FF tunes don't make any sense

[Scaff]

Greyout mentions that increasing the stiffness of springs at the rear increases weight transfer in the rear.

I said that incresing the stiffness in the front increases weight transfer in the front, but you said that was incorrect. I then read what you told me too and greyout says the same thing about the rear.

So I was correct in the first place right?

Yep - my bad on that bit (serves me right for posting at one in the morning), I was thinking in regard to lateral transfer not longituduinal transfer.

The tires arent very clear (This photos was not taken for this purpose), but there is no droop in the wheels whatsoever, they appear to be perfectly plumb, zero camber.

I dont know if thats because the Impreza has a different suspension to the Racnge Rover.

Would the Impreza's wheels droop in mid-air?

The Impreza would not be an ideal candidate for seeing it, running with a minimum of unsprung weight and well sorted suspension. Its why the Range Rover is such a good candidate, heavy wheels and tyres result in a very high unsprung weight (to drag the suspension down when airborne) and soft suspension (which will minimise it). Certainly GT5 does have some better candidates to look at (nice picture however).



Scaff

 

[JackWilson]

Yep - my bad on that bit (serves me right for posting at one in the morning), I was thinking in regard to lateral transfer not longituduinal transfer.

No Problem

The Impreza would not be an ideal candidate for seeing it, running with a minimum of unsprung weight and well sorted suspension. Its why the Range Rover is such a good candidate, heavy wheels and tyres result in a very high unsprung weight (to drag the suspension down when airborne) and soft suspension (which will minimise it). Certainly GT5 does have some better candidates to look at (nice picture however).
Scaff

I didnt think it would be as good a candidate as the Range Rover, but I knew I had a picture of an Impreza mid-air and didnt think the wheels drooped.

There's a Range Rover Concept in game and also one of the Pajero Rally Raid car, I'm gonna have a look at them.

 

[shirakawaa]

You should check with a Fiat 500 (old). Its rear suspensions architecture make camber become significantly positive upon unloading in real life:

 

[JackWilson]

You should check with a Fiat 500 (old). Its rear suspensions architecture make camber become significantly positive upon unloading in real life:

What type of suspension does that have?

 

[shirakawaa]

Front: double wishbone with transverse leaf spring
Rear: semi-trailing arm with coil springs

By the way, cars with swing axle rear suspensions (an archaic independent suspension architecture) should also suffer from this problem (that is, if they're correctly modeled in GT5). Example: Mercedes 300SL, VW Beetle, etc. But seeing how well these cars handle in the game, I doubt it's the case.

 

[JackWilson]

Stock Suspension. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ FC Suspension (Max Ride Height, Minimum Spring Rate)

Stock Suspension. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ FC Suspension (Max Ride Height, Minimum Spring Rate)



As you can see there is no droop or camber change at all, in the stock suspension or the FC suspension.

I tried the Range Rover Stormer Concept too with stock suspension and the wheels, when in the air, were clearly at 0 camber, completely parrallel with each other.

 

[Greyout]

Not to stick up for PD too much, but keep in mind that visual affects programing does't always match physics engine programming.

On the other hand, I am quite sure they didn't get that right

 

[SolidSnake7735]

As Kintaro of Golden Boy anime would say. "Soo educational "

 

[che69velle]

For myself one of the clearest examples is the Shelby Cobra.

The combination of high profile, thin tyres, leaf spring rear suspension and light weight/high torque, should mean that a full throttle launch should see the rear of the car step out of line to a massive degree. You should have to correct it or risk a spin.

Instead what happens is you simply get wheel spin and the car launches in a straight line, no correction required.

Scaff

The cobra has a jaguar independant rear suspension not a leaf sprung solid axel

The car model's are coded for the tire size they were scaned at when PD computerized them, case in point the shelby gt350r. The tires are narrow profile and low grip and even changing to slicks don't improve it. then there's cars that have loopholes like the m3 csl, it has a soft compound tire from factory so pd coded it to have more grip, which results in it having more grip than it should on other tires.

 

[chuyler1]

swing axle rear suspension? LOL, that car has no rear axle!!!!

 

[shirakawaa]

swing axle rear suspension? LOL, that car has no rear axle!!!!

