I didn’t aim for lap times when making my cars, I was going solely for the Driving Factors piece of the prompt. Really just meant to be fun more than fast. Obviously flubbed the 10s entry but I’m hoping the 90s entry proves one of the most fun to drive.
I haven’t had much time to test before this, does 45 profile in 2020 behave the best in terms of balance of rigidity versus contact patch? I tend to default to 35 profile purely to reflect IRL conditions for “sport” tyres (which I now agree overperform on lateral grip).
My experience with BeamNG is very limited, but I did do some amount of testing for the Wraith.
I seem to remember that it either stopped getting faster once I went below 45 profile or got ever-so-slightly faster, but became twitchy and less forgiving at the grip limit, which would have gone against my goals.
The Wraith being a relatively firm ride to begin with, I suppose there could be different results with different suspension setups. I haven’t done any testing to speak of in 2020, so I just used a value I knew was OK.
Lowering the tire profile increases the size of the contact patch, thus providing more grip, but too much grip on the front end (relative to the rear) leads to oversteer.
Also, what body set is the Wraith H built on? I thought it was either a '69 Mobula or '70 Fiesta body set, but now I know it’s neither of those - so what is it based on exactly, then?
Actually, it turns out to be the '67 Muntu - it’s an odd choice for a modern FR coupe, considering its age, but it’s small and light, which matters quite a lot in a time trial.
alright finally had a look at your GR Supra replica. It looks pretty damn good actually, I’ll double check the engineering later but styling wise, I’ll definitely take it! Thanks a bunch for doing that.
EDIT: Decided to go full send and try to replicate more a later model with a more powerful designation since that’s the entire idea of this segment in 2020. Slightly retuned to reflect the gear ratios more accurately after searching on some forums, the wheel sizes have been corrected, as has vehicle weight and balance. Also added a couple more slight details for fidelity so here we are. For reference this is the base for what I’m exporting to Beam.
A90_GR_Supra_-_J29_DB42.car (78.9 KB)
This means that all the cars have now been completed, and, funnily enough, since I have some parental leave, and looking after this baby isn’t actually an unmitigated disaster as I figured it’d be (just an intermittent one 
) I might have time to start chipping away at driving and footage gathering.
2 Likes
That Toyota document I found is probably the spec of the original launch model, which doesn’t actually seem that powerful
yeah agreed, it started out at “335hp” but the version they provided to journos probably quoted that figure but was sandbagged by a good 50hp or so judging by just how badly it destroyed the Z4 in a drag so I figured screw it, everybody’s got a much more powerful version by now.
I have now fixed the weight distribution and weights for the 10s entries. I have decided for ease of conversion to add an 80kg driver (me) to the total weight, but not change the weight distribution.
I liked the look and I wanted track width. It’s not perfect, the exhaust doesn’t really fit in the channel, but I think it came out looking alright.
It wasn’t so much over/understeer with the side wall testing, more about whether traction loss would be gradual or sudden - one is easier to catch and play with, the other is game over.
That Supra vs Z4 drag sure was something. Hadn’t seen it before.
I am pleased to report that I’ve completed the first of 9 segments of testing: the 10s cars at Bathurst. I originally intended to start from the 90s and work my way up but ended up doing it in reverse order.
The good news is that I know this track like the back of my hand, so you can be confident there is no learning bias in the results. Nonetheless I would read the actual track times down to the nearest tenth, maybe give it a margin of error of 0.3s or so depending on just how much of a handful the car was.
The results so far confirm that the many restrictions I put in place did help narrow down the range of characteristics so as to garner more meaningful insight into some of the factors behind export handling behaviour, namely tyre model. I’ll finish the other decades at Bathurst before further comment.
11 Likes
Further update:
Have been fixing up the 00s entries now and have run into a very curious issue. In one of the entries, the engine refuses to load. It loads in the Engine tab fine, but it won’t load into the trim, and therefore when exported, the engine is just… missing. The car behaves as if there is an engine there but for the purposes of weight calculations it is not, which throws the weight balance of the car way off and I can’t actually meaningfully adjust the car’s weight balance using the usual trick of moving the heaviest component i.e. the engine. Weird. Anyone encountered this before?
Have you tried selecting the drivetrain option again and see if it loads up then? This is similar to what I get when the car didn’t quite finish exporting because of game crashing.
by now we’ve tried about everything, clicking through all the tabs, hiding and unhiding the engine, etc. Reimporting, re-exporting and reimporting etc.
