5 replies to this topic

[BrunoB]

Asymmetric cars - fundamental bug in NR2003 and all its mods.

To the best of my knowledge no NR2003 setup wizard have acknowledged (or discovered) that there is a hardcoded bug burried deep inside the NR2003 game engine.
The cars are fundamentally asymmetric in the horizontal plane!

What does that mean?
It means that if you create a 100% symmetric setup then the cars behaviour is different if you want it to turn left or right!
Because the grip values is different in leftward corners than in rightwards!

The reason is surely because the orig NR2003 was created to be driven on ovals.
And on ovals you only turn to the left.

But if we compare this to the real world then no chassis of real Nascars (or Indycars for that matter) are made fundamentally asymmetric.
Its only the special asymmetric oval setups for these cars which makes their behaviour asymmetric.

OK I can hear you say that this is no big deal for this great sim and all its mods.
Not much difference if the asymmetry is in the car itself or introduced in the setups (as in real cars).
But the problem explodes when we are taking our virtual cars away from ovals and place them on a road track with rightward corners.
Because then we have to create an assymetric setup to try to make the car behaving (kind of) symmetric!

I admit this is not really a problem for a rookie driver or a more casual driver because they probably cannot feel the difference between the "normal" car control difficulties and this general behaviour issue.
But why have no of the more experienced drivers discovered and targeted this issue?
Probably because they have driving skills which are able to compensate for the problem.
But at least in my mind there is no doubt that even aliens would benefit of implementing changes caused by this discovery.

You can download the full article as a Word document here:
https://dl.dropbox.co...g in NR2003.doc

BrunoB Racelab

Attached Files

•  Assymetric cars - fundamental bug in NR2003.doc   562K   1032 downloads

[jgf]

BrunoB, on Apr 12 2012 - 04:14 AM, said:

...
But if we compare this to the real world then no chassis of real Nascars (or Indycars for that matter) are made fundamentally asymmetric.
Its only the special asymmetric oval setups for these cars which makes their behaviour asymmetric.

...

I don't know when Indycar abandoned the visibly asymmetric suspensions, but NASCARE definitely still uses them.  Larger, more affluent, teams build specific cars for short tracks, speedways, superspeedways, etc.;  the chassis are certainly similar but definitely neither identical nor symmetric.

From https://www.autocentr...hassis-For-0001

"While tight regulation applies to most parts of the vehicle chassis, the suspension system represents a relatively unrestricted zone that leaves plenty of opportunity to engineer inherently faster race cars. After establishing subtle compromises between vehicle aerodynamics and handling performance, detailed suspension design and tuning begins. Chris Hussey explains that consistent tire patch contact with the track can be maintained under all racing circumstances, by adapting suspension characteristics to align the driving behavior of the car to the size and banking of the oval track. "Besides focusing on designing for lighter vehicle weight, we modify the suspension design, all within the design boundaries imposed by NASCAR. We tune caster and camber in the asymmetric suspension along with toe, and design the right amount of toe out that can be gained on the inside tire in a turn. Following our strategy to bringing more engineering in-house, we take full control over chassis/suspension engineering and safeguard our technical expertise."

If you perceive a difference in grip for left hand vs right hand turns in the sim with a neutral setup I would first suspect a software issue.  Also consider a psychological/physiological issue;  being right handed, I find it easier to make left turns than right ones ...in real life and in sims.  A quick test in a sim is to create a completely symmetric, neutral setup and do a few medium speed laps at a speedway;  just warm up the tires, don't try for hotlaps, and note the tire temps.  Start over and do equal laps in the opposite direction, again noting tire temps.  If the chassis physics are modeled symmetrically, and your driving is consistent, tire temps will be congruent, though reversed, for both runs.

I do feel there are physics issues with the GTP mod since these cars should be significantly faster on a superspeedway than a Cup car, yet they are invariably slower.

[BrunoB]

First my postulate that real oval cars are not created/build asymmetric.
I think its a matter of choosing the words because as I clearly points out in the article the setups for real oval cars ARE (ofcourse) asymmetric.
But maybe I have not taken the thought fully to its end because as I understand your quotation the asymmetry is more drastic than giving the camber, caster, pressure and springs asymmetric values (as was my thought) but also changing/redesigning some swingarms and stuff. Thats OK with me. It just shows that they are taking the setup settings further asymmetric than I thought:-)

But then to this other one where I to the contrary of the above think that you havent completely got my point (hard but nice unfair postulate :-):-)

jgf, on Apr 12 2012 - 08:58 AM, said:

If you perceive a difference in grip for left hand vs right hand turns in the sim with a neutral setup I would first suspect a software issue.

