make any difference. Surface irregularities would create an oscillation around the calculated mean values.

    In this track, the surface characteristics are the same throughout the whole track. There is no difference in construction between any part of the loop.
    In it, the upward loop should reach near 10g normal acceleration values (98 m/s2). However in rF2 normal this value drops dramatically after having reached around 4-5g.

    As in other cases, the negative acceleration in the mid of the loop (vertical) is around 2,5 g. 1 g is due to gravity but where are the 1,5 times gravity difference coming from? For sure not from drag which does not reach those values at the speed that the car carries when the problems arise. It is clearly coming from some bad calculation of either force magnitude or its direction which ends up pulling down the car heavily.

    Screenshot_20171227-021549.png
     
    Last edited: Dec 27, 2017
  7. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    The tube is funny.


    For those who want to try. EuskoTestTrack_v1.10
    Everything is WIP since I am just doint trials for the moment.
     
  8. Raintyre

    Raintyre Registered

    Joined:
    Jul 28, 2013
    Messages:
    226
    Likes Received:
    187
    What about tyre loads when the car is going down the loop?
    I mean, it is true there is a sudden loss of tyre load when the vehicle climbs the loop. I tried with other vehicle and it also happens to me. But, do you obtain also any loss of tyre loads when the car is vertical and going down?
     
  9. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    My feeling is that the problem is the same. The only difference between both cases is gravity. When going upwards gravity wants to slow the car and makes the "no tire load" effect much more noticeable but you can still feel it when going downwards.

    It was late when I finished and tested the track so I didn't check telemetry. I just drove and made the video so that you could check the result.

    I am now enjoying in the snow out of home so I cannot drive properly. I didn't bring my steering wheel with me. I have my computer though. I'll check the telemetry this afternoon.

    In any case it is clear that rF2 physics fail to work as in real life when the car is not in near horizontal position.

    There shouldn't be a limitation for this considering that the problem is apparently regarding the reference system being used for aero load calculations. In my spreadsheet it doesn't matter any of these things and so could be done in rF2.

    Note that the car with lowest aero (MCL 650) was able to make a double loop in the tube. Something that I dont think the Formula ISI would be able to do with standard aero settings.

    I already mentioned @Michael Borda to see if he maybe could explain the situation. I am very unlucky in receiving any official response to any physics related issue I have pointed in the past.
     
  10. dadaboomda

    dadaboomda Registered

    Joined:
    May 15, 2016
    Messages:
    712
    Likes Received:
    202
    I have try the 1.1...... Very fun. With my Boxmaster fully loaded torque +X7 and power +X12.... awesome :p

    Thank's Spaskis.
     
    SPASKIS likes this.
  11. Raintyre

    Raintyre Registered

    Joined:
    Jul 28, 2013
    Messages:
    226
    Likes Received:
    187
    I wouldn't say it is exactly a RF2 engine bug yet. Probably the issue is caused by the vehicle's physics, which perhaps are not ready to cope with 10G's of normal force.
    Have you checked ride heights telemetry at the loop? Perhaps ride heights are reaching negative values, due to strong force against the road, and that can drive to unexpected effects on aerodynamics, suspension and perhaps also on tyres. Remember that RF2 tyres deform under load. Try with harder suspensions until ride heights are always positive.
    I think you should take in account tyre roll resistance for your drag calculations, since tyre friction could be respectable when you push the tyres to those extreme loads.
     
  12. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    The problem was also present when the radius was negative and hence normal force would be caused mainly by aero. In that case the argument of negative ride heights fails.

    This is the track I refer
    https://forum.studio-397.com/index.php?threads/euskotesttrack.58457/page-3#post-924772
     
  13. Michael Borda

    Michael Borda Car Team

    Joined:
    Oct 21, 2011
    Messages:
    187
    Likes Received:
    143
    The hat collision detection begins to fail around 60 degrees of angle and continues until about 120 deg. Check the TGM bristle display with CTRL+= and you will see that the bristles no longer contact the ground properly. So when that happens it starts to fall on the undertray, which drops the load on the tyres. I don't really know what it would entail to fix this, TBH, but I don't think the fix is trivial...
     