The main problem we have is that the engine itself loads, but the engine block simply isn’t being put in the car, like so:
It just… refuses to turn up
Having discussed this with Riley unless someone can come up with a solution to this shortly I unfortunately will probably have to drop this entry and call up the next least ineligible entry.
Update: have elected to recall @Xepy 's 00s entry, which turned out to be a real pocket rocket to keep Howler’s 90s entry company up the fast end of things. It has a bit too much rear tyre, but I just so happen to have used to 10s entry results to optimise to the ruleset and found that actually you don’t need that much tyre to handle a lot of power… at least on a race track. Will discuss this further later!
The new roster:
4 Likes
I have now completed the 2nd of 9 segments of testing, the 00s at Bathurst! Cars placed in vague order of raw power, with a couple of exceptions to prevent too much colour blocking
This segment was almost purely comprised of tourers, interestingly enough, in very stark contrast to the S2000 replica I threw in this lot…
5 Likes
If the Wraith starts feeling too skittery by comparison, feel free to add 1-2cm to the rear tires (for an unranked tester, I mean) 
I found it quite transformative coming out of slow corners, not that big of a difference at speed probably.
one comment I will make is that with the change in Beam export tyre model, this test is demonstrating that what matters most is 1) the profile 2) the ratio between front and rear widths. With what I’ve learnt from this test, I made a 1250kg 450hp mule (the maximum power and power to weight ratio) car which was only allowed 235 on the rear (and that was with 2 penalty points)… and with 215 on the front it controls just as well if not better than a lot of the current entries with much much wider tyres. It was also the fastest overall by a huge margin. I also made a 1600kg entry which was allowed 285 on the rear, and well, the corner speeds weren’t any faster but it was slower to accelerate due to being heavier. All of this on 40 profile tyres front and rear.
I’ll put in a much more detailed commentary later, but while this is a very incomplete physics model, there are some relevant aspects to the arguments about how much tyre one really needs in real life track and performance driving. I am still debating whether I need to go full send and invest in widebody and get up to 295-305 when I big power the Civic (which will happen next year), or should I keep it on 265 square for lighter response and better handling. The current tyres can handle the current power no problem… that said this is also FWD so that’s another consideration altogether.
90s troubleshooting
Okay so aside from the unfortunate issue with Riley’s car in the 00s, we were going quite ok with the meshes exporting and all that. But the 90s yielded some interesting problems:
@Riley 's car has some, uh, parts that generate meshes where they shouldn’t. I hope this doesn’t cause me to crash into stuff. Also that lip for some reason generating a node box far lower than it should is going to cause braking issues. This is all part of testing, of course, so I’ll report on that in detail when I get to it.
@NormanVauxhall 's car generates a mesh that’s lower than the car itself. I think this is a known issue with that body. It doesn’t stop driving normally but on bumps, steep elevation changes and heavy braking there may be some sparking.
@Xepy this is why I said don’t go ham on the interior. Each interior fixture counts for a surprising amount of weight and so this is what happens when I try to get the weight to spec (kerb weight plus 80kg driver). As a result the minimum weight the car stays intact is at 1130kg which is equivalent to a 50kg penalty, so, let’s say, me and my wife riding passenger for this particular test.
4 Likes
Oop I basically just used a car I had already. I can send you one without interior if you like.
Actually you know what, you might as well! I won’t be getting to drive it for several hours yet.
EDIT: so Xepy did send me an interiorless version and the weight didn’t exactly change much at all. Maybe 6kg at most. I’ll take that 6kg but still, clearly my original impression was a bit wrong. Maybe this body is just heavy in general because it’s node dense?
3 Likes
Next update: thanks to having parental leave, I’ve managed to squeeze in the rest of the driving of the 90s at Bathurst, so can declare testing at Bathurst is complete.
Again photos not reflect of vehicle performance, but I had to make a bit of a joke because somebody submitted a Soviet New York Taxi that looks like it has no business being at today’s test
I’m going to make some graphs and charts and hopefully string together a video or something and that’ll probably take me a week or so. As to when I can get the other 2 segments of testing done, I honestly don’t know because I’ll be back at work by then. But at least the exports are all taken care of and I’m aware that the tyre model appears to have changed again with the new Ellisbury update, but nonetheless we should get some relevant info.
11 Likes
Or rather, an ex-Soviet copy of a New York taxi:
Thankfully, aesthetics is not a judging criterion here, but it could redeem itself if it drives well enough in your hands.
As for my entries, I will most likely revisit them to account for the changes implemented in 4.3, now that I am currently running the open beta instead of the 4.2 stable release.