Because the only reason I use so much time "translating" the Papyrus help text (and calling it from the mouth of the horse itself:-) is solely because this is the closest of a proof you can get!
Because if the NR2003 cars were not fundamentally asymmetric there would not be any asymetric difference in gripchange from low to high speed corners.
Exactly how it is in the rear!
This fundamental asymmetry has absolutely nothing to do with "a software issue" or a "psychological/physiological issue" (like kinda placebo effect).
Or formulated rhetorical: why do you think Papyrus need to make this strange asymmetric distinction between how the front grip changes and the rear grip does?

Please dont take this as Im patronizing you. Because Im pretty satisfied that at least you give some feedback/questions my findings.

[jgf]

Dredging up an old topic here but I've been digging through my old N2003 stuff recently (it was too cold to go outside and enjoy winter this year).  In the track.ini are settings for "track type" and "chassis type" which seem to reference hard-coded "suspension" (handling) parameters;  all the oval settings apparently replicate asymmetric suspensions (stagger?  this was a setup option in early versions of Nascar Racing), the "road course" option probably being the only one capable of true symmetric setups.

I haven't experimented with these much, beyond removing restrictor plates, and there is some controversy over what exactly these settings control.  The following are from my notes (sources unknown ...long forgotten):
----------------------
"You can try any chassis type at any track by putting in one of the numbers below.

1 = superspeedway with restrictor plate
2 = speedway
3 = short track
4 = road course
5 = superspeedway without restrictor plate"
-------------------------
"0 = short track
1 = mile oval
2 = medium oval (1-2 mile)
3 = long oval (think this is the 2 mile)
4 = super speedway
5 = road course

Chassis types:
0 = ?
1 = super speedway
2 = speedway
3 = short track
4 = road course

From testing at Talladega -

Chassis 0 - high drag, huge torque, tires never wear out, top speed around 190.
Chassis 1 - low drag, low HP, fixed rear springs, shocks and spoiler, top speed around 185.
Chassis 2 - medium drag, top speed around 210.
Chassis 3 - medium drag, top speed around 205.
Chassis 4 - medium drag, top speed around 205.
Chassis 5 - minimal drag, doesn't like to turn or stop, top speed around 240."
-------------------------

From those Talladega results I would think Chassis 0 = short track, Chassis 1 = restrictor plate, Chassis 5 = superspeedway with no plate; the others - ???, too little data.

As you can see there is some confusion just between those two quotes, and I've no idea whether the track or chassis parameter, or some arcane combination thereof, tells the sim which physics modifications to apply.  (For example, what if you designate Daytona as a medium oval and select the road course chassis?)

And how much, if any, of this applies to the GTP mod is anyone's guess.  Redline modified the physics so greatly that all this may be entirely irrelevant.

[ZiggyM]

jgf, on Mar 14 2014 - 08:27 PM, said:

(sources unknown ...long forgotten):

----------------------
"You can try any chassis type at any track by putting in one of the numbers below.

1 = superspeedway with restrictor plate
2 = speedway
3 = short track
4 = road course
5 = superspeedway without restrictor plate"
-------------------------
"0 = short track
1 = mile oval
2 = medium oval (1-2 mile)
3 = long oval (think this is the 2 mile)
4 = super speedway
5 = road course

Chassis types:
0 = ?
1 = super speedway
2 = speedway
3 = short track
4 = road course

From testing at Talladega -

Chassis 0 - high drag, huge torque, tires never wear out, top speed around 190.
Chassis 1 - low drag, low HP, fixed rear springs, shocks and spoiler, top speed around 185.
Chassis 2 - medium drag, top speed around 210.
Chassis 3 - medium drag, top speed around 205.
Chassis 4 - medium drag, top speed around 205.
Chassis 5 - minimal drag, doesn't like to turn or stop, top speed around 240."
-------------------------

Those are my words  :-)

At the time, I was referring to the chassis by the 'speedway' or 'short track' names but after looking and testing, they would be more correctly called the 'banked' or 'flat track' chassis since chassis 2 is used the banked tracks, including Bristol, while chassis 3 is used at Pocono, Indy, Homestead (original), Martinsville and Richmond.  Track length doesn't matter so I no longer call chassis 2 the 'speedway' chassis.  :-)

I've done more testing since that write up.  Your results may vary.

CORRECTION and UPDATE 02/12/15! - I have changed the track# (at an oval) and saw no difference in anything so I'm thinking it might just be used to identify the track so the AI knows which ratings to use and also for the 'officials' to know how long to keep the yellows out?  I have no idea how it's determined if the yellow will fly or not but I have been at short tracks, had a big mess and no yellow, while at a bigger track, somebody spins into the infield and the yellow flies.  It might determine that but checking yellows is difficult and very hard to get people to join in for.

I changed the track# at a road course and the car pulled to the left!  Also, the tires wanted less psi on the left side.  The car looked to be sitting at an angle, right side lower than the left, from the cockpit. This is the first time I tested with the track # 0 at a road course and the first time I encountered the pull to the left at a road course.  With every chassis and the track set to 5 (road course) I never felt the car pull left.