To be fair, the old Fiat 500 although camber becomes significantly positive on suspension unloading, doesn't have swingle axle rear suspensions. The Beetle (recently included in GT5) is supposed to, however.

http://en.wikipedia.org/wiki/Dante_Giacosa

[...]He overcame the stability problems associated with other rear-engined designs by identifying that the problem was not the weight distribution of the cars, but the simple swing axle rear suspension used in those designs. His answer was to use a semi-trailing arm type of rear suspension that eliminated the large change in the camber of the rear wheels that was inherent with the simple swing axle suspension system.

This paragraph is for the Fiat 600, but the Fiat 500 had the same suspension design.

 

[Scaff]

The cobra has a jaguar independant rear suspension not a leaf sprung solid axel

Coil Sprung live axle it would actually appear, the earlier AC models did run leaf springs, I'm not aware of any '60s Cobras running independent rear suspension however.

http://www.carfolio.com/specifications/models/car/?car=4950

It also doesn't change that the car in GT5 is not behaving as it should do under a full throttle launch.

The car model's are coded for the tire size they were scaned at when PD computerized them, case in point the shelby gt350r. The tires are narrow profile and low grip and even changing to slicks don't improve it. then there's cars that have loopholes like the m3 csl, it has a soft compound tire from factory so pd coded it to have more grip, which results in it having more grip than it should on other tires.

Which I'm sure you would agree is an issue, as all changing compound does is increase lateral-g by around +0.2, and it does it regardless of any other difference between cars.

The end result is wildly different cars (with different suspension layouts, CoG, PMI, etc) having exactly the same later-g figure for the same tyre compound.


Scaff

 

[Rotary Junkie]

Coil Sprung live axle it would actually appear, the earlier AC models did run leaf springs, I'm not aware of any '60s Cobras running independent rear suspension however.

Wellllllllll... It appears the 289 cars at least ran independent rear suspension with transverse leaf springs (as did the AC Ace).

And the 427s ran IRS with coils.

At least this is my understanding from a quick dig around.

As for the lateral G remark... The difference may be the same between cars but different cars definitely have different levels of grip for a given tire.

 

[Scaff]

Wellllllllll... It appears the 289 cars at least ran independent rear suspension with transverse leaf springs (as did the AC Ace).

And the 427s ran IRS with coils.

At least this is my understanding from a quick dig around.

Well the web would appear to be a minefield of contradictory info on the subject in that case.

The only 427's I've seen info on running IRS have been continuation models (the specs for which are as wide and varied and the companies that have made them over the years)

As for the lateral G remark... The difference may be the same between cars but different cars definitely have different levels of grip for a given tire.

A very large body of testing done here at GTP doesn't reflect that at all...

https://www.gtplanet.net/forum/showthread.php?t=160821

...with an extreme example a MINI Copper S '07 producing the same lateral-g figure on CM tyres as a McLaren F1 and a Merc CL600 '00 and an Audi S4 '03. All give 0.86g on CM tyres.

Then you pop a set of RH tyres onto just about any car and you get 1.3g.

Yes some cars do give slightly different lateral-g for the same tyre (unless its RH), but they are the exception.

Do you honestly believe that a MINI Cooper s will produce the same lateral-g as an F1 if they have the same compound of tyre? Despite having wildly different CoG, PMI, tyre sizes, etc, etc.


Scaff

 

[oopssorryy]

Again, I do think it's possible to tune with real world methods, it's just much harder to do so. The key I've found so far is Damper settings. Turn in is always great with real world like settings, but exit is no so great. Toe helps, but having low front compression and high rear extension really helps keep the weight on the front tires, and there by, keep front grip and the rear loose. My Celica is running about a second faster than my old set up, I have a TRD Celica that's running .5 seconds faster than it's previous set up. I also have a Corrola Levin BZ-R '98 that's running right with the '99 Celica.

All times on Eiger Nordwand Short Track, all aids off (ABS @1), and I am a below average driver (subtract around 2-4 seconds off my times to get around what many of you make, found this out during the shootouts)

'99 Celica: 1.13.329@450pp
'98 Levin BR-R: 1.13.328@450pp
'02 TRD Celica: 1.11.717@488pp

Sorry if this is off topic, but I believe it is possible to do. Are the physics flawed? Yes. Are they so bad they can't be reasoned with? Not at all.

