Physics The Physics discussion thread

[RC45]

I have been working on "the definitive C5 Corvette mod"... but with the AC 1.9 sound update revelations about Kunos yet again breaking modded cars I may switch over the rF2 instead. But either way, here is a little teaser LOL.

Each years correctly badged and detailed Fixed Roof Coupe (1999 & 2000) C5 Z06 (385bhp 2001 & 2002, 2003 & 2004 405bhp versions) - with 50th anniversary details, 2004 Z16 version all in stock form with fast street setup and slick tire hard cor setups as well as lots of different real world modded C5 Corvettes.

Then maybe I will also do the Coupe and Vert versions as those 3d models are ready.

The 2001 Speedworld Challenge Z06's are also ready to go.

My chassis notes & photos & measurements

The 385bhp 2001 C5 Z06 WIP

Speed TV World Challenge Series Z06

I am about 100 hours into the mod... using lots of access to real data, C5 Speed TV World SAE aero white paper, real cars, experience and custom made and purchased 3D models as well as multiple engine type recordings (stock LS6, modded LS1 & LS6, modded LS7, roots blower, centrifugal blower, stock pipes, long tube headers, Borla exhaust, Billy Boats, Corsa. But if Kunos is going to keep breaking the core game why bother?

Although to be honest, as I have been working with real chassis data I have noticed that the AC physics engine is very very very far from being a real representation of real world physics using real world chassis data.

I have had to alter the chassis layout of the C5 Corvette just to match the bizarre behaviour of the AC physics engine. Then when I make well documented changes to the setup, the AC car does not behave the way real cars do.

I guess I will just keep my install at 1.8.1 and finish the mod for personal use, as there is no way I will keep chasing Kunos moving sound goal posts.

 

[NightEye87]

But if Kunos is going to keep breaking the core game why bother?

Would the update to their FMod stuff really break everything? I have no idea how FMod works, but I can't imagine doing some updates will break everything or make it hard to fix if it does. I know that the major update a year or so ago did break stuff, yes.

Although to be honest, as I have been working with real chassis data I have noticed that the AC physics engine is very very very far from being a real representation of real world physics using real world chassis data.

That might be worth reporting, then. AC's physics engine should calculate all movement and its coherent responses, so if something is not right they need to know about this.

Would be a shame to not get those mods out to the public. Good luck

 

[RC45]

Would the update to their FMod stuff really break everything? I have no idea how FMod works, but I can't imagine doing some updates will break everything or make it hard to fix if it does. I know that the major update a year or so ago did break stuff, yes.

That might be worth reporting, then. AC's physics engine should calculate all movement and its coherent responses, so if something is not right they need to know about this.

Would be a shame to not get those mods out to the public. Good luck

Would a sound change by Kunos break everything? No, but it would mean that I would have to keep revisiting a finished mod to continually make it work with the ever moving goal posts Kunos presented - no thanks, what a lot of wasted effort.

The mods will be public, just not in Assetto Corsa form - probably rF2.

Reporting these real world physics behaviors to Kunos will do nothing, there is no way Kunos will admit their toy engine is fundamentally flawed in a way they already know it is - the game may behave like a car simulator responding to real world physics data, but it is not.

It is just a game engine that can be manipulated to have cars that FEEL like the real thing, but that FEELING is not based on changing values based on accurate real world input, it is based on changing values to non-real input values in a trial and error way until the in-game car FEELS like the real car.

That is NOT how a simulation engine is supposed to work.

If AC was at all accurately responding to real-life data, then after I use REAL chassis data to establish the the car BASELINE, I should just need to change that chassis DATA with REAL WORLD values to get a predictable response, I should not have to CHANGE those input values to OTHER numbers to get the response I expect.

It turns out that AC is a cute game for creating an ILLUSION of a REAL car, but the DATA used for that ILLUSION is not the real data but rather a massaged and manipulated version of the data.

It really is a pity, because once you have the data massaged into its no-longer-real-form, the resulting illusion is really good and a lot of fun to drive.