I haven't tested offline races and know smaller tracks have more caution laps, 4 or 5, while bigger tracks have 2 or 3 and road courses give the "we're going green" after 1 whether there are lapped cars of not.

I haven't seen a difference in tire pressures, handling, tire wear or fuel mileage when changing the track # so I have to think it's something that doesn't affect the car.

As for the pull to the left....  Maybe the engine torque makes it pull left?  I haven't tried testing that yet.

Test results are below.  Top speed and HP were fro the NR2003 Replay Analyzer and drag is how many seconds the car too to coast from 170 to 140 MHP on the back stretch of Talladega.  I come off T2 and hit the clutch and let it coast, then check the replay for the time and speed.  Higher numbers are 'good', lower numbers are 'bad'.
I have yet to check the body height to try to guestimate the amount of downforce each chassis provides but since chassis 0 has the most drag and is the easiest to recover (nearly impossible to spin that thing) and chassis 5 has the least drag and will snap around with little effort, I'm pretty sure the drag is caused by downforce.

Test track - Talladega, clear skies, 70 degrees F and no wind.  Top speeds in MPH, drag number rounded to nearest second.

Chassis 0 - Mystery car - Huge torque, huge drag, likes the psi 42 all the way around at all tracks, tires never wear out unless you have pits set to 4x and slam the brakes and hold them there.  Then you will get them to wear.  Idles at 800 RPMs while the other chassis idle at 1000.  Get really bad fuel mileage.  Has the same HP for all chassis.
Cup - 745 HP, top speed 189, 5 drag
GNS - 745 HP, top speed  187, 5 drag
CTS - 745 HP,  top speed180, 7 drag
PTA - 745 HP, top speed 192, 5 drag

Chassis 1 - Restricted Aero - low drag but fixed spoiler, fixed rear shocks, fixed rear springs, left side track bar can not be higher than the right side. Likes the psi at 28/48 at ovals, 48 all around at road courses - stable at high speeds, low grip at low speeds.
Cup - 370 HP, top speed 189, 9 drag
GNS - 342 HP, top speed 181, 9 drag
CTS - 550 HP, top speed 181, 6 drag
PTA - 625 HP, top speed 212, 7 drag

Chassis 2 - Banked track - Medium drag, downforce on the nose, rear feels a bit light, likes the psi at 28/48 at ovals, 48/48 at road courses, stable but a bit 'twitchy', sensitive to steering input -
Cup - 635 HP, top speed 208, 7 drag
GNS - 623 HP, top speed 206, 8 drag
CTS - 562 HP, top speed 186, 6 drag
PTA - 625 HP, top speed 210, 6 drag

Chassis 3 - Flat track - Medium drag, even downforce, maybe more side force, likes the psi at 22/34 at ovals, 34/34 at road courses, stable, easy to recover -
Cup - 637 HP, top speed 203, 7 drag
GNS - 622 HP, top speed 202, 7 drag
CTS - 562 HP, top speed 185, 6 drag
PTA - 625 HP, top speed 207, 6 drag

Chassis 4 - Road course - Medium drag, even downforce, maybe more side force, likes the psi at 22/34 at ovals, 22/22 at road courses, stable, easy to recover -
Cup - 640 HP, top speed 204, 7 drag
GNS - 622 HP, top speed 203, 7 drag
CTS - 562 HP, top speed 185, 7 drag
PTA - 625 HP, top speed 207, 8 drag

Chassis 5 - Unrestricted Aero - Low drag, stable at high speeds, low grip at low speeds, likes the psi at 28/48 at ovals, 48/48 at road courses, left side track bar can not be higher than the right -
Cup - 706 HP, top speed 232, 11drag
GNS - 622 HP, top speed 226, 11rag
CTS - 574 HP, top speed 190, 7 drag
PTA - 625 HP, top speed 212, 6 drag

PTA car appears to be least effected by the chassis # although I have to admit I have the least experience in it and it took many attempts to get any data because I'd wreck or nearly wreck every lap. Just not at all used to it.  :-(

Chassis # affects when the tires go yellow and red as well.  Chassis 4 tires go yellow at 220 and red at 240. I think chassis 1/5 have to be at 240 to go yellowand 270 to go red I can't recall and haven't recorded those numbers yet.

So, those are the numbers I've come up with.  I hope more people share their results!  

Edited by ZiggyM, Feb 13 2015 - 12:55 AM.

[Ziggy Moonglow]

After looking into the "NR2003 Physics Editor" and playing with it a bit, I'm pretty sure chassis 0 is the pace car.  I noticed that no matter which physics are being used (Cup, GNS, CTS or PTA) and no matter which chassis is assigned by the track, the game is always loading a car with chassis 0 so it's got to be the pace car.  And the way the pace car goes from 0 to 100 in no time, all that torque makes sense, along with no tire wear.  When it goes back to it's stall, it probably gets filled up with fuel, but I've never made it stay out forever (which could be done with the "!yellow" command every couple of laps) to see if it ever runs out of fuel.