 

[CSLACR]

Well the web would appear to be a minefield of contradictory info on the subject in that case.

The only 427's I've seen info on running IRS have been continuation models (the specs for which are as wide and varied and the companies that have made them over the years)




A very large body of testing done here at GTP doesn't reflect that at all...

https://www.gtplanet.net/forum/showthread.php?t=160821

...with an extreme example a MINI Copper S '07 producing the same lateral-g figure on CM tyres as a McLaren F1 and a Merc CL600 '00 and an Audi S4 '03. All give 0.86g on CM tyres.

Then you pop a set of RH tyres onto just about any car and you get 1.3g.

Yes some cars do give slightly different lateral-g for the same tyre (unless its RH), but they are the exception.

Do you honestly believe that a MINI Cooper s will produce the same lateral-g as an F1 if they have the same compound of tyre? Despite having wildly different CoG, PMI, tyre sizes, etc, etc.


Scaff

One thing that should be noted, is that the G-meter in GT5 has not been proven accurate relating to cornering speed, at least that I know of.
I do know different cars can have wildly different grip, a quick spin in a GT500 NSX then GT-R will make that clear as day.
I actually think the g-graph is more of a feel-good thing thrown in, with no real indication of performance. But I haven't tested it tbh, because I've seen different cars have very different grip levels.

Of course the big problem is that you can't upgrade anything other then tire type, so any car that has crappy handling stock, can never really handle well. By "handling", I strictly mean the ability to go around corners quickly, not ease of use or under-over steer.

The one really important thing I don't think many people have realized, is that forward drive from spinning tires is wildly low in GT5.
Example: https://docs.google.com/spreadsheet...GQ4cUZaQVhVTEtvRGdDNjRDbTFGZ0E&hl=en_US#gid=0

Almost every car here has slow 0-60, and 1/4 mile times. Why? Because they can't get anywhere off the line, due to lack of forward traction. Almost every single one of them reaches a real-world or higher speed at the 1/4 mile mark, all are going too fast by the mile mark, and many have wildly high top speeds. Keep in mind, these cars have 0 miles, and no oil change, so they can't be made to have any less power, and they still lack traction, and many still go way too fast.
It affects FWD even more, because FWD has less traction to start, which is one reason it's so hard to get a properly exiting FF in GT5.

I'll have to check if the GTR and NSX have the same g readout while cornering, if they do, the g-meter is meaningless. (Or at least is in some cases)

Again, I do think it's possible to tune with real world methods, it's just much harder to do so. The key I've found so far is Damper settings. Turn in is always great with real world like settings, but exit is no so great. Toe helps, but having low front compression and high rear extension really helps keep the weight on the front tires, and there by, keep front grip and the rear loose. My Celica is running about a second faster than my old set up, I have a TRD Celica that's running .5 seconds faster than it's previous set up. I also have a Corrola Levin BZ-R '98 that's running right with the '99 Celica.

All times on Eiger Nordwand Short Track, all aids off (ABS @1), and I am a below average driver (subtract around 2-4 seconds off my times to get around what many of you make, found this out during the shootouts)

'99 Celica: 1.13.329@450pp
'98 Levin BR-R: 1.13.328@450pp
'02 TRD Celica: 1.11.717@488pp

Sorry if this is off topic, but I believe it is possible to do. Are the physics flawed? Yes. Are they so bad they can't be reasoned with? Not at all.

What you need to understand is what works for you, or what you find "good" isn't considered good by everyone. Obviously you can use real world principles, and some people will even like their cars more set up that way, but it's clearly not the ultimately fastest way to go, not by a long shot.

I understand I can use real world settings, but if I can go faster by making a ridiculous tune, I think that pretty clearly indicates a problem, don't you?

 

[Scaff]

One thing that should be noted, is that the G-meter in GT5 has not been proven accurate relating to cornering speed, at least that I know of.
I do know different cars can have wildly different grip, a quick spin in a GT500 NSX then GT-R will make that clear as day.
I actually think the g-graph is more of a feel-good thing thrown in, with no real indication of performance. But I haven't tested it tbh, because I've seen different cars have very different grip levels.