But if I can't take real data and use it unaltered then AC is just a toy that is great for pretending to drive a real car, it is not a meaningful simulator tool.

Ehm, why bother posting if you're that bitter about AC? It's not like your post will change anything about Kunos update plans. And how long has it been since the last "game breaking" update, that being the last fmod changes, a whole year almost? Compared with other games AC is really backwards compatible (looking at you Cities:Skylines)

Why bother posting? To save someone else that does NOT live, eat and sleep Assetto Corsa, (but just occasionally keeps up to date with the AC culture and life style), the trouble of spending time on a mod based on the early public announcements by Kunos that the game is long term mod friendly, only to come find out it is not.

 

[garyjpaterson]

Reporting these real world physics behaviors to Kunos will do nothing, there is no way Kunos will admit their toy engine is fundamentally flawed in a way they already know it is - the game may behave like a car simulator responding to real world physics data, but it is not.

It is just a game engine that can be manipulated to have cars that FEEL like the real thing, but that FEELING is not based on changing values based on accurate real world input, it is based on changing values to non-real input values in a trial and error way until the in-game car FEELS like the real car.

That is NOT how a simulation engine is supposed to work.

If AC was at all accurately responding to real-life data, then after I use REAL chassis data to establish the the car BASELINE, I should just need to change that chassis DATA with REAL WORLD values to get a predictable response, I should not have to CHANGE those input values to OTHER numbers to get the response I expect.

It turns out that AC is a cute game for creating an ILLUSION of a REAL car, but the DATA used for that ILLUSION is not the real data but rather a massaged and manipulated version of the data.

It really is a pity, because once you have the data massaged into its no-longer-real-form, the resulting illusion is really good and a lot of fun to drive.

But if I can't take real data and use it unaltered then AC is just a toy that is great for pretending to drive a real car, it is not a meaningful simulator tool.

I'm interested as to what you think specifically are the issues with the AC engine? I'm not saying you're wrong, I mean if there are real problems then its definitely an issue. I'm just curious as to what the specifics are, as I can't say I've encountered any significant problems myself.

 

[SportRotary]

It turns out that AC is a cute game for creating an ILLUSION of a REAL car, but the DATA used for that ILLUSION is not the real data but rather a massaged and manipulated version of the data.

My experience is a bit different. I have an FD RX-7 track car, and I've measured and made a kinematic model of all of the suspension points on my actual car. As a result, I found roll center heights, camber curves, KPI, scrub radius, mech trail, motion ratios, etc. Looking into the suspension.ini file, the Kunos kinematics model matches my kinematics model very well. I've edited the Kunos RX-7 model to have my (from my real track car) spring rates, ARB rates, ride height, alignment, damper data, torque curve, etc. I didn't touch the kinematics data, I'm just using the Kunos data since it matches my model so well. The resulting car in AC handles and responds just like my car, and I can run almost identical lap times at VIR compared to real life. Now, the one thing I haven't messed with is the tire data of the original Kunos car, but it seems to behave similarly to my tires (if they were creating any sort of "illusion," I suppose it could be done here). So, in summary, real world data = Kunos data and real world feel = Kunos feel (in my experience).

What sort of result are you getting when you use un-manipulated real-world data on the C5?

 

[RC45]

@RC45 , that's an awful lot of text and an awful lot of photos just to say that you are not going to work on mods for AC anymore. If you're going to move to RF2 then just leave, instead of telling everyone that you're going to leave.

Seems like attention seeking to me. In your own words, "why bother"?

Why bother? To save some one else the trouble of starting down the AC path. It is that irritating to start something, come up for air an find out the core is changing yet again. And its not even the important stuff in the core, its a 3rd party sound add on engine efor goodness sakes.

Most dancers around are OK with the floor they are dancing on. If the problem is actually related with AC software and you could investigate and report in official forums, I guess that would be a good job. But it seems that you are blaming the floor for hindering your performance at the moment. Go post it to AC official forum physics bugs subforum and experienced guys will tell you if you are right or wrong, and hopefully you'll have good discussion there. If you already didn't @RC45 ? This is not that place to do that, it is silly here.

