Porsche 911 RSR Loss of downforce

Discussion in 'Bug Reports' started by Matheus Machado, Nov 12, 2019.

  1. Matheus Machado

    Matheus Machado Registered

    Joined:
    May 4, 2017
    Messages:
    62
    Likes Received:
    30
    At the rF24 I had a very weird bug regarding downforce with the Porsche.

    When I got in the car for my night stint I noticed the car was very oversteery compared to practice and the stint I had done earlier in the race before the server restart. As track conditions were very tricky I blamed the behaviour on that as well as possibly being tired at first. But when rain started to fall, it was very similar to the first period of rain, and the car was completely different. The car had no damage and the setup was the same as before. After I got out of the car I went to double check ride height values to see if the setup was indeed the same and it was very similar when the car had little to no aero load. So it was an indication that it was the same setup. Although the car was producing less downforce overall, especially at the rear, which would explain the car behaviour which was more noticeable at high speed corners, such as the porsche corvette complex.

    That is the long story short. Now I will show the actual values and explain the calculation methods I used to get downforce values and signs of oversteer.

    So first, I calculate downforce using suspension force (as called in motec), subtracted by the static force when the car is stopped, with weight transfer considered for the average longitudinal acceleration above 200 kph when on full throttle. I do that for front and rear as well as a combined channel. So this not using the previously available downforce channels. Aero balance is simply the amount of front downforce compared to the total downforce produced by the car.

    To check if it was a mechanical grip issue, I checked the amount load transfer % that was happening in the front of the car. Load transfer is calculated simply by subtracting FL - FR and RL - RR or vice versa.
    upload_2019-11-12_19-3-41.png
    These are pitlane ride height values comparing the affected car and normal car.

    Average load transfer at the front increased from 0,55 to 0,56 on the affected car, not being compatible with the oversteery behaviour. Even with tire pressures at the front increased by 30 Kpa cold and load transfer at 0,57 the car continued to be oversteery, and the data shown below is already with the increased front tire pressure.

    With that in mind, here are my findings.

    First this is the aero data comparing a lap under similar conditions in the first stint and in the second where the issue was present (colored).
    upload_2019-11-12_18-53-12.png
    upload_2019-11-12_18-54-34.png
    I think it is quite clear that the car lost as much as 1/3 of rear downforce looking at this data. There was also downforce loss at the front, although, less significant as the rear was producing more downforce as a starting point. Notice as well how overall downforce decreases.

    upload_2019-11-12_19-0-0.png
    Interestingly the rake was more positive which should have the exact opposite effect on the aero balance, where the car would at the very least produce more front downforce, and overall being able to produce more downforce. Aero balance shifted to the the front as expected and shown before.


    As the car had a crash during the night, I thought it might have been some issue with the repair of the rear wing, so after the race I removed the rear wing of the car (crashed in reverse at the first chicane) and got these results.
    upload_2019-11-12_18-44-48.png
    This is comparing a car with no rear wing (grey) and the car with loss of downforce (colored).
    It shows that the rear was even worse without rear wing, so my initial theory of the rear wing not producing downforce anymore was not valid. I offset the graph on purpose to match speeds for comparison reasons.

    upload_2019-11-12_19-12-20.png
    After repairs the car regained downforce as expected, and showing more downforce than the affected car (colored).

    upload_2019-11-12_19-13-29.png
    Now the colored data shows the car from the stint it was normal, compared to the repaired car which I removed the rear wing. Results are within expected margin of error.

    upload_2019-11-12_19-16-4.png
    This is comparing a normal car (colored) to one which I had scratched the rear bodywork. There is a small decrease, although not as drastic as the bug being reported. It is interesting to note that no damage was showing in the hud in this case, so it is not something you can repair, unless you crash the car more severely.

    It is also interesting to note that the understeer angle of the car decreased by 2.1 degrees. Another indicator of the oversteer.
    The understeer angle is a math channel that basically compares how much steering input you have compared to the calculated required angle under ackermann cornering. So it is a objective measure that can be used to quantify understeer.


    I plan on getting my teammates motec logging, although I think they dont have the suspension channels enabled which would leave their analysis much more to the subjective side, although they also had the same decrease in laptimes I had. Which is an indicator they also had the issue.

    If any additional data is required, contact me.
     
    Last edited: Nov 13, 2019
  2. Daniele Vidimari

    Daniele Vidimari Registered

    Joined:
    Apr 20, 2016
    Messages:
    659
    Likes Received:
    593
    Great post.
     