Of course the big problem is that you can't upgrade anything other then tire type, so any car that has crappy handling stock, can never really handle well. By "handling", I strictly mean the ability to go around corners quickly, not ease of use or under-over steer.

If the HUD G-meter had been the only thing used I would agree with you, however the datalog was also used..

We don't have an actual 200' skidpad to calculate lateral g force with in GT5, but what we do have is a g "meter" and a datalog. For the values I came up with, I created a delineated scale and taped it under the HUD g-force bar graph, and also used a scale against the datalog graph during replays as verification. The measurements were taken on the widest part of the TGTT, by turning a continuous steady-speed circle after warming the tires. Lateral g force was recorded up to the point where the car started to skid and could no longer hold the established circle. I also ran laps "on the edge" to verify the numbers, and repeated all the tests twice. (Note that I rounded the numbers to the nearest .05, due to my screen resolution).

...now of course a margin for error will always exist with these things, however I'm sure you would agree that the testing is being conducted in a reasonably repeatable manner, and the result of the RH tyres alone indicate something is not right with regard to the tyre model.

I have also tested a few of these using known radius corners and backwards calculating lateral-g from the cornering speed and on the cars I've test so far (which I acknowledge is not many) they do corroborate the body of testing done above.

The one really important thing I don't think many people have realized, is that forward drive from spinning tires is wildly low in GT5.
Example: https://docs.google.com/spreadsheet...GQ4cUZaQVhVTEtvRGdDNjRDbTFGZ0E&hl=en_US#gid=0

Almost every car here has slow 0-60, and 1/4 mile times. Why? Because they can't get anywhere off the line, due to lack of forward traction. Almost every single one of them reaches a real-world or higher speed at the 1/4 mile mark, all are going too fast by the mile mark, and many have wildly high top speeds. Keep in mind, these cars have 0 miles, and no oil change, so they can't be made to have any less power, and they still lack traction, and many still go way too fast.
It affects FWD even more, because FWD has less traction to start, which is one reason it's so hard to get a properly exiting FF in GT5.

Oh, I more that recognise it, its the main point I raised in regard to the Shelby Cobra test, the tyre model (and I suspect the suspension model) don't create any lateral movement under a full throttle launch. All that happens is the tyres spin like mad and don't grip, and you are right that its worse with FWD.

No torque steer or step out occurs at all, and I suspect the root cause of both is down to either a lack of tyre deformation or a very basic tyre deformation model. I also believe that it may be one of the reasons why rolling starts are by far and away the norm.

I'll have to check if the GTR and NSX have the same g readout while cornering, if they do, the g-meter is meaningless. (Or at least is in some cases)

Maybe not if the datalog and reversed calculated lateral-g figures corroborate the HUD g-meter.

What you need to understand is what works for you, or what you find "good" isn't considered good by everyone. Obviously you can use real world principles, and some people will even like their cars more set up that way, but it's clearly not the ultimately fastest way to go, not by a long shot.

I understand I can use real world settings, but if I can go faster by making a ridiculous tune, I think that pretty clearly indicates a problem, don't you?

100% agree.


Scaff

 

[oopssorryy]

I can see that the system is severely flawed, in the way that it is easier to make a unrealistic tune that will go fast, than it is with a real world approach. But if you put the time in I'm getting much faster times with realistic tunings. I know I should be lifting the inside rear tire and lift off understeer, but I'm not really getting either. This does show serious issues, issues that should have been dealt with before launch. What I'm trying to get at is, I used to run an odd tune, with a stiff front and soft rear on all of these cars, but am now turning faster lap times. Maybe it just works for me, but are you 100% it doesn't work for you?

The biggest thing I'm finding so far that damper settings are super important in this situation, were once they were highly debatable on if they were more for personal preference or actually really affected handling characteristics. The spring rates give you a good base, but if you can't keep the weight on the front tires coming out of a corner, it's useless. It took me hours longer than it usually would to tune a car, but once you get the right spring rates with correct supporting suspension settings, the car is much easier to deal with in and out of corners.

Again, I see that there are many fast FFs setup in strange ways, and that they should barely be able to get around a track. I know this is wrong, but for some reason it works. At the same time I don't think these set ups are getting all they can from these cars.

 

[CSLACR]

If the HUD G-meter had been the only thing used I would agree with you, however the datalog was also used..

