14 replies to this topic

[BrunoB]

I have to admit that Im pretty ashamed that I have until now not discovered (or realized) that NR2003 has a severe bug in the modeling of the FrontSway/FrontAntiRoll.
But since nobody else in the community has cried out then maybe I instead should be a bit proud:-)
And when Im saying the bug is in NR2003 I have only researched how the bug is unveiling itself when you use the popular GTP mod.
Actually I have never been interrested in the original Nascar 2003 sim. Only in the exciting GTP mod.

Anyway. While creating my manual "BrunoBs new approach to racing setup" I mainly used my sim experience from this GTP mod.
And because the only needed driver ability in this new approach is to destinguish between under/oversteer then its ofcourse pretty urgent to be able to "feel" if the sim is under/oversteering as an answer to a change in the setup parameters.
Note: The manual dissapeared with the Blackhole site but a short discussion: https://srmz.racesimcentral.net/inde...?showtopic=5198
Some people have actually PM´ed me about the manual so Im looking for a re-release on a new site.

Cause.
I think Papyrus when they created NR2003 has simplified the modeling of Front and RearSway so much that they have merged Front&RearSway to 1 parameter which is only able to raise or reduce the REAR grip! Nothing is happening with the Front grip!
To make it clearer I think we can focus only on the FrontSway parameter, because changing RearSway dont really matter in NR2003/GTP.
If you change the RearSway (from say min to max value) then you will only feel that something (very subtle) has happened. Its quite impossible to feel "what" has happened. Has the outcome given more or less Rear grip?
Dont let the psychology (anti-placebo) fool you. Like you know what car behavior change SHOULD happen and because something (subtle) feels different then it MUST be the suspected outcome.

Problem.
OK. Back to the Front Sway, and my proof of the pudding.
Take your most stable GTP (or NR2003?) setup and set the RearSway to some middle value (say 0.775).
Then set the FrontSway to the softest/minimum value (1.000).

Test the setup on a drifting friendly track. The "3_forty_9 townie" from WindyCityRacing is highly recommendable. I love it:-)
Depending on the rest of your setup (the spoiler!) the car will probably be very loose and Oversteering (my postulate:) because of low REAR grip.
And now to the proof.
Go into pit. And change the FrontSway to the hardest/maximum value (1.938).
Go out and test it.
The car will now be VEEERY more Understeering and tight. And this time it is easy to feel that its BECAUSE of much higher REAR grip.

But this modeling of changing a FRONT value is plain wrong!
Its OK that a car with min Front Sway should be OVERsteering, and
its OK that a car with max Front Sway should be UNDERsteering.
But its plain wrong that you can feel that Papyrus has modeled this change in car behaviour by altering the REAR GRIP.
Because WE HAVENT CHANGED ANYTHING IN THE REAR.

Its my oppinion that everybody except the most green NR2003/GTP rookies can easily feel that this actual grip-change is in the REAR.
The change is not in the front as it should be because we have only changed something in the front.
BrunoB´s Racelab

[Arturo Pereira]

HI Bruno

Quite interesting finding for sure !! About what you posted at the Blackhole, feel free to post it here. I will be surre to make the thread sticky

I have never been good with setups, but about the front/rear sway bar I always thought they worked as it would be logical.

Most of the time, with the GTPs, I rarely changed the front ARB. I delt with over/understeer by reducing the rear ARB AND by reducing bump/rebound. It would be interesting to test all this stuff, so I guess I have homework to do

[John Woods]

Well, okay! GTP being my second favorite sim this is interesting for sure. Here is my monkey wrench, in case anyone wants to consider how it might apply in the context of this thread or GTP or GPL. I think Niki is making the point in example 3 that if front ARB and camber settings are out of sync everything else goes out the window. No implication about cause of BrunoB's observation, but noted to assure it is considered and eliminated as the case. <BR><BR>Edit: Did enough testing to really confuse myself completely. Increased rear grip effect is noted when following BB's example, but seems oversteer with lower front bar setting can be overcome by adjusting other factors such as springs, camber, weight distribution, etc.<BR>Pop-up instructions in garage menu note "increasing front bar tightens car," btw.

Attached Thumbnails

• 

Edited by John Woods, Dec 15 2010 - 11:04 PM.

[fortine_oo]

BrunoB

Unlike an oversteer where the car swaps ends, there's not much to feel when the car is pushing, it just doesn't turn.
The front is subtle, the rear is just more dramatic.
Your description of the lack of "front feel" is what happens if the caster value is under 3, and especially if it's under 2.

[BrunoB]

TvO, on Dec 15 2010 - 08:09 PM, said:

Hi,

I've never tried NR2003 myself, but isn't it possible that putting the Front Sway into extreme values will result in the front losing grip? Percentage wise, the rear will become more grippy but in absolute terms the rear grip stays the same, right?

Tom.

Hey Tom
Good argumentation. But
1. The GTP use a lot of reduction in the different min/max setup values compared to stock NR2003. So min/max SwayBar isnt that extreme.
2. If people follow my simple test then (postulate:) evebody can actually feel that the change in grip HAPPENS IN THE REAR. In absolute terms not only relatively/percentage.

But after I started this discussion I found this in the "Redline GTP Setup Guide":

"In n2003, there is ***no*** modeling of the loss of grip from a stiffer sway bar.

For some reason this fact seems to go unknown among the community, and many fast drivers and well known "experts" have sworn by ultra-soft sway bars, believing incorrectly that this gives them increased lateral grip, which is simply not true".

[BrunoB]

Conserning John Woods quote from the Niki Lauda book.
I cant use it to anything at all.
Lauda is refering to real GP racing and Real cars.
And we are discussing a kind of "simulation" of real cars.
And eventhough we are discussing one of the best simulations (IMO) it has nothing (much) to do with the real world.Or rather you cant deduce much from real world car setup because there are so many shortcommings and simplifyings that even the best socalled "sim" should be happy if it is able to give the human driver a kind of experience that vaguely can be related to anything "real".
People with another oppinion should try to count in the above mentioned Setup Guide how many times the Redline authors are refering to some physics force or setup parameter that is not simulated in the NR2003/GTP "sim".
And I think the Redline group does know at least something about the NR2003/GTP physics engine:-)

By the way. Just an inspiration to you folks.
If you are able to cope with (use positively) this SwayBar/AntiRollBar bug:-) then you will be able to make some veeeery grippy setups.
But as I mentions in the discussion about "BrunoB´s New Setup Approach" this kind of setups demands a driving style that feels more like rallye.

[John Woods]

As noted, I tried BrunoB's front bar experiment and found the same result. So I agree about the effect seeming to be more, actually a lot more, rear grip. One reason for posting the quote was example 3 referred to changing a front value to effect a change normally made on the rear, and that changing the rear as is typical would only worsen the already poor effect.

Anyone reading V. R. Metaphysical Aesthetics knows my opinion about real vs. virtual, or analog vs. digital, is close enough to BrunoB's to be considered identical, (at least so far, I think so anyway), but, we don't know, (do we?), what the GTP, or GPL, models really look like or have in them, only that they are alleged simulations of some kind of virtual replica's portrayal of something real/analog, that we do know something about.

BrunoB, I second Arturo's recommendation. Post here when ready, please.

[BrunoB]

John Woods, on Dec 16 2010 - 12:51 PM, said:

As noted, I tried BrunoB's front bar experiment and found the same result. So I agree about the effect seeming to be more, actually a lot more, rear grip. One reason for posting the quote was example 3 referred to changing a front value to effect a change normally made on the rear, and that changing the rear as is typical would only worsen the already poor effect.

Anyone reading V. R. Metaphysical Aesthetics knows my opinion about real vs. virtual, or analog vs. digital, is close enough to BrunoB's to be considered identical, (at least so far, I think so anyway), but, we don't know, (do we?), what the GTP, or GPL, models really look like or have in them, only that they are alleged simulations of some kind of virtual replica's portrayal of something real/analog, that we do know something about.

BrunoB, I second Arturo's recommendation. Post here when ready, please.

You are right. We dont know whats inside "the machine".
I have ealier been on the track myself that I deduced sim car behaviour from how a racing car was responsing in the real world.
Most often with some success. But thats only because creators of better sims (like Papyrus) behind all kind of shortcommings and simplifyings actually tries to mimick/simulate at least some of the real cars responses.

But then I realized that there are NO wheels and springs and dampers in ANY sim.
There are only some programmed analogies and algoritms often called "the physic car model" or something. Which in a more or (often) less sophisticated way tries to make the car respond in a predictable way.
The most obvious area to prove that even the better simulations has a lot of shortcommings and even quite false responses is in the drifting/sliding area. In short the tyre slip-model.
One of the best and most sencere discussions (conserning rFactor) has now disapeared with the original RSC forum but it was called the Drifting Issue. And I think it was a bloke from Holland (Niels something?) with a lot of insight in precisely this area (even with some inside info of REAL data from tyre producers) who was able to argument rather understandably (I was only able to follow the mathematic argumentation some way:-) WHY THE TYRE MODEL DIDNT WORK.
And eventhough the drifting ability in NR2003/GTP is a bit better (IMO) then its still far far away from realism.
I havent tried the iRacing sim but it sounds like this NR2003-2.0´s drifting behavior has a strange familarity with its predecessor.
Try to ask even an iRacing fanboy how the cars behave in the grass!
I dont want to mention GPL in this area because I want to stay polite:-)
----
By the way. Conserning my setup manual "BrunoBs new approach to racing setup" then I tried to Upload it (zip) to the Download section in this forum.
But there are only 2 categories: GPL and P&G.
And I fancy GTP(NR2003) so much that eventhough the manual is "generic" (works for all sims) then it dont belong to GPL or P&G ;-)
BrunoB

[Arturo Pereira]

Hi Bruno

Do you mean you could not upload it here? Anyway, you send me the manual with a PM and I will post it at this forum.

[BrunoB]

Arturo Pereira, on Dec 17 2010 - 07:20 AM, said:

Hi Bruno

Do you mean you could not upload it here? Anyway, you send me the manual with a PM and I will post it at this forum.

PM sent Arturo
When you have placed the manual in the forum I will change the dead Blackhole links round at the net to the new adress. OK.
Kind regards BrunoB

[BrunoB]

Hello Arturo
The download for the "Brunob´s New Approach To Racing Setup" is working perfectly.
https://srmz.racesimcentral.net/inde...t=0

Thank you very much for your help
And happy Christmas/New Year.

Edited by BrunoB, Dec 25 2010 - 06:16 AM.

[Lee200]

Sorry to be late to this thread, but I just discovered it as I don't have NR2003.  I do know a bit about GPL's physics and AFAIK, Kaemmer used GPL's physics engine for NR2003.

In GPL, both the front and rear roll bars are operational although I don't think the physics properly models how real world roll bars work.

It is my understanding that NASCAR cars use a beam axle at the rear end with a Panard rod to locate the axle laterally.  There is no roll bar for the rear as the two opposing wheels are already rigidly tied to each other by the solid axle.

Perhaps, NR2003 simulates this by not applying the rear roll bar setting?  

Lee

[BrunoB]

Yeah. Perhaps, Perhaps

[jgf]

GPL was the basis for Nascar Racing 4.  NR2003 was all new.