https://www.beamng.com/threads/rollover-physics.71590/ This has been discussed before but I don't feel like it was properly adressed, the discussion trailed off pretty quickly to modifying suspensions and how most real life rollovers happen due to collisions or going off the road. The OP was and I am interested in stock vehicles going on 2 wheels or rolling over with no external factors, just extreme manouvers on flat asphalt. A thing that comes up a lot with cars rolling is the moose test so some important things need to be noted about it first: (Here is the wiki page for the moose test: https://en.wikipedia.org/wiki/Moose_test) 1. The moose tests are officially done with the car fully loaded - that means every seat is filled by a person and the trunk is stuffed with cargo for a worst case scenario. There are specifics and other things here that I am too lazy to look up and write down 2. Both inner tires leaving the ground is way beyond just failing the moose test, most cars do not even lift just their front wheel. 3. This scenario is happening to new cars and small cars, it's not limited to archaic suspensions or massive SUVs There's now a small moral debate here about point 2: should BeamNG.Drive represent normal cars that are roadworthy and actually safe or suboptimal cars that can behave dangerously? My 2 cents is that the specific thing this game dragged me in with 9 years ago was the promise of crummy cars. To be even more specific heres that one quote from the super old vehicle handling video: "BeamNG.Drive isn't all about race cars and race tracks, it features production vehicles on everyday roads which are often in poor condition. Some cars can behave dangerously when driven to their limits." I don't think I would be here or even like cars in the slightest if I hadn't watched that video. It doesn't feel like that promise has arrived. Now let's do a test. Take a roamer, put a 200kg box on its roof, and empty the fuel tank to raise that CoG as high as you can. This feels like it'll be more than what the moose test does in terms of loading. But if you do this you'll notice the roamer still struggles to even lift its front wheel, let alone the rear one. I think you need to put the weight somewhere at 400kg before it'll actually flip and that's way beyond whatever the moose test does. Now do this on the cherrier tograc and you'll see similar results. I hope it comes across that I'm no scientist, everything here is based off the feels(TM) and the extent of what I've actually proven is that the gavril roamer is a very good SUV. I haven't actually figured out if a car with 200kg on the roof not flipping is unrealistic cause BeamNG doesn't simulate people sitting in the car, the CoG might still be in a good spot. I can only say is that it feels wrong, and it does feel wrong. Thank you for coming to my TED talk, this is why I think the cars in BeamNG should be easier to flip
200kg "feels" like more than what the moose test does? You stated yourself every seat is occupied as well as having cargo. So what do these people weigh, 30kg? "The car to be tested has one belted person in each available seat and weights in the boot to achieve maximum load." This would imply that they load vehicles to their GVWR which can be anywhere between 15-40% of their curb weight. This is their payload capacity which is generally more than "400kg" even for sedans. 400kg to flip the roamer tracks.
The important thing here is, that's 200 on the roof! This changes the CoM more than passengers and cargo whose masses are gonna be below the windowline. The second important part was removing most of the fuel, even though this means less weight the CoM goes up making the car more prone to tipping over. But I get that you don't vibe with the napkin math. I did new testing based off vehicle weight instead: Hopper ZXT-6 (M) dry weight is 1570kg with fuel 1604kg diff 34kg GVWR (estimated, 40%) is 628kg GVWR minus diff = car should be loaded with 594kg on top of a full tank to reach theoretical GVWR Oil filled barrels weighing in at 195kg is a pretty good way to reach this, I put 3 in the car (needed to tape the 3rd one into the trunk) and tested on gridsmall pure. Testing was done by accelerating to 60kmh and then doing a slalom at different intensities, alternating doing said slalom with throttle and without throttle so it could be tested with weight shifted to different places This was repeated with equivalent weight calcs for the roamer (v8 4WD XT Auto) and also the 90s pessima (default fugly red one) they did not tip over or lift any wheel off the ground in these tests I will say again: these cars not being a meanace and operating safely at their (unofficially calculated) maximum loads is not unrealistic, this entire thread amounts to me wanting the cars to be more dangerous which is not unrealistic either, there are real cars that fail the moose test and slaloms like this. I think making these virtual cars more dangerous aligns with what BeamNG.Drive aims to be, I base that off what they've previously said If and how the BeamNG team achieves this is up to them. The least they can do to further this cause is to make driver and passenger weight simulation not be exclusive to the 3-wheelers, seriously what is up with that. The closest analog is cargo boxes in the seats which max out at 25 kilos, it's that or manually dragging oil barrels into the seats and strapping them in like a very, very american baby which was tedious (PS. did you go into the game and do the tests yourself? You should go into the game and do the tests yourself. Both this nerdy one and the napkin math one.)