I will check it out, for certain. Without it being tested that I know of, I don't know if the data logger even accurately records g-readings for cars.

...now of course a margin for error will always exist with these things, however I'm sure you would agree that the testing is being conducted in a reasonably repeatable manner, and the result of the RH tyres alone indicate something is not right with regard to the tyre model.

I have also tested a few of these using known radius corners and backwards calculating lateral-g from the cornering speed and on the cars I've test so far (which I acknowledge is not many) they do corroborate the body of testing done above.

My main question is if the g-meter changes appropriately according to cornering speed or not. I'm not saying it doesn't, just that I've never tested it. It wouldn't surprise me to find any terrible inaccuracies in this game anymore.

No torque steer or step out occurs at all, and I suspect the root cause of both is down to either a lack of tyre deformation or a very basic tyre deformation model. I also believe that it may be one of the reasons why rolling starts are by far and away the norm.

It might have something to do with LSD function as well.
Tire sidewall for slicks is calculated in the game as higher, but I think that is all, nothing more in-depth, I agree. Even that size increase in diameter of tires I believe is simply a small % increase, rather then any true calculation, and like all tires in GT games, is a mandatory set figure.

Maybe not if the datalog and reversed calculated lateral-g figures corroborate the HUD g-meter.

Very true, I'll be looking into it.

I can see that the system is severely flawed, in the way that it is easier to make a unrealistic tune that will go fast, than it is with a real world approach. But if you put the time in I'm getting much faster times with realistic tunings. I know I should be lifting the inside rear tire and lift off understeer, but I'm not really getting either. This does show serious issues, issues that should have been dealt with before launch. What I'm trying to get at is, I used to run an odd tune, with a stiff front and soft rear on all of these cars, but am now turning faster lap times. Maybe it just works for me, but are you 100% it doesn't work for you?

Practice makes perfect.

Maybe it just works for me, but are you 100% it doesn't work for you?

If there were a way to be more then 100% sure, that's where I'd be. You simply can't remove enough under steer (particularly offline) without it.

There are of course, cars that are exceptions, which have neutral handling at stock settings, or close to it. these cars can be worked with, and you could potentially get realistic settings running fast laps, but the bulk of the cars (90%+) need a ridiculous setup to achieve their maximum potential. And of course, the only reason these exceptions can use realistic looking settings, is because they don't need massive amounts of under-steer dialed out of them. It actually gets a lot more extreme for faster drivers then myself, I actually use more under-steer prone tunes then many drivers on my own pace, let alone the really fast guys.

 

[oopssorryy]

Hmmm, I think I understand what your getting at now. Your saying I should get massive amounts of oversteer from the settings I'm running now, yes? I will say that the Celicas handle well already so maybe it wasn't a good candidate. I also see that I should be lifting the rear inside tire under hard cornering, but I am not. The only way to really simulate the lift is to put a lower grade tire on the rear. The rear end just retains to much grip to get it to swing out. Having large amounts of toe helps but far to much makes the car unstable....

This, and other point you have all made, have really be hammered home today as I tried to tune a Corrola Matrix (not it's really name in game, but it's to long to remember). It's entry was great but it's exit was crap, and it shouldn't be. The rear end should have been slipping out. Because of this I started testing other things, and I think that the problem is that PD really doesn't want us to flip the cars. There may be some sort of aid we can't turn off that keeps all four wheels on the ground. It would help explain why it's so hard to flip a car at Trial Mountain. By keeping all tires on the ground the rear retains grip and understeers. I know this doesn't explain why reverse settings work, but this is all I can really come up with. Sorry for sounding like a thick headed idiot before and de-railing the thought process. Please don't mind my posts, and continue discussing the bugs in the game.

 

[Rotary Junkie]

Actually, oopssorryy, that's exactly the theory behind why I started doing it in GT4.

 

[CSLACR]

Hmmm, I think I understand what your getting at now. Your saying I should get massive amounts of oversteer from the settings I'm running now, yes? I will say that the Celicas handle well already so maybe it wasn't a good candidate. I also see that I should be lifting the rear inside tire under hard cornering, but I am not. The only way to really simulate the lift is to put a lower grade tire on the rear. The rear end just retains to much grip to get it to swing out. Having large amounts of toe helps but far to much makes the car unstable....

