1. Trouble with the game?
    Try the troubleshooter!

    Dismiss Notice
  2. Issues with the game?
    Check the Known Issues list before reporting!

    Dismiss Notice
  3. Before reporting issues or bugs, please check the up-to-date Bug Reporting Thread for the current version.
    0.36 Bug Reporting thread
    Solutions and more information may already be available.

Oil starvation goofiness

Discussion in 'Troubleshooting: Bugs, Questions and Support' started by NGAP NSO Shotgun Chuck, Apr 12, 2023.

  1. NGAP NSO Shotgun Chuck

    NGAP NSO Shotgun Chuck
    Expand Collapse

    Joined:
    Mar 6, 2015
    Messages:
    1,446
    My vanilla customized Moonhawk that I use as a default has issues with random oil starvation, especially on undulating roads. However, I also discovered by using the engine thermals debug app that sometimes it will oil starve quite significantly, enough to start racking up a tiny amount of damage, just by flooring the throttle from a stop. (It has a 2.64:1 axle ratio and rides the clutch up to about 25MPH; dumping the clutch and spinning through first gear seems to prevent this). The odd part is, it seems to be completely random when this happens, and it seems to be coming from oil starvation XY (which I assume is excessive lateral G-force) rather than oil starvation Z. As I mentioned to start, it will also throw an "engine is starved of oil" message in canyon runs, though this also seems to be unpredictable; on one hand, I did manage to trigger a momentary hiccup by getting light over a crest, but on the other, much more severe lightness and even a couple of jumps didn't faze it previously, and previous oil-starvation incidents may not have matched particularly well with topography.

    It's on Radial T/As with a race oil pan, there's no way it should ever have trouble with excess lateral force.

    Climbing hills seems to cause problems with the game's oiling system in general. A Bolide 320 GTT, when climbing a hill, can enter a state where even a slight squirm side-to-side will trigger oil starvation, though not enough to actually damage the engine; this is difficult to replicate. More significantly, on a previous occasion when I was messing around with the Shutoko mod map, I found that an all-vanilla Wendover with a maxed-out 4.4L V8 will very repeatably oil starve, enough to cause damage, when traveling uphill at speeds over 220 MPH; backing off to below 220 would prevent this. It only seemed to happen in one specific tunnel on outbound Wangan and not on any other uphill section, so I assume a very specific climb angle triggered it.
     
  2. Agent_Y

    Agent_Y
    Expand Collapse
    Jbeam/QA support
    BeamNG Team

    Joined:
    Jul 10, 2020
    Messages:
    10,434
    Yeah I have similar issues with my Moonhawk, hasn't stopped when they remade the engines. Sometimes it even somehow starves of oil from standing still, but that's very rare. Several other cars I have had similar issues with, but I don't remember which ones in particular. Idk if the Moonhawk has the worst issues or if it's just the fact that I use it the most.
     
  3. NGAP NSO Shotgun Chuck

    NGAP NSO Shotgun Chuck
    Expand Collapse

    Joined:
    Mar 6, 2015
    Messages:
    1,446
    Other cars have the issues but so far the Moonhawk definitely seems to be the worst and I wonder if it's my incompetent floaty suspension tuning that's at least partially to blame. Though to be honest, the Moonhawk has a ton of weirdness anyway. The 309 that absolutely refuses to rev or produce power in any form even though the smallest engine should theoretically rev the highest when tuned up, the 448's weird second power peak above 6000 with every upgraded long block, the Special still having its rev limiter set super low despite being able to safely reach that second peak (besides that it's unironically a good car now), and so on.
     
  4. Agent_Y

    Agent_Y
    Expand Collapse
    Jbeam/QA support
    BeamNG Team

    Joined:
    Jul 10, 2020
    Messages:
    10,434
    Yeah I feel like the engine rework was super rushed and lots of things were overlooked with the new engines
     
  5. Diamondback

    Diamondback
    Expand Collapse
    Vehicle Systems Lead
    BeamNG Team

    Joined:
    Apr 8, 2014
    Messages:
    1,966
    What you report sounds like a tiny amount of jbeam instability.
     
    • Agree Agree x 3
  6. NGAP NSO Shotgun Chuck

    NGAP NSO Shotgun Chuck
    Expand Collapse

    Joined:
    Mar 6, 2015
    Messages:
    1,446
    Was what I figured but I wasn't quite sure. I know cars will have a tiny, imperceptible amount of vibration even at rest.

    For what it's worth, the remastered VM301 Pessima seems to be rock solid in this regard. Even with deliberately abusive J-turns, I can't get it to oil starve in any significant way.

    The Wendover on the other had has an extremely bad case of this for which the main trigger seems to be wheelspin, especially in first gear - seems like the "floatier" the front end gets and the more you change steering angle after the wheels start spinning, the more likely it is to oil starve. (A half-J-turn in certain sports variants will very reliably put the front end on ice and kill oil pressure.) Letting off the gas will stop the oil starvation, unless it didn't happen while the wheels were spinning, in which case letting off might start it instead. A more powerful engine also seems to make it slightly more likely, to the point where I wondered if this wasn't a part of the reason for no V8/manual Wendover in the official config list.
     
