Physic discussion thread

Oh yeah, that’s a nightmare crest and I now remember spinning there many times in the Retro even when being extra careful. Very odd.

F-Retro looks very polished and immersive in VR so I hope it gets the diff treatment soon. And whatever is causing the high speed spins.

@Maser V6 are there recent diff graphs from the F-Retro? Would be good to have some showing the low speed snap-to-good-turning that was also noticed in the Retro thread months ago. Just for reference for later use.

 

Oh no I just meant visually. I love looking at the gauges and suspension. Just holds up really well... until you drive it .

Ok good to know that there are graphs. Those grass-tarmac tests tell one story, but I think the only way to know for sure that the sudden low speed bite is at least partly on the diff is to look at tire speeds. I say this, because we also get tire changes with updates.

 

purely for speculation this at this point is bottomed. too early (1st bump?) and for some time prior to this for the main loss of control??


This may be the point just before the bounce of rear ?(2nd main crest/bump???) but its so difficult to get replay at correct place to take image
Just by max rear spring, all else default it becomes possible. (ha but not easy)
Springs are soft and underdamped for here in default. This is of course in addition to acknowledged aero issues
Car is unsettled long before the main crest
Edit Just ran the vintage to see clearer the suspension compression. It recovers well from the compression in the dip prior to rise. then as going up hill just before the sign on the left a small compression tales place 1st dip/ then fastly followed by main compression and jump
Will need to check the Jackie stewart pic as they seem to get perfect "air"at a different spot possibly.? possibly/probably track changed?

 

Hmm. If it does bottom out and hits the ground then that would disturb downforce wouldn’t it? But then I don’t know if that should upset downforce balance or just decrease downforce evenly for a moment.

 

Good evening, gents. If you blocked me, a gold star. Take the F classic white one there in the pic round Bathurst - but do it with steering assist on. Jump at first straight, very good.

Great game engine. Its my guess a lot of things which happen get corrected elsewhere.

for reference...

"The MADNESS engine isn't just about superficial visuals but also has a global reputation for highly-detailed simulation and physics. The tyre model, physics, and handling are the most advanced in the games industry—so advanced that professional racing drivers use it to practise for real-life races."

Now, the US spells practice with an S, but anyway...

Another great example of these things is Guapore and the new LM Merc. Its like a greyhound doing a run around a mini compact disc.

Understeer, over steer, ...its driver assist so maybe dont steer, anyway, as anticipated. Very well done.

I trust you lot will get to the bottom of things. Guapore, one of my fave tracks, is a well-made track across several iterations and indeed several games and the new cars the merc is prob the best perhaps, but its a fresh slate car afaik.

Both are amazing drives and are probably a conducive cross-section or reference point. Can put it against iracing, raceroom and even ACC and it holds up brilliantly.

Old Bathurst trick is go neutral round that first bend, slide. I will say this though - the default is as good as the WR holder. This, who was he?.. this crystal eminence fellow hrm. Turning down the ffb and of course default, got me very close if not for a spin out on the last turn. Which lol - lets face it, was because I turned down FFB in the first place!!

but best settings are 69 and 89. because: computer game and preference

 

Thousands are dying daily with this virus, likely many more for some time.
Lifes to short

No thanks

 

I would think so, reminds me of skirted f1.
As Crisom's images have show though. Its more aero than suspension by some larger degree.
but not seperate imo

Gt1 Merc aero is fine/superb here. You can have a Mark Webber moment if you wish through setup or (rr bumper prevents the full Le Man flip!!) blindingly flat fast and full control. ran out of time last try, somewhere in the setup maybe even full air with controlled landing. The car is fine though in default. Taken over from the Brabham as favorite At the moment.
Back to work for a few hrs now

 

it won't.

It's a little comedic relief, yes. But look at how we use plugins in Skyrim compared to 10 years ago. It's only a matter of time when learning.

 

Please turn your attention to the hair pins at hockenheim, even on gt1 Merc, a God level SIM car, havoc with turn in then acceleration if you cut or chop around the curb. Let's get some moxy on that riff and get some optics on that diff. Need as many graphs as possible on this one.

One of the best things about gaming though and world's is of course reaction to the world space. I'd be hesitant to make it inert, got to keep it screaming and shredding out of such turns.

Diff preload up. Suspension tune rebound downward. Any better or this may not apply

 

I just tried the Lotus 72 at Bathurst and got a massive snap in that fast left-hander right before the downhill section starts.

Here's the corrective steering input and shortly after about a million Gs from the crash. Everything after that is me limping home to complete the lap so Second Monitor stores it



Not sure what to look for in telemetry. I checked that it wasn't the differential opening at high speed. But I noticed that there's a downforce tab in Second Monitor. That's quite a drop.



I apply opposite lock at the bottom. All downforce is gone within 0.2s. But the thing is that even with front wing at 3 and rear wing at 5, it's quite rear wing dependent at this speed. Front and rear downforce values are visible next to the bottom graph. About 20% more rear downforce at high speed. So if both drop to the same values, it's going to be front balanced right?

I'm not sure where to look next. There must be a reason for this. There's up to 100mm ride height at the rear when this happens instead of the usual 50-ish. Suspension travel drops really quickly.



That makes me think that the crest is just too much with the default setup or that I should simply approach it from a safer angle. But should the car respond with a snap like this in the first place? There was absolutely nothing I could do to save it, the rear just shot out.

Also I have to say that this level of analysis is way over my head. I'm not an expert in aerodynamics or suspension. I wanted to clarify that. But it's a fun thing to look at and it would be interesting to hear if someone has a factual explanation.

 

Quite an interesting discussion going on in this topic, even tough most of it goes completely over my head. The lack of on-throttle oversteer on slower/medium corners and the sudden snap-back oversteer on hairpins is something that does bother me quite a bit on AMS2, maybe more than it should, so it's nice to see some conversation going about the subject from people who actually seem to know what they're talking about, instead of the ignorant ramblings on the matter that I posted some time ago (sorry).

But I did need to use the ignore button in order to be able to enjoy the conversation... holly molly

 

I don't know if "knows what he's talking about" is an accurate description of me, but I did want to investigate the root cause rather than settling on "it's just the Madness feel". I've tried it in game before, but using telemetry popped into my head just last week. And then after posting some graphs I heard that it's an old subject on the beta forum.

It's interesting how people focus on different effects from the problematic loose power / tight coast differential setup. You mention the lack of on-throttle oversteer which is valid. I started to wonder why there was no liftoff oversteer after GamerMuscle said that on multiple videos. I noticed that there is if you drive the open differential cars. There is a lot in the Formula Trainer without being snap oversteer.

(There might be some liftoff oversteer at above ~80km/h in the newest cars with LSDs, but it's very little and likely coming from some other variables such as their tires responding to weight transfer differently.)

After this last update I really noticed the snap-back oversteer at slow speeds you mentioned and had to dig in. Sure enough it can be explained by the differential.

 

Not tried 72 yet as such there but plenty have.
Nice to see a graff on downforce, should shed some light elswhere.
Ref the Brabham as a car that has ground effect, (and the lower the more the effect) need to run more for visual, but gave the impression (pic running on ground) its kind of sucked low for longer distance going up hill, and remains so (front) over both bumps.
Speculation of course but replays need looking at this end.

 

Oh yes - I must be the miscreant.

Well thats why I can set you straight and it won't hurt your feelings. The guy who is doing it doesn't even know what he is on about - he does not know how to use his software - and all that jazz- Rezia pay people to do that and they are already aware of the problem.

The other thing is, it should not be affecting your enjoyment of the game - its a niggle - and YES its a GAME!

Ahem. So all that is approaching autistic levels...guys... seriously, its holiday time and the BEST thing you can do in your life (and I know messages even my long ones only take a few mins to write)... is spend time making graphs and focusing on esoteric numerical/autistic-level things?

But I am playing the game when I get a moment, doing other things, and checking in on progress. Its hardly unhealthy levels of commitment - though I fully understand this attitude is not endearing whatsoever in text - sorry to hurt your feelings!

But I mean thats the best you can do with your life? Make a useless graph for people who already can and do make much better graphs etc. And heck even have access to basically an OPEN game engine afaik.

You have waited months to play this game properly, whats a few more? They did what they did in about 1 year which is pretty good as far as engines go and indeed, learnt etc. They have I am guessing another 1 or 2 years of working on the game and adding features etc, its hardly something to bog them down and drop the plot.

This graph nonsense is absolutely whimsical, sad and a little autistic.

Get a REAL car - go WORK on it - get an interest besides going after fictitious games/graphs in ficticious ways. It doesnt matter if that car you get is a 500 dollar banger guys, haha, no really, its better than doing what you're doing...