I don't know what settings you're running right now, so I don't know if it should over steer or not.
The Celica over steers on entry (kinda) and under steers on exit stock, I believe, or it's at least balanced that way somewhat.
I tuned a Celica SS-II after the FF shootout, but long lost the tune.

This, and other point you have all made, have really be hammered home today as I tried to tune a Corrola Matrix (not it's really name in game, but it's to long to remember). It's entry was great but it's exit was crap, and it shouldn't be. The rear end should have been slipping out. Because of this I started testing other things, and I think that the problem is that PD really doesn't want us to flip the cars. There may be some sort of aid we can't turn off that keeps all four wheels on the ground. It would help explain why it's so hard to flip a car at Trial Mountain. By keeping all tires on the ground the rear retains grip and understeers. I know this doesn't explain why reverse settings work, but this is all I can really come up with. Sorry for sounding like a thick headed idiot before and de-railing the thought process. Please don't mind my posts, and continue discussing the bugs in the game.

Honestly, I think you may be right.

I was driving the Alfa something something race car '67, offline last week, and at stock it felt really neutral, but wouldn't really swing the rear without me intentionally sliding it.
Once I changed to an uber setup of max/min across the board, it got looser, but even still felt like the game was keeping the rear in-line to a point for me.

In short, I do think there's a hiden driving aid in GT5 that keeps the rear from coming out too much, and this might actually be the online/offline difference, being that online doesn't have this aid.

None of this changes the way settings work, unfortunately though, because online, the ride height/suspension stiffness settings still have the same effect for me as offline, I just need less of it. Depending on the car, like a crazy over steering monster, the GT500 NSX, I even go backwards, with a stiffer rear and higher rear to stop the over steer.
But the "nanny" I think I feel offline, isn't there online.

 

[oopssorryy]

This unknown aid could explain the different physics, if it is real. I think it controls the body roll/ wiht shifting, or at least heavily effects it. With some of the settings I have run the front end of the car should be dipping both forward and into the coner, but it doesnt. I think it limits the amount of weight that can transfer. Again I'm not sure how it would make the reverse setting work....

 

[Rotary Junkie]

My thought in GT4 was that because the inside rear was effectively "glued" to the ground the game sorta "pulled" it down when it should've lifted, pulling weight onto the inside rear.

 

[JackWilson]

That inside rear tyre is definately stuck down with some super industrial adhesive.

No matter how stiff you make the rear Anti-roll bars or springs, you will never get that off the ground.

I think that is our problem.

 

[shirakawaa]

The problem above might also be related to the inability to roll over cars if not with super sticky tires or by hitting heavy obstacles (at Trial Mountain for example), and that when they do, cars appear to unnaturally set back by themselves on their four wheels as soon as they can.

 

[chuyler1]

Without downforce, most of the cars in this game would flip over with racing tires.

 

[CSLACR]

I haven't seen too many cars flip in real life on smooth pavement, on any tire compound, for what that's worth.
Lifting the inside rear is a far cry from lifting half an entire car to roll it.

Also, I'm not even sure down force would do anything to really prevent a car from flipping on it's side, though I can see it helping a car flip, due to extra grip.

But I do think there is a stability control offline in GT5, what I'm not sure about is exactly how it works. All I know is that a car can feel very neutral, but if you push it even further, it goes back to under-steer, and it's probably the reason some people don't believe you need to do certain things to increase rotation, dependent on driving style alone.

 

[chuyler1]

If you drive in rush hour traffic it is a daily occurrence. People slam on the brakes in their SUVs, turn the wheel and end up blocking traffic for hours. But the dodge ram in this game can take turns at almost the same speeds as a sports car? LOL. You can't even get NASCARs to flip over when you have a pileup at 200mph.

 

[praiano63]

That inside rear tyre is definately stuck down with some super industrial adhesive.

No matter how stiff you make the rear Anti-roll bars or springs, you will never get that off the ground.

I think that is our problem.

Like Obama say : YES WE CAN. Nordschleife km 1 turn sequence no kerbs.

PEUGEOT 206 RC '03 // 500 PP // RACING SOFT

><(((((°>°°°°°°°°°°°°°