Perhaps you are unaware that Kunos are not happy when communicating with people that critique their product - hell just look how the maketing & licening guy reacts LOL. These are not physics bugs, they are simply physics engine shortcuts.
The game is fantastic, the feedback is awesome - and the resulting "play" is as realistic as any game/sim has ever been, but trying said with raw unaltered data appears to not work as intended.

 

[RC45]

My experience is a bit different. I have an FD RX-7 track car, and I've measured and made a kinematic model of all of the suspension points on my actual car. As a result, I found roll center heights, camber curves, KPI, scrub radius, mech trail, motion ratios, etc. Looking into the suspension.ini file, the Kunos kinematics model matches my kinematics model very well. I've edited the Kunos RX-7 model to have my (from my real track car) spring rates, ARB rates, ride height, alignment, damper data, torque curve, etc. I didn't touch the kinematics data, I'm just using the Kunos data since it matches my model so well. The resulting car in AC handles and responds just like my car, and I can run almost identical lap times at VIR compared to real life. Now, the one thing I haven't messed with is the tire data of the original Kunos car, but it seems to behave similarly to my tires (if they were creating any sort of "illusion," I suppose it could be done here). So, in summary, real world data = Kunos data and real world feel = Kunos feel (in my experience).

What sort of result are you getting when you use un-manipulated real-world data on the C5?

Well, what I have so far (also comparing to data they use in their C7 Stingray model) is that in order to get the feel response as expected camber rates need to be exaggerated well beyond real values - into the -3 and -3.5 rage - in order to get the results with real world envelope ranges -2.5 to -1.9 (depending on tire type) the tire width and wheel offset need to be altered outside the proper envelope range.

The beauty of the C5/6/7 platform is that the stock cars have the built in adjustment capacity for height, camber, toe, caster by way of adjustable suspension pickup points right on the stock suspension - going from -0.5 to -2.5 can be accomplished by kicking out the lower a-arms using the factory eccentric or camber plates or if you have the upper a-arms shimmed just remove some shims.

The point being that I was looking for AC to be able to be manipulated by values from the real world operating envelope - not an exaggerated subset to get the correct "feel".

Now granted, once the "feel" is dialed in, the SIM is spot on - I too can place my in game version of my car around asphalt as it would be expected to be in real life, with the same types of steering and throttle inputs.

I am not discounting AC's results when DRIVING the car, I am just not seeing the same set of inputs needed you are seeing.

Perhaps the tires are the magic glue.

 

[SportRotary]

Well, what I have so far (also comparing to data they use in their C7 Stingray model) is that in order to get the feel response as expected camber rates need to be exaggerated well beyond real values - into the -3 and -3.5 rage - in order to get the results with real world envelope ranges -2.5 to -1.9 (depending on tire type) the tire width and wheel offset need to be altered outside the proper envelope range.

The beauty of the C5/6/7 platform is that the stock cars have the built in adjustment capacity for height, camber, toe, caster by way of adjustable suspension pickup points right on the stock suspension - going from -0.5 to -2.5 can be accomplished by kicking out the lower a-arms using the factory eccentric or camber plates or if you have the upper a-arms shimmed just remove some shims.

The point being that I was looking for AC to be able to be manipulated by values from the real world operating envelope - not an exaggerated subset to get the correct "feel".

Now granted, once the "feel" is dialed in, the SIM is spot on - I too can place my in game version of my car around asphalt as it would be expected to be in real life, with the same types of steering and throttle inputs.

I am not discounting AC's results when DRIVING the car, I am just not seeing the same set of inputs needed you are seeing.

Perhaps the tires are the magic glue.

So it sounds like, when real parameters are used, the car understeers and has wonky steering feedback? That's interesting, but it's encouraging that it can be fixed by tweaking a couple of things. Definitely double check your measurements; being a few mm's off on the location of a suspension point can change the roll center heights by a lot, which will upset the balance of the car. The resulting scrub radius and mechanical trail values are also really important for steering feedback.