    Last edited: Jan 3, 2018
    Raintyre, Pauli Partanen and SPASKIS like this.
  14. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    Thank you very much for the reply Michael. I'll check what you suggest. Does Ctrl+H work in single player or should I do it in dev mode?
     
  15. Michael Borda

    Michael Borda Car Team

    Joined:
    Oct 21, 2011
    Messages:
    187
    Likes Received:
    143
    CTRL+H is a dev mode only thing. For the record, I did manage to run the FISI (physics untouched) upside down on your track, with a tweaked setup. It's difficult because it scrubs a lot of speed near that 90º angle, I think the bristles may even sometimes apply forces in the wrong direction when approaching 90°. Finally, I just looked into the last issue about tyre spring rates / linearity and stuff you mentioned before. The problem appears to be that reported tyre deflections in the telemetry are actually still based on some .TBC calculations, and the real rates with the .TGM are not as linear as the telemetry would suggest. Furthermore, the issues with 'ride height' are caused by the way our garage setups are done, rather than a model issue as such. Setup is still applied to the .TBC model, so if there is a discrepancy between the .tbc and .tgm (which happens more with extreme setup ranges), it ends up giving you an "incorrect ride height". Setups are also applied after tyre adjustments. So if you lower pressures, the ride height is automatically extended to compensate. We do plan to look at various things in the future and make them more intuitive.
     
    Mulero, McFlex, SPASKIS and 1 other person like this.
  16. Lazza

    Lazza Registered

    Joined:
    Oct 5, 2010
    Messages:
    12,346
    Likes Received:
    6,572
    :D:D:D
     
  17. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    Thanks again. Very much appreciated. Knowing the limitations I will design the track according to them. I will maintain doable loops. I really enjoyed the tube. I am willing to test it with my new hardware when I get back home.
     
    Pauli Partanen likes this.
  18. dadaboomda

    dadaboomda Registered

    Joined:
    May 15, 2016
    Messages:
    712
    Likes Received:
    202
    For information:

    I have huge bugs that require me to do ctrl + alt + del in the track ett:

    for example, with the boxmaster fully loaded + 20 × more powerful and torquey than the original version, rf2 bug when I do the little loop (> 250 km/h :p).

    Frankly I have the impression that it is a physics bug rather than a track bug.
     
  19. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,426
    Remember that that car is not official and hasn't been sufficiently tested. The physics are the same for the cars where those issues don't appear.

    It might be more interesting to use official cars and if you find problems there report them to their authors which might find your findings more useful.

    The extreme loads and circumstances of this test track might be a good to spot undesired issues regarding hat and regular collision targets.
     
  20. dadaboomda

    dadaboomda Registered

    Joined:
    May 15, 2016
    Messages:
    712
    Likes Received:
    202
    I do not try to test the limits of the game. I have reached them despite myself. It's not my job and I'm not part of the S397 team.

    I actually reached the limits with the boxmaster (ISI van), official car, on which I made only extremely minor changes.

    About 30 laps of Nordschliefe performed (600 km), even with a large number of big shocks with and without AI, I had no form of bug.

    Your track is extreme, it shows problems that there wouldn't be on other circuit.

    The changes made :

    Larger brake disc, higher braking torque, strongly much higher brake cooling, torque and power greatly increased by COEFFICIENT: the physics of the engine was not modified. Significant increase in engine braking. Increase in clutch torque and bulk torque. Very strong increase in the final gear ratio.

    I think that's about it all :

    The physics of the boxmaster remains the physics of the boxmaster.

    Then the most important :

    Upgrade to the boxmaster FULLY LOADED. Setup: maximum braking power, and very important, 50/50 braking distribution. Finally, maximum body height (24.5 cm / 25.5 cm) for fun and entertainment.

    With this configuration, whose physics Boxmaster have almost not been changed and superficial, speed reached greater than 250 km / h between the arrival and the first looping : arrived at third (or more I didn't remind) of the looping, the image crashes completely, the game must be stopped via CTRL + ALT + DEL.

    I do not think you're interested, but I can put you the modified boxmaster on Dropbox. By exactively deactivating tire wear and using the fully loaded upgrade associated with the setup I propose, it's absolutely great. Incredibly genial. I have not yet changed the quad, but it's still a great ISI car that unfortunately is not used by anyone.
     

Share This Page