Is Suspension Tuning Backwards? - A Test with RX-8

[///M-Spec]

Good discussion in here. 👍

I didn't get a chance to try any of this last night, but should have a couple of hours to confirm everyone's prior testing. I do have some important suggestions for anyone trying to sort this out.

1) IMO, the best configuration to test with is a car with 50/50 weight distribution, IRS and equal sized tires/wheels all around. In theory, this type of car should be the most neutral, have the least amount of tendancy towards over or understeer and thus make for the best baselines. The Mazda RX-8 is like this. The BMW 330i is also another good candidate. My initial tests were done with a base M3, which has much wider tires in the rear than front (255 vs. 225).

2) Testing should be done on a flat surface as cambers and big elevation changes will throw off results. New York is very good for this as it has a lot of constant radius turns, in addition to a fairly circular skidpad at Columbus Circle. Seoul is another good one.

3) As Greyout mentioned, each test should be done with a single change to a single variable. Changing two things at a time would cast doubt as to which change caused which effect.

EDIT: 4) Also please try to note if you are experiencing over/under steer while WOT, braking, or steady state cornering.


M

 

[Greyout]

its unfortunate that you can't get a replay at the driving park. To get a photomode pic or even a paused replay, showing the car from the side when one end of the car is sprung much stiffer then the other, would be a quick indication of any backwards labelling in the setup screen.

has anyone tested sway bar effectiveness? 1/7, or 7/1, and its effect on balance?

 

[///M-Spec]

I just had an idea to visually test the "spring rates are labelled backwards" theory. (I'm at work, so can't try it)

Take a high-powered car to Vegas drag strip. The exterior view at the strip should show a side-on view at launch. The amount of squat under WOT should tell you right away if the hard or soft spring settings are correct.


M

 

[Greyout]

if someone does do this test, keep in mind that the pitch/dive angle wouldn't change a HUGE amount, the biggest difference would be the point of rotation.

Although the game doesn't mention it, as the springs get softer, they would have to get longer to maintain the same ride height. This means not only does it compress more under weight, but it would uncompress more as weight is unloaded.

notice in the drawing below how the cars are all pitched at the same angle, but they just rotate around a different point because of the differences in springs.

Edit: at least they would be if I could draw straight

it would be a good test though, because you can do a photo-replay on the drag strip from any angle, and really study the picture.

 

[boostdemon]

you have to remember that its not just springs.... it has to be proportional especially in regards to stay bar size and each oposing corner is balanced to its other... e.g. outside rear wheel vs. inside front wheel.

its all about roll centers

 

[awdrifter2]

Deleted.

 

[awdrifter2]

Deleted.

 

[///M-Spec]

Thanks, awdrifter2. So much for that theory...

Now I'm very eager to get home and try my hand at this as well.

(watches PC clock)




M

 

[Greyout]

hmm, thats some bad news...

test the sway bar effects of over/understeer

1/7, 7/1, everything else at the minimum with ride height at the maximum...

 

[Maturin]

Didn't GT3 have the same issue with less stiff = more grip? I can't find my copy, but I think it did.

Does the Japanese copy of GT4 work like this?

 

[Greyout]

Didn't GT3 have the same issue with less stiff = more grip? I can't find my copy, but I think it did.

Does the Japanese copy of GT4 work like this?

That might explain why I could never get my integra to rotate in GT3... at the same time, with a extreme-stiff rear end on a Silvia, it would drift like CRAZY in GT3 ( I think...)

Strangely enough, Japanese tuners look at things a little different then N/A and europe.

When a Japanese race team running a FWD car, such as spoon, tunes their race cars, they don't run super-stiff rears w/ soft fronts.

They actually run stiff front ends, which allows them to run a much lower ride height. The advantage of a stiff front, in their eyes, is reduced ride height (I guess most racing series over there don't have minimum ride heights like here in the US), which lowers CG.

They then correct for the significant understeer with tire stagger - they run big, beefy, sticky tires up front and less agressive tires in the rear.

With this tire stagger, they still are able to get the rotation they want, despite the chassis tuning. (and corner well with the low CG)

Maybe PD wasn't paying too much attention when getting some consulting, and only heard part of the feedback

either way, at this point, it appears that there is no other explination for the spring rate other then physics model error.

If the sway bars perform properly, i.e. stiff rear = oversteer, then I think my head will asplode.

 

[stormtrooper026]

I know you probly already have your answer but just in case the answer is, NO that is the way supention is suposed to work check out M-specs tunning info, he can explain it.

 

[Maturin]

I know you probly already have your answer but just in case the answer is, NO that is the way supention is suposed to work check out M-specs tunning info, he can explain it.

What are you talking about?

 

[Greyout]

...

 

[///M-Spec]

I know you probly already have your answer but just in case the answer is, NO that is the way supention is suposed to work check out M-specs tunning info, he can explain it.

Thank you for placing your faith in the guide, but it needs an update to cope with the changes in GT4. I have some theories at the moment, but not enough research to support it.