  3. D Nichols

    D Nichols Registered

    Joined:
    Jun 18, 2019
    Messages:
    32
    Likes Received:
    16
    We had similar issues at the same corner. I will see if the driver involved was running MOTEC at the time.
     
  4. CrimsonEminence

    CrimsonEminence Registered

    Joined:
    Jun 18, 2019
    Messages:
    442
    Likes Received:
    439
    I always thought, the Porsche RSR is way to sensitive at the rear, after 2-3 laps at Le Mans, also in situations, where rake should work against some bad effects...thankfully you logged numbers now!

    It would be great, if S397 could take a look over all their GTE, because i have the feeling, the M8 also has this problem, but i have not logged anything, unfortunately...
     
  5. Matheus Machado

    Matheus Machado Registered

    Joined:
    May 4, 2017
    Messages:
    62
    Likes Received:
    30
    Indeed the car tends to become more oversteery as the stint goes on, although when the downforce was normal I was able to run minimum rake and downforce with still reasonably set springs. It required mostly damper tweaks and aggressive toe.

    Tyre wear which under normal conditions would be 2x higher at the front was equal for most of the stint and at the end the rear was getting around 2% more wear than the front. So that didn’t help as well.
     
    CrimsonEminence likes this.
  6. Matheus Machado

    Matheus Machado Registered

    Joined:
    May 4, 2017
    Messages:
    62
    Likes Received:
    30
    I would gladly take a look at those logs.
     
  7. Matheus Machado

    Matheus Machado Registered

    Joined:
    May 4, 2017
    Messages:
    62
    Likes Received:
    30
    And regarding rake, I got myself mixed up with the images at the time, it indeed increased with more speed, as expected for a car with rear downforce loss.
     
  8. CrimsonEminence

    CrimsonEminence Registered

    Joined:
    Jun 18, 2019
    Messages:
    442
    Likes Received:
    439
    This helps, indeed! I use VERY conservative settings for dampers, hard front, soft rear in general, to generate more stability, but it has almost no influence on the "general" behaviour in high speed corners. But at least, i don't lose it that fast...Laptimes have nothing to do with reality though...weakest car in the GTE grid in rF2, which is ridiculous.
     
  9. Matheus Machado

    Matheus Machado Registered

    Joined:
    May 4, 2017
    Messages:
    62
    Likes Received:
    30
    I think that is beyond the point of this thread, but saying the Porsche is the weakest car is stretching it imo.
     
  10. Matheus Machado

    Matheus Machado Registered

    Joined:
    May 4, 2017
    Messages:
    62
    Likes Received:
    30
    And comparing to real life is imo, useless, as the tire behavior and track conditions aren’t simulated close to what happens irl, so it is pointless to just mirror real life for bop purposes.
     
  11. CrimsonEminence

    CrimsonEminence Registered

    Joined:
    Jun 18, 2019
    Messages:
    442
    Likes Received:
    439
    Stretching what?
    It is, test it by yourself...you can't go the M8s pace at all...may it work on long run? Yes! May it be realistic...no.
    It is pointless to simulate a GTE package, if there is no simulation close to real life. It should definetly equal some things in real life, or we can call it GT-Sport (i don't have anything against GTS, as a disclaimer :D ). I personally don't need balancing for fancy eSportz purposes. Too much praise for rF2, all the time, if this is the way, we should look at it.

    Just my opinion to that...maybe it is unpopular, i can understand that, but for me, rF2 is about realism, even if i'm aware about some flaws within the simulation of conditions at race weekends :)
     
    Last edited: Nov 14, 2019
  12. Matheus Machado

    Matheus Machado Registered

    Joined:
    May 4, 2017
    Messages:
    62
    Likes Received:
    30
    Again, not the point of this thread.
     
    VirusGR42 likes this.
  13. CrimsonEminence

    CrimsonEminence Registered

    Joined:
    Jun 18, 2019
    Messages:
    442
    Likes Received:
    439
    Yes right.

    Realistic behaviour of rear downforce of the RSR it is.^^ And it doesn't work right...therefore it doesn't perform, as expected.
    Therefore it doesn't perform. Period.
     
  14. D Nichols

    D Nichols Registered

    Joined:
    Jun 18, 2019
    Messages:
    32
    Likes Received:
    16
    I have asked the driver if he was running MOTEC. I also have my own log somewhere from a practise session, where I was just testing the default setup. I will see if I can remember which one it is, and I will send it also. Would you be willing to share the maths you used to get those results?