Its not science man, it means you need more meaning in your life - and thats the crux of my attitude on this.

If the game is something not good enough for you, then check in at the next patch - they are more than capable of nutting things out now they are aware - but to be honest its not this huge impact on gameplay that you believe. Maybe pre October it would be, but not now.

FAILING ALL THAT - go and move onto the next major problem you think the title has before you can actually feel as though you can enjoy it.... but maybe leave out the graphs for real scientists, computer programmers and engineers and the parameters they would set up for these things.

A result is not necessarily valid. Just because of a graph. And of course many opinions exist, but you may want to respect my stance on this before you go throwing stones yourself. And with this I will probably check out of this discussion. I am not one for wanton derision on my part and I can do without toxic internet culture.

The only reason I stuck around in this discussion - never mind that guy(s) I don't want to talk to them - please I am far too cool, one look at me would reveal that - I wanted to see if his 'work' was valid.

But its farcical.

__

So solve it for the unknowns - have a round table. Solve for an equation that you are trying to disprove.

That would be a dog thing to do - leave it there.

Actually I see these things I think they mean nothing, graphs ect, just scrambled results from scrambled results! maybe...

But maybe need an extra virtual torque converter in there with a reference table and LUT kind of setup - whatever it funnels down to, convert it (correcting it ultimately), and dissipate the numbers that do not make sense in whatever conjoining equation is devised. It would be corrective and weightless. If possible. It would connect to the steering wheel/column and the dif. It would switch on in the situations. No one would know.

And those I would take the median of those 'troubles' and design it to fix those. Where it did not work I would put dampeners into the equation - and ALSO I would scramble it +/-20%.

40% is better when you do not know what you're doing - in order to 'guesstimate' but here they could do 30 or 20% - they are not actual cars, just representations, no one will mind.

So thats the round-table. They will get it fixed somehow. Maybe that is doing what a brain may do but faster, its parsing the results and saying that can't exist and using an array of values. I do not know how the engine works. Torque converters are great for isolating (and exposing) certain numbers/forces, nullifying them, amplifying them.

In other words - as inelegant a solution as a new torque converter on a separate line is... and who likes complexity OR length heck I do not - you guys maybe I don't know because we knew it was happening, brought out the graphs too early. It could be practical or theoretical. We do not know.

 

See the above quote - this is not a learning thread though - its a stab in the dark thread.

You must know me by now though unless you stupidly blocked me - physics has a LOT of constants and equations. Functions, arrays, coefficients, laws (things that are worked with but may not even be applicable) and theory's.

IF it can be done in the unreal engine I guarantee you it can be done here. Physics is maths, and with that and economics we resolve there world into a picture and describe it for how and what it is... and by extension we logically devise engineering things. At each stage one specialized person comes in sorts something out.... but we all already knew the problem so we didn't need the graphs.

That was just whimsical. Kind of cool though.

And if you want to learn physics don't delay. There are a number of documentaries, online resources, khan academy, youtube physics (general and interesting things), heck even chemistry and biology can be interesting. It helps if you have done it at school but honestly its not that difficult the low end stuff. And its applicable every day.

IF (probably like these guys ) you are technically minded - AND you learn well by reading, as in if you pick up a manual and its not daunting (and forget my verbosity, concise and clarity) then you will have no problem grasping most physics except maybe the equations, but not all are difficult and a lot is easy to visualize and interpret. Solving things is just one part of it.

With my virtual torque converter - you just need a magnitude (and some look up tables to know which section to go to), which it interprets (the fluid) and then it gives a reaction. A mere tool, maybe, but it can play on the geometry which is set up by the user, which contributes to the displacement of the objects.

Its not as difficult because there is already a heap of telemetry, and there's already an output - lol - they are JUST shaping it a little more. I am sure an engine like that would have some kind of centripetal force.

Its a model. Of course in science they have a method. So what I did there was give a theory, and its based on the hypothesis that the two wheels spinning may not be easily solvable (needing a re-write or plugin) by just changing a few numbers.

Its a theory based on incomplete data (I don't know the engine) - so it has ZERO exact application just in its current thing - its merely an idea - so absolutely what I just said in these long posts is in any way shape or form descriptive or instructive, whatsoever - its theoretical. It may be impractical or useful.