What are you talking about?

Please.. just let it go for now.


M

 

[Greyout]

Thank you for placing your faith in the guide, but it needs an update to cope with the changes in GT4. I have some theories at the moment, but not enough research to support it.

heh, you can delete my thread, I don't think it has anything to do with GT4

 

[///M-Spec]

heh, you can delete my thread, I don't think it has anything to do with GT4

I'm confident we can make it work in the context of GT4. This seems like too big an oddity for it not to be explainable. We've got a lot of people in this thread looking into it. At the very least, I'm waiting until sukerkin puts his 2 cents in. And since he's faster than me anyway, I'm taking his words for more like a buck.

My working theory is the soft front rates are causing front suspension geometry to change too much, thus causing the understeer. This actually has some applications in real life. For example, in SCCA Solo II stock classes, front sway bar is "free". You may think it would make a car push with a thicker front bar, but many cars (like M3s) with a strut front suspension actually picks up grip because it keeps the front end from compressing too much and changing the camber.


M

 

[westwood]

were simply missing something. its simply not possible for PD to have gotting something like this backwards. theres some variant being missed here.

 

[Greyout]

My working theory is the soft front rates are causing front suspension geometry to change too much, thus causing the understeer. This actually has some applications in real life.

The closest thing to this in the applications I have researched more closely involve the suspension compressing so much that they hit the bumpstop, or the upper A-arm hits the fender. When that happens, the front spring rate goes up to infinity (well, not really, as the rubber bumpstumps can be compressed slightly, but for ther sake of my point, I'm sure you understand).

if a front suspension was too soft, I agree that would cause the understeer when it reached the end of its travel... But from a suspension geometry standpoint, no camber curve I'VE ever seen is so agressive that it goes too far positive under compression. (Hondas for example, have pretty agressive camber curves (traditional short/long arm setup) and when the suspension is compressed as far as the chassis will allow, it gives you maybe 3 to 4 degrees. When you factor in the body roll counteracting the negative camber gain from compression, the out-of-wack geometry idea doesn't fit IMHO - at least when looking at the amount of body roll (or lack of) in the cars.

The minimum spring rate on many "full race" suspension package in GT4 is at or even slightly above stock spring rate.

When I was running full-stiff spring in the rear, and somewhere in the middle in the front, I noticed that the car practically didn't roll at ALL in turns. with the soft front, the front left should compress more, with the car pitching foward and rolling to the point that the inside rear comes off the ground.

I think the limitation of body roll and the game's refusal to allow the unloading of a tire on the inside is causing this.

I wish I could just go home and play with it, instead of sitting here at work thinking about it

edit: all my tests were done at maximum ride height, so I don't think the above mentioned was a problem.

has anyone tested the sway bars yet?

 

[///M-Spec]

The closest thing to this in the applications I have researched more closely involve the suspension compressing so much that they hit the bumpstop, or the upper A-arm hits the fender. When that happens, the front spring rate goes up to infinity (well, not really, as the rubber bumpstumps can be compressed slightly, but for ther sake of my point, I'm sure you understand).

Yeah, hitting the bumpstops is a good theory. That would effectively give you front a front setting of like a 99. Maybe the bug is that the front tires are crashing into the top of the fender lining?

if a front suspension was too soft, I agree that would cause the understeer when it reached the end of its travel... But from a suspension geometry standpoint, no camber curve I'VE ever seen is so agressive that it goes too far positive under compression.

It is very common in the front strut world of BMWs and Porsches, where static camber is set to very aggressive levels because MacPherson struts have horrid camber curves. My street/auto-x M3 was set to -2.5 and many people run -3.0 or more. Incidently, the stock progressive springs on that car was something like 87 lbs/ 1st inch. while a typical club race car runs 450/500.

The minimum spring rate on many "full race" suspension package in GT4 is at or even slightly above stock spring rate.

When I was running full-stiff spring in the rear, and somewhere in the middle in the front, I noticed that the car practically didn't roll at ALL in turns. with the soft front, the front left should compress more, with the car pitching foward and rolling to the point that the inside rear comes off the ground.

I think the limitation of body roll and the unloading of a tire on the inside is causing this.

I wish I could just go home and play with it, instead of sitting here at work thinking about it

Me too. About to go in 10 minutes.. but then have to fix/eat dinner, play with the kid, spend time with pregnant-demanding wife... etc..


M

 

[Greyout]

my wife is going out of town tomorrow for a week

so tonight = no GT

but for the next 7 days = GT night and day! (except for that whole "work" thing)

 

[JaeTea]

If the sway bars perform properly, i.e. stiff rear = oversteer, then I think my head will asplode.

I haven't seen it work to the effect it does in a real life application.

SUPER rear stiff and almost zero front would send you into an oversteer almost immediately at turn in.

If it does in fact work with a stiffer rear causing more oversteering then I can't begin to understand how so such a soft rear spring would cause more oversteer.