    Spitballing this topic within our team, we wonder if this is related to the diffuser stall? As we know, if you get too close to the car in front, then it can lead to that cars diffuser stalling. This has already been raised here: https://forum.studio-397.com/index.php?threads/bmw-m8-gte-downforce-problems.64055/#post-1006005 (although its a real life effect, its not a bug as such).

    Our question/concern is on the implemenation of this stalling. Is it happening when it shouldn't? In other words the game is simulating diffuser stall when the conditions for it to happen have not been met. That could explain why the game is suddenly simulating a loss of downforce. Any diffuser stall would have a small impact on the splitter, as the overall underbody air speed is reduced. Although it would be clearly concentrated on the rear; which your graphs show.

    Another issue some of our drivers have noticed is the setup of car changing between race sessions and/or driver swaps. Both in the VEC Spang race, and at this Le Mans race, two drivers noticed this problem. Firstly at Sepang when Brandon took over from me, he said the car just did not behave in the way it had done just minutes before when he was putting in some laps to prepare. Then at Le Mans, Marius said the cars behavior changed between the sessions (practise, qual and race etc). Both are experienced enough to be able to tell the difference in real road changes, and setup changes. Both said the setup had changed. This was also reported by another driver on the S397 discord after the race, but I forgot that drivers name. Has your setup altered at the point you took over the car, leading to it not handling the same way?
     
    CrimsonEminence likes this.
  15. Matheus Machado

    Matheus Machado Registered

    Joined:
    May 4, 2017
    Messages:
    62
    Likes Received:
    30
    At first I asked the driver who "rejoined" the race if he had loaded the correct setup. He sid he did and data can confirm it, as you can see with pitlane ride height measurements and top speed being very similar.

    If you look at front and rear ride heights (on a high speed straight) you can see this. If the rear is higher than it should be, while producing the same top speed it is indeed the same issue happening. Just be sure to check if ride heights at pit limiter are the same as a "normal" car.
     
    D Nichols likes this.
  16. Stefan_L_01

    Stefan_L_01 Registered

    Joined:
    Nov 18, 2012
    Messages:
    595
    Likes Received:
    386
    The RSR has a 3rd sping which will affect longitudinal weight transfer, no?
     
  17. Matheus Machado

    Matheus Machado Registered

    Joined:
    May 4, 2017
    Messages:
    62
    Likes Received:
    30
    It doesn’t have 3rd springs. If it did, it wouldn’t affect it.
    It would change the spring stiffness calculation, but as the sim already gives the load of the suspension mount it is already taken into account.
     
  18. Stefan_L_01

    Stefan_L_01 Registered

    Joined:
    Nov 18, 2012
    Messages:
    595
    Likes Received:
    386
    Imo rf2project plugin showed 3rd spring. Motec plugin does not show 3rd at all so you dont know.
    Anyway your back calculation of aero is nice yet complex and based on some corrections which may may include some unknown variables.
    If I understand you correctly, you have recorded during the race motec data and now you compare it to offline data of same setup, same fuel load, same tire temp / pressure, same ambient temp?
    So please show pure Suspension forces on straights right before braking at same speed. The difference should be same, no complex correction needed. And of same lap as engine torque may may increase with heating up. And same mixture of course.
     
  19. Matheus Machado

    Matheus Machado Registered

    Joined:
    May 4, 2017
    Messages:
    62
    Likes Received:
    30
    The intention of this downforce math channel was never to get scientific article levels of approximation of the downforce value. I use it to compare aero balance between different setups. So all the calculations were done with this in mid, not to consider that many variables which are way below the scale of the value produced, and most of them are not even there in rf2.

    Regarding 3rd spring, the 911 rsr definitely does not have one and the data acquisition interface does provide the 3rd spring position in cars it is available.

    And as you can see in the post, I have compared to an earlier stint which the car was normal, as well as offline data with different levels of damage.

    All the parameters you mentioned were similar and even if not, the difference is way above anything even the worst case of external factors could produce.
     
  20. Stefan_L_01

    Stefan_L_01 Registered

    Joined:
    Nov 18, 2012
    Messages:
    595
    Likes Received:
    386
    There are 2 variables in the Plugin interface called mFront3rdDeflection, mRear3rdDeflection ( no Forces) and they show changing values for the RSR so I assume it has a 3rd spring.
    Anyway, interesting find. At slow Speed they seem to run together (no Forces shown in pit lane however), so yes I get your Point About aero.
    I´d like to compare with telemetry Output from myself (participated too that race but breakpoints and grip of others, huh...) but I dont have your set of Course...and did not record during race myself. Still it would interest me to compare it by an offline stint
     

Share This Page