Keep it up though! I would love to see a quality C5 mod in AC. My friend has a well-prepared C5 FRC autocross car that I have a lot of seat time in.

 

[RC45]

I'm interested as to what you think specifically are the issues with the AC engine? I'm not saying you're wrong, I mean if there are real problems then its definitely an issue. I'm just curious as to what the specifics are, as I can't say I've encountered any significant problems myself.

With what I have done so far, I am finding that every more exaggerated values need to be used to get the expected handling results - and I see this even with AC's own C7 Stingray.

In my opinion, and its just my opinion, if the over all engine is working as it should, and the chassis layout is entered accurately, then adjustments to the chassis using real world envelope values should result in the chassis showing real world results both in physics suspension layout and dynamic vehicle behavior.

I am not see this until the input adjustments are exaggerated. As I say, form what I see even the Kuno C7 Stingray suspension.ini uses exaggerated values to get the car to behave so well in game.

Oh well, haters will be haters and I expect the general Assetto Corsii to begin attacking in 3.. 2... 1... LOL

 

[David Dominguez]

With what I have done so far, I am finding that every more exaggerated values need to be used to get the expected handling results - and I see this even with AC's own C7 Stingray.

In my opinion, and its just my opinion, if the over all engine is working as it should, and the chassis layout is entered accurately, then adjustments to the chassis using real world envelope values should result in the chassis showing real world results both in physics suspension layout and dynamic vehicle behavior.

I am not see this until the input adjustments are exaggerated. As I say, form what I see even the Kuno C7 Stingray suspension.ini uses exaggerated values to get the car to behave so well in game.

Oh well, haters will be haters and I expect the general Assetto Corsii to begin attacking in 3.. 2... 1... LOL

I'm trying hard to understand the issue but the description is so vague... what values do you need to exagerate? and by how much? what's exactly the handling issue/issues? Also, have you edited the tyres?

 

[RC45]

Yes i can understand where you are coming from but if you are gonna have a page by page volley of this issue then it warrants its own thread, what you have said doesnt really help in the development life cycle after you have replied about it ten times to the several people that have replied. The problem is noted, if you want to discuss it further then make a new thread for it else this will get too bloated.

Ok - I am game, but where should such a thread be opened/put so as a) not to be shut down because it critiques Kunos product and b) where it is not going to be over run by the Kunos cheer leading squad?

I am all for a meaningful physics discussion - but I don't believe it can be done without folks getting all bent out of shape.

*edit* -- see above for an example of an such a response.

I'm trying hard to understand the issue but the description is so vague... what values do you need to exagerate? and by how much? what's exactly the handling issue/issues? Also, have you edited the tyres?

My description I snot vague in any way shape or form.

You do not need camber of -3.5 front and rear for a C5, C6 or C7 chassis to respond very well - except in the Kunos universe that is. This is just one example. And I believe your question/response is technically clogging this thread

 

[David Dominguez]

You do not need camber of -3.5 front and rear for a C5, C6 or C7 chassis to respond very well

Also, have you edited the tyres?

That's why I asked that. All suspensions/aero data is very welcome always, but at the end the only stuff that is in contact with the ground are the tyres. And if they are off... everything is.

 

[Stereo]

My description I snot vague in any way shape or form.

You do not need camber of -3.5 front and rear for a C5, C6 or C7 chassis to respond very well - except in the Kunos universe that is. This is just one example. And I believe your question/response is technically clogging this thread

You didn't mention which physics parameters you tweaked to get "realistic" handling with "unrealistic" values... unless you just mean you're running -3.5 camber instead of -2.0 ingame.


If you want to change the ideal camber angle of the tires and the size of the performance envelope you should be adjusting DCAMBER0 and DCAMBER1, not the chassis geometry.

 

[RC45]

You didn't mention which physics parameters you tweaked to get "realistic" handling with "unrealistic" values... unless you just mean you're running -3.5 camber instead of -2.0 ingame.