 

[awdrifter2]

Deleted.

 

[awdrifter2]

Deleted.

 

[Maturin]

Negative toe = toe out?

That's already been established.

 

[awdrifter2]

Deleted.

 

[sukerkin]

I'm confident we can make it work in the context of GT4. This seems like too big an oddity for it not to be explainable. We've got a lot of people in this thread looking into it. At the very least, I'm waiting until sukerkin puts his 2 cents in. And since he's faster than me anyway, I'm taking his words for more like a buck.

[Sukerkin looks up from running licence test]"Mmm ... wha ... ?" .

Many thanks for the vote of confidence my friend 👍

I'll be looking into this from an empirical standpoint over the weekend when I have some cars and cash to work with. I'm thinking that any inadequacies in the suspension algorithms will probably show up fairly quickly as my DFP hasn't been shipped yet so I'm driving with the DS2 (can anyone say "Nurburgring" & "Fishtail" & "Morgue" in the same sentence) !

If they've improved the sensitivity of cars in GT4 to spring rate changes that'll be a good thing as it was something of a blunt instrument in GT3 (as oppossed to the rapier that is damper tuning ).

However, if they've really messed up how it's been implemented I'll be shocked. I know that nobody properly playtests games these days but abrogating the laws of physics in a supposed simulation would be a very poor show.

My guess is that what they've actually done is implemented some aspects of the physics engine that weren't turned on before so we're going to have to adapt accordingly. As someone has already stated there was code for camber change etc in GT3 that was simply disabled (I remember reading an article on this years ago wherein they explained that they didn't think people would be able to drive the cars with all the 'realism' factors switched on ROFL)).

I'm actually looking forward to this if it means that some of the 'unreal' tuning set-ups of yore that some used in GT3 really will make you crash in GT4 👍.

Got to go ... bosses looming and looking disapproving (they foolishly believe that the power stations and sub-stations of this country are more important than GT4 ... silly mortals ).

 

[Mr Whippy]

Not sure about it being excessive suspension travel...

Even with a big ride height, lots of travel, and a 5.2kg/mm rate ~ 52000N/m spring rate, a bmw m3 has an almost 1:1 motion ratio.

Hence, IRL, the stock bmw m3 spring rate at the front is about 23500N/m, or 2.35kg/mm.

If THAT is too soft, then this problem we are seeing should be exhibited all the time!

However, a rate of 52000N/m on a lightweight M3 CSL with stage 3 lightening (what I used to test), and 5.2 front, 10.4 rear with everything else at 0 or 1, left the car with oversteer traits...

What was bizzare though, was the fact a VERY stiff front, and a soft rear was quite slidy.

Either way, i wasn't happy with the level of control the suspension had. It seemed to make very little difference in my 1200kg, 480bhp M3 CSL... OK, it drove better with normal settings, but the oversteer/understeer traits hardly changed with huge ARB differences.


With "normal" settings though, because I assume RACE suspension can't be used to make a soft setup full stop, the CSL did follow normal tuning traits, but up to a limit... after what I would consider "normal" ranges, the suspension went a bit crappy, and what was going on wasn't clear!

I'd suggest testing with an rx8 with no tuning, just a selection of tyres and racing suspension... Too much power brings power-oversteer into the equation!



Really, GT4's whole suspension setup is crap imho. GT3's was more clear... ideally we should have a realtime pic of the car as we adjust stuff, so we can see it alter, and can visually SEE that the wheels point or camber the right way...

To be honest, I think GT4 is right, within normal ranges it seems ok, and I expect they can't have made such an obvious bug. I think that it's probably to do with the damping (as we don't know what the values mean at all (can we see shaft speeds in telemetry readout???)), or it could be bottoming and we all know a bottoming rear suspension would cause instant oversteer!!!
Like I say though, not having any visual aids or decent telemetry or on-screen debug info for suspension makes it really tough to know what is happening when testing your setups...
Big thumbs up to GT4, loving it so far, but big thumbs down to the interface and clarity, and mainly, the ease of use of the suspension setup!

Dave

 

[sukerkin]

I concur; Damping compatability is also going to be my first point of study Mr. W (unless someone else gets there first of course ... subtle hint to those not at work !).

If GT4 no longer has the 'magic' dampers of GT3 (which appeared to largely rate themselves to the springs so you only had to worry about relative values) then that seems an obvious candidate for mysterious, counter-intuitive behaviour from changes in spring rates e.g. soft springs + stiff dampers = effectively increased spring rate leading to understeer or oversteer depending on which end of the car is in question (crosses fingers in the hope that it's really that straightforward ).

 

[Eric.]

...Well, after reading all of that, I'm going to grab some aspirin.

Good luck on figuring this out. I wish I knew more about suspension tuning to help.

With the soft rear, could it not be transferring the weight to the back quick enough, thus causing the oversteer? Just my two cents (worth half of everyone else's in here )