  7. default0.0player

    default0.0player
    Expand Collapse

    Joined:
    Nov 30, 2018
    Messages:
    2,057
    If the jbeam structure clunk against each other or have some instabilities, the engine might vibrate(excessive G) in a microscopic amount that can only be observed when using 100x slowmo. How about adding a low-pass filter in the oil starvation logic?
     
    • Agree Agree x 1
  8. Diamondback

    Diamondback
    Expand Collapse
    Vehicle Systems Lead
    BeamNG Team

    Joined:
    Apr 8, 2014
    Messages:
    1,966
    The data is already heavily smoothed (which is a low-pass filter), but at some point the only fix is to make the jbeam stable.
     
  9. Agent_Y

    Agent_Y
    Expand Collapse
    Jbeam/QA support
    BeamNG Team

    Joined:
    Jul 10, 2020
    Messages:
    10,434
    I always assumed the Moonhawk engine had a stable Jbeam but apparently it doesn't, here's the cause of our issue
    sdffaw.png
    This happens for every single engine, for the V8s more than for the I6

    Meanwhile the Wendover is stable when static or driving straight, but as soon as you start cornering or accelerating hard it becomes unhappy.
     
    #9 Agent_Y, Apr 13, 2023
    Last edited: Apr 13, 2023
  10. Szymon2007

    Szymon2007
    Expand Collapse

    Joined:
    Nov 30, 2017
    Messages:
    747
    i think the entirety of the moonhawk is a tad bit unstable
     
    • Agree Agree x 1
  11. Agent_Y

    Agent_Y
    Expand Collapse
    Jbeam/QA support
    BeamNG Team

    Joined:
    Jul 10, 2020
    Messages:
    10,434
    Yeah that's the force transmitted from the engine through the transmission, transmission mount, frame and body to the fenders it seems. This car needs a full Jbeam rework.
     
    • Agree Agree x 2
  12. NGAP NSO Shotgun Chuck

    NGAP NSO Shotgun Chuck
    Expand Collapse

    Joined:
    Mar 6, 2015
    Messages:
    1,446
    Looks like American Road? Some jittering could be due to distance-from-zero, which tends to happen with megamaps in Beam.
     
  13. Agent_Y

    Agent_Y
    Expand Collapse
    Jbeam/QA support
    BeamNG Team

    Joined:
    Jul 10, 2020
    Messages:
    10,434
    I thought that got fixed in 0.28
     
  14. NGAP NSO Shotgun Chuck

    NGAP NSO Shotgun Chuck
    Expand Collapse

    Joined:
    Mar 6, 2015
    Messages:
    1,446
    I haven't tried it yet; when I first updated I had to deactivate all my mods to get anything to load. But somehow I doubt it; it is/was a consequence of floating point imprecision
     
  15. Agent_Y

    Agent_Y
    Expand Collapse
    Jbeam/QA support
    BeamNG Team

    Joined:
    Jul 10, 2020
    Messages:
    10,434
    As far as I understand they fixed it by calculating position from the vehicle side, not the map. So the map is now affected by this inaccuracy rather than the vehicle, and a lot less can go wrong since it's nowhere near as complex.
     
  16. NGAP NSO Shotgun Chuck

    NGAP NSO Shotgun Chuck
    Expand Collapse

    Joined:
    Mar 6, 2015
    Messages:
    1,446
    How does that work itself out? Do the positions of objects and decalroads change with every load? Do they jitter around? Does the map itself jitter around under you?
     
  17. Agent_Y

    Agent_Y
    Expand Collapse
    Jbeam/QA support
    BeamNG Team

    Joined:
    Jul 10, 2020
    Messages:
    10,434
    I have no idea, I'm just saying what I've seen other people say. I haven't tried any large maps yet either, will try today.
     
  18. Szymon2007

    Szymon2007
    Expand Collapse

    Joined:
    Nov 30, 2017
    Messages:
    747
    It's on the Pikes Peak map
     
  19. Bauer33333

    Bauer33333
    Expand Collapse

    Joined:
    Jul 25, 2020
    Messages:
    267
    I don't know how it is now, but before 0.28 they did. But the position changes where usually too small to be noticed, only thing they broke where train tracks on huge maps like Roane Country since those need to be lined up perfectly for trains not to derail
     
  20. Diamondback

    Diamondback
    Expand Collapse
    Vehicle Systems Lead
    BeamNG Team

    Joined:
    Apr 8, 2014
    Messages:
    1,966
    The floating point accuracy issue never affected the physics (for reasonable distances) since those always used 64bit. It was only the visuals that vibrated and yea, that's should be fixed now (I'm surprised I read so little about that, given the amount of complaints we got about that before)
     
    • Like Like x 2
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice