To Reiza: What physics do you actually want?

@steelreserv
Can you plot in telemetry longitudinal and lateral forces provided by the tires on development build?

If so, could you provide a graph of corner exit with tires spinning on (as an example) the opala old stock?

Thanks.

 

Well let me clear up a few things:

First, i didnt say the cars accross the same build were directionless. I said differences between builds seem to be. If anything, the cars in the same build share too much the same traits.

Second i dont think my question is enigmatic at all. Your long answer starts by stating you have access to information that tells you the radial tires in the lotus should be "stiffer" than the bias ply. Source for this data? Because as far as i know, and it has been stated by people like Niels Heuskinveld, tire data is extremely hard to come by, specially from 1986. IF anything, the sidewalls should be softer than a bias ply, because that was one of the biggest benefits of a radial tire, which allowed more side to side movement, with the yes, caveat that the loss of grip could be more abrupt,but NEVER to the degree exibited in the game. A radial tire also provided more traction than a bias ply, due to the same softer more flexible sidewalls, again, at the expensive of possible snapiness or "wheel hop".

Now as for the Lotus 98T, i actually have a mod for AMS1 featuring that car, that has been years in the making. I have books about that car, and the turbo cars in F1 in general, i've seen a lot of footage, and i have many many kms fine tuning those physics and that car in particular. I dare to say nobody knows that car better than me in the simracing world. The reason why i used the PCars2 car as an example, and not my own, was because it is much more relevant since we are talking about the SAME PHYSICS ENGINE.

So again, in my opinion, the way the Lotus 98T behaves, with the Historic/vintage tires on, in PCars2, is a better representation of that car than the one you provide.

Now the kicker, that car in that game has much MORE wheelspin than the one in AMS2. But fundamentally, it has more lateral integrity, and crucially, you can control either the wheelspin, or the rear sliding, because this is much more progressive in that car, while at the same time, being a lot more pointy and suffering much less understeer on corner entry.

This has nothing to do with how much i have to coast on the throttle, but what happens if i DON'T. Real cars are driven by feel, not by memorizing how much i can squeeze the throttle or i am dead, specially high powered machines like this. I am not saying they should be easy to drive, i am saying they should be much more PROGRESSIVE and comunicate a lot better with the driver.

Lateral and longitudinal grip in the grip circle are not separate, but they are also not mutually exclusive. You can have imense wheelspin, and stil maintain some lateral integrity. What happens now in the game is that this lateral grip disappears instantly whenever there is any weight transfered to the back, almost regardless than where i am in the rev or torque range. I made experiments in which i realized it was very easy to flip the car sideways even being way off the maximum torque band, just by applying any degree of throttle and steering input. So no, i dont agree with you, it certainly doesnt feel anything related with torque, but rather with simple weight transfer, and then the problem gets compounded by torque.

As for dumfries lap, his car is NOT skidding the fronts into corner exit. IF they would be doing that, his tires would be killed in a few laps, that would be a highly undesirable trait, that has no place in a specially built machine like an F1. His car is describing smooth arcs around corners and yes, he only gives it full beans going down the straights, but that doesnt stop him from flooring the throttle hard when he comes out of them. In the Patrick Tambay onboard we can see that even better.

As for the direction of the physics, your post got confusing. I didnt get if you are trying to say that you make compromises for drivability for the sake of the users ,or if you are fully modelling the car in all its perceived difficulty, and then letting people setup the car to change its charater?.. because the later is impossible now, as i already alluded to.

 

Let me begin with saying I know an onboard video without telemetry can only tell you so much, and a lot has to be interpretation, but since you brought the video up IMO it's fair game. I'm also still interested if there's maybe a flaw with my throttle inputs. I can only compare the same cars in differnt sims, which have different stock setups, but eg in AC the same car does not snap oversteer like in AMS2.

That being said, having just watched the video, the car looks quite stable when Dumfries is already on the throttle shortly after the apex on most corners, even out of the 1st gear corner after the long straight, and the slow right hander before the start/finish straight. The snap oversteer AMS2 exhibits seems to bee missing here. His wheel movements for the most part don't look like he is catching the rear of the car and he is certainly not countersteering while the engine sounds like a good amoput of throttle is applied - but again, that cannot be said for certain.

With all we know and don't know about Dumfries car setup in this video, and since both the car and the track are in AMS2, do you think you could recreate this lap which was about a 1:27 or 1:28 going by Youtube's video timer? In the sense of, match the lap time and accelerate the car without the need to countersteer out of corners? Maybe even provide a telemetry trace of steering, throttle and rear wheel slip for peolpe who are mor knowledgeable than myself?

I'll be giving it a go myself later today and see if I can match the lap.

 

1.28 is indeed a slow lap, with a full tank.