If you want to change the ideal camber angle of the tires and the size of the performance envelope you should be adjusting DCAMBER0 and DCAMBER1, not the chassis geometry.

Unless the vehicle you are working with is deigned to do just that. Changing camber on a Corvette chassis actually involves moving the chassis pickup point.

But we are clogging this thread - where is the RD AC physics discussion section?

 

[David Dominguez]

.

 

[RC45]

Thanks for this.

You didn't mention which physics parameters you tweaked to get "realistic" handling with "unrealistic" values... unless you just mean you're running -3.5 camber instead of -2.0 ingame.


If you want to change the ideal camber angle of the tires and the size of the performance envelope you should be adjusting DCAMBER0 and DCAMBER1, not the chassis geometry.

Unless the vehicle you are working with is deigned to do just that. Changing camber on a Corvette chassis actually involves moving the chassis pickup point.

The point being that I looked at the C7 Stingray suspension.ini that Kunos have and they have the rear tie rods and front steering rods in the incorrect place - I believe they have done so that they can manipulate camber values without moving the suspension pickup points.

Rear lower a-arm camber eccentrics circled in red (you normally replace with fixed plates)

Front lower a-arm camber eccentrics circled in red (you normally replace with fixed plates)

Front upper a-arm pickup points circled in red - these will be shimmed to move the upper a-arm out in conjunction with lower camber plates to get ideal static camber - still using wheels centered over the steering pivot point.

Rear fixed upper a-arm pickup points circled in red, both behind the rear axle line parallel to the ground. Lower rear also fixed but on an angled plane tot he ground with the forward adjustable pickup above it.

 

[David Dominguez]

You didn't answer my question. What have you done with the tyres?

 

[RC45]

You didn't answer my question. What have you done with the tyres?

| am using Kunos V7 tires set to the correct wheel and tire sizes. Per Kunos own statements that tires can be used complete one car to another.

I just happened to choose the C7 Stingray tires as the compound is the same as what we would use on the C5 and C7.

But the tire model is not the problem. Being that some folks literally take the entire C7 wheel and tire and bolt it on the their C5 or C6 with certain predictable results.

I have been running C6 Z06/Grand Sport wheels and tires on my car successfully for years now - in compounds ranging from Toyo R888/Michelin PSC2 all the ay to Hoosier A7's.

It is my understanding that the suspension pickup points used the suspension file are in relation to the wheel center - is this correct?

 

[mclarenf1papa]

Thanks for this.


Unless the vehicle you are working with is deigned to do just that. Changing camber on a Corvette chassis actually involves moving the chassis pickup point.

This statement alone proves the issue with your car....you. If you don't know the function of DCAMBER you can't claim Kunos is using fake values to provide real behavior, because you don't know what those values represent.

Stereo was saying that instead of changing the car's actual camber measurement (i.e. leave the suspension completely unchanged from the measurements you took), that you can change the tires' optimal camber to provide the "correct" feel even with unchanged suspension. Since every tire is different and has different characteristics, this isn't an unrealistic change to make. To adjust the Camber/Grip relationship, you adjust DCAMBER 0 and 1. This formula can allow you to visualize changes (it will give you a graph of the curve...IIRC y axis is a grip multiplier, x axis is camber in radians): 1+(x*DCAMBER_0+(x^2)*DCAMBER_1)

| am using Kunos V7 tires set to the correct wheel and tire sizes. Per Kunos own statements that tires can be used complete one car to another.

I just happened to choose the C7 Stingray tires as the compound is the same as what we would use on the C5 and C7.

But the tire model is not the problem.

It is my understanding that the suspension pickup points used the suspension file are in relation to the wheel center - is this correct?

So evidently it is... (because if the size is different that means the construction is different and therefore every characteristic of the tire, regardless of compound, will be different in some way)

And yes suspension points are based on the wheel center.

 

[David Dominguez]

Well, maybe that's one of your issues, thinking that using whatever Kunos made will work straight into your car...