Time is money though so I bet you the most elegant solution will be the one to go with. Even if that has no use or other uses, there will be only one best way to do it, I am guessing.

Anyway the cool thing about centripetal force is that its a combined effect. Maybe a canned one, don't know. I detected a difference at the apex for some reason and its possible the engine may understand the apex - well if its drawing an infinite line/plot - thats great, because as the car is turning its moving towards something via its wheels. Not nuts and bolts but it can be measured and probably accounted for.

At this point theres static force and various frictions and coefficients allowing for spin-speed so to speak. Reducing coefficients there helps the car turn. But when working out the values the max is the min, for its purpose. Values in the formula like the maximum is assumed. So its all plug and play when finding a new value.

The minimal static will allow slide. So a lot of this will just correct coefficients. And velocity.

Those are not my forte by any means, but models are a little bit more down my alley.

This does not mean a whole lot, but its an insight into the way things are going on under the hood.

 

I have been focussing hard getting at least some of this liftoff oversteer working in AMS2.
The reason is that because of some leg injuries Im driving with a fully handcontrolled wheel SRW-S1 (no pedals) - so to still have some finecontrol of the thr/br I cannot turn the wheel more than about +- 15-20 degrees.

Therefore I HAVE to use a liftoff oversteer way of getting the car to rotate when I enter a corner.
My setups does therefore allways use about max diff coast values.
But because there is something wrong(IMO) with the diff implementation in AMS2 then I also have to exaggerate the close to not existing liftoff oversteer in AMS2 - by a brutal shifting down in an (often) too low a gear at entrance of any crucial corner.
Explanation: Its the engine braking that makes it possible provoke it

ByTheWay: So it it is possible to find a kind of liftoff oversteer in AMS2 - but because the effect is so nimble you have to be brutal to evoke it

 

Not quite a lament to discussion thread about car behaviour, customer feedback, questions, on it, be it beta or main. Yes validated by and acknowledged by devs, with data and graphs and dedicated testing by many borne out of passion for the sim. Some extracts.

They did

correct

Lots of players inc not in forum!

As Graphs logs/updates tweeks to diff for instance also aero show.

Many have said also hence discussion

feedback from all corners and graphs from many sources show

It would be nice, but

fantastic,you were not the only 1 Now hope it does.

we all know and see this with constant updates and fixes even 2 or more on same day (beta)feedback reports testing even over holidays

Gladly done

wow

yep

Graph results/dedication testing discussion and more

Hope

thats was the good intention (as you know!! and thanks for your contributions

Oh dear

We know, with help also hence having a forum and testers
Coupled with feedback/discussion/reports/date/verification of findings/public comments/beta testers/.

It has.
Even your earlier post/feedback/likes. you just dont see it.

?

Was work in progress here. Been somewhat put off.

I think this is Qualified answer, given MC

misunderstanding

car behaviour /physics?

Please, sadly too late?

? Factual representation of "felt behaviour"

Ok @Raceracerace
No its a Shame the thread has turn to this. kind of -
Not quite a lament to discussion thread about car behaviour, customer feedback, questions, on it, be it beta or main. Yes validated by acknowledged by devs data and graphs and dedicated testing by many borne out of passion for the sim. Some extracts.

 

I have noticed you mentioning this before. I believe everyone should be able to drive sims to a reasonable extent of course. It’s cool that hand controls exist at different price levels to make steering wheels possible with leg motor issues.

Liftoff oversteer is expected from slight modulation of the throttle in many cars (not all!). It makes it easier to adjust the driving line with small throttle corrections. I can see why its lack is especially frustrating for you. Breaking rear traction with aggressive downshifting or hard braking to get the car rotating often means having to finesse a slide which I assume is challenging with your range of steering angle.

Have you driven the ARC Camaro or Super V8 a lot? They are built around a locked rear and I find them easier to control in a small slide than these uh... more elegant race cars . No offense to spool fans.

By the way. If I wanted to be pedantic for a bit, the physics are fine. It’s just the various car components that could still use improvement . But people often throw everything under the physics umbrella so that’s the term we use.

 

Hey everyone!
Yesterday I did a AMS 2 video with the F Ultimate, racing the A.I @ Interlagos! It was a good experience.. if it only had not ONE big issue, concerning the handling model in general!

See the video up here:

Tell me what your driving experiences are with AMS2 down in the comments!

Thank you all!

Ralonso