In this video, we can even see Patrick's legs:

Two good laps from Nakajima. True, its the 99T, but still same era, same tires, same engine. you can clearly hear the throttle application mid corner, and how he feeds the throttle after the apex of almost every corner:

 

Oh man I get pretty nervious when we are talking too much about my favorite the L98T - because any change in the cars physics or TM will by garantee make it more tame/boring.
I dont know if this car simulates the RL L98T exactly 100% - but as it is now its extremely challenging and fun.
Hehe and its not exactly useable for all the casual drivers
Just like it wasnt in RL
So please Reiza keep your fingers away from this car bomb - PLEASE

 

I agree slightly that the L98T in the video dont look quite so brutal as in AMS2.
But I prefer a bit more brutal than the opposite

 

This is exactly what i mean. Especially in modern cars the throttle is not an on off switch. You can pretty easily control tire spin with your right foot. And thats exactly how most sims make me feel. Im memorizing not driving by feel.

 

Now, posting again this video from Nakajima, notice how he has absolutely no problem with corner entry understeer, as its aparent when he even has to slightly counter steer at the entrance of some corners, like turn 1. Also, notice how he can apply some throttle, not full, but enough, to still keep or increase his speed, without any drama around such long corners. In the slowest corner of the circuit, you can see him counter steer quickly, because the car there told him that the rear wanted to step out, again, a very lose car, but the rear never really does so, he just felt that the rear was sliding from just cornering alone, without him increasing his throttle input, and even in such a slow corner the car is neutral to oversteery, and yet, you never see the rear step out massively even once in this whole lap:

 

I absolutely agree, and it's one of the reasons my physics work has had such an emphasis on driveability and communication, without losing anything in the realism department. Real cars talk to you the whole time, thats one of the pleasures of performance driving.

 

Hold on, why is the diff locking at all? This is a limited slip diff, which as the name implies, it allows a limited speed difference between the wheels. It should never be fully locked. This is not a detroit locker or any sort of device. Even either the clutches fully compressed, or by gear action (depending on diff) there is always wheel speed differences, unless your diff is setup in a strange way, because you never want it to lock fully in the middle of a corner, no matter how much throttle you apply.Detroit lockers are a thing of the past for a reason, even cam and pawls werent able to completely lock the wheels.

 

Again, not being pedantinc, but that advice is absolutely contrary to the literature about this car. I happen to have the real gear ratios they used, and the first few gears are very short between them, You can even tell this by the videos in adelaide. This is because you abolutely didnt want these engines to fall under their optimal torque and power range, or else you would have nothing in terms of power, followed by everything, and you would be left there spinning the wheels and have all sorts of hairy moments.

No, it is much easier to control the rear and the wheelspin when you have the torque at your command under your foot, and when you can just go quickly throw the gears to minimize wheelspin, and get quickly to the top speeds, something i tried to faithfully reproduce.

 

We are very adaptable and can become complaiceant as to how much we adapt. and cover error. did it for over a year with Bt44

This reminds me of a rl porsche I posted in beta debate to show it will wheelspan happily whereas the newly released car did not (possibly still does not) . + others

Video with pedal inputs to show/replicate exact issue with car

Is relevant.

Yes a look here would help (grip balance) + imo from comments here throttle response.

Shortened post to.
3 things pop out here to me in this thread. from comments

Long/grip
Lat/grip

Throttle/ response. ( on this We have been here before and worth checking imo )

I think vid examples of these would help as not every player has the same experience (again).
for whatever reason.

 

Equal amount of torque being transfered to both wheels is not the same as locked wheels. Locked wheels is when no amount of speed difference is allowed. Or else an open diff is not an open diff. An open diff is called open precisely because it allows an infinite amount of wheel speed difference between the two wheels.

Having a locked diff on acceleration is only optimal if your car is always traveling in a straight line, in a perfectly flat road. In any other situation, this is not the case. (some debate its not even the case on the first one, but thats for another day). Torque is only diverted to the slipping wheel, if there are wheels slipping. If this isn't the case, its abolutely desirable to have more torque transfered to the outside wheel, hence vector steering being a nice feature to have. IF one wheel happens to lose traction, then the limited slip system is there to limit this power loss in the smoothest way possible, and thus NOT afecting the trajectory of the vehicle too much mid corner. So in other words, unless you want the rear to kick out suddenly (sounding familiar here) the last thing you want is for your differential to lock both wheels (again, binding both wheels to the same rotation speed) mid corner.

So i assume you are either interpreting the graph wrongly, or something is upp with that differential. I assume the first, because just because the wheels have the same speed that doesnt mean the diff is locked, as that would obviously naturally happen if the car is moving perfectly in a straight line, with no asymmetrical ondulations on the road.

I repeat, no limited diff of any kind is ever locked, and a limited slip diff what it does is transfer MORE TORQUE from one wheel to the other, when the mechanism is forced to do so due to wheel speed difference.

But now i wonder...could this be the reason for all these problems? Because if you believe that both wheels should essentially be bind together in almost all situations, this would explain somethings.

 

I have done this real-life to AMS2 comparison before and seen a similar outcome.

When you drive a lap at a similar pace to a real-life lap, the car behaviour seems very similar. I dropped the clutches to 4 and softened the front ARB two clicks then drove this lap. The car was very easy to control; had no real understeer and a few twitches on power but nothing unmanageable. But I am driving nowhere near the limit in AMS2. I am barely touching 60-70% of the available lap time.

 

IRL they were probably scared to push 110% due to crashing. In AMS2 if we crash we just note what we did wrong and move on. Believe there is a mental factor

 

What tires did you use, how much fuel did you have, and what was your exact laptime by the way?