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

[///M-Spec]

I've got this problem partially figured out, after spending almost 4 hours on it last night.

It appears everything does work according to conventional tuning wisdom... provided you use N1, N2 or N3 tires. Bear with me, this is going to be a long and involved explaination. But if you've read the whole thread up till now, you're probably dedicated enough to go through this anyway.

Here is my test setup:

Car: RX-8 Type S (J) Bone Stock
Track: New York Normal
Tires: Sport Medium, N2, N3
Mods Bought: Full Race Suspension from Mazdaspeed shop
Computer Settings: ACM and TCS all set to ZERO
Steering Type/Mode: Logitech GT Force/Simulation Steering

Part One - Shaking the car down stock

First thing I did was to get used to the car on the track. I ran about 15-20 laps in Practice Mode and started putting down consistant laps in the 1:52:xxx range (I'm confident I left several seconds down by Central Park South, so there is definatly room for improvment. I'm sure some of you faster guys can dip into the high 1:49s). The car was absolutely Bone Stock.

Columbus Circle: Light understeer under throttle, very light lift-oversteer
Turn 1a (Broadway to 42nd Street): Strong understeer under braking
Turn 1b (42nd to 7th Avenue): Moderate understeer under throttle
Turn 2 (7th Avenue RIGHT to 54th): Moderate to light understeer under braking
Turn 3 (54th LEFT to Avenue of the Americas): Light understeer to light oversteer
Turn 4 (AoA RIGHT to 57th Street): Light oversteer under throttle
Turn 5a (57th LEFT to 5th Avenue): Almost neutral - cornering attitude adjustable with throttle, with tendancy towards understeer at high steering lock
Turn 5b (5th to Plaza Hotel): Same as 5a
Turn 6 (Plaza Hotel RIGHT): Neutral under braking
Turn 7 (Plaza Hotel LEFT to Central Park South): Light understeer
Turn 8 (Central Park South RIGHT into Columbus Circle): Neutral under braking with tendancy towards light understeer

The purpose of this is to establish (with me behind the wheel) going as fast as I can in about 15-20 laps, how the car predominately behaves. My conclusion is the RX-8 is a fairly neutral car with a light to moderate tendancy towards understeer --which describes most modern sports cars. My other conclusion is the RX-8, like many cars can be made to understeer AT WILL simply by pushing too hard on the car while under WOT. No surprises here. My other conclusion is the RX-8 in GT4 does a good job of behaving like the real thing (which I drove last year).

NEXT: N2s


M

 

[Greyout]

I don't think its an issue of extremes. I was playing with my S2000 this morning. The spring rates are on the high side, but not maxed out (like 13 or 14 or something). Making small adjustments to one end by 3 to 4 kg/mm produced the same backwards results. No changes were made to the dampers.

The sway bars are very ineffective also - only the slightest change from 1 to 7... I THINK that they are correct though (7 in the rear leading to oversteer - but hardly)

I was playing with the dampers also going 10/10 at one end and 1/1 at the other, to see how the car behaved at turn in. I remember being confused, but not sure what happened, I'll have to do it again

 

[///M-Spec]

Part Two - N2s

My next step was to see what the car would do with the race suspension in Default, Stiff F/low dampening, Stiff F/mod dampening, Stiff R/low dampening and Stiff R/mod dampening.

During this test, I inadvertantly (while using the SET A, B and C buttons) put N2s on the car. Imagine my surprise when the car suffered severe traction loss using the same braking points coming down Broadway into Turn 1a

No problem, though. Let's see what the car does with N2s: I spent most of my time here and I can happily report that with THESE tires on the car, a STIFF FRONT will PROMOTE understeer in situations where there was light understeer. A STIFF REAR will PROMOTE oversteer in situations where there was previously light oversteer.

In the case of my full test with STIFF REAR + MOD DAMPENING:

Springs 5/7
Shock Bump 1/4
Shock Rebound 3/3
All other settings DEFAULT

The car behaved like this:

Columbus Circle: MODERATE to STRONG oversteer under throttle, MODERATE lift-oversteer.
Turn 1a (Broadway to 42nd Street): Still strong understeer under braking
Turn 1b (42nd to 7th Avenue): LIGHT TO MODERATE OVERSTEER under throttle --sometimes STRONG oversteer. I even spun it once.
Turn 2 (7th Avenue RIGHT to 54th): Still moderate to light understeer under braking
Turn 3 (54th LEFT to Avenue of the Americas): Light understeer to light oversteer
Turn 4 (AoA RIGHT to 57th Street): Light oversteer under throttle
Turn 5a (57th LEFT to 5th Avenue): Still almost neutral - cornering attitude VERY adjustable with throttle, with STILL a tendancy towards understeer at high steering lock
Turn 5b (5th to Plaza Hotel): Same as 5a
Turn 6 (Plaza Hotel RIGHT): Neutral under braking
Turn 7 (Plaza Hotel LEFT to Central Park South): MODERATE oversteer
Turn 8 (Central Park South RIGHT into Columbus Circle): Neutral under braking with tendancy towards oversteer.

If I reversed the settings, the car understeered in the places where you would expect. In some cases, especially under WOT, the car pushed very badly. N3s behaved very similarly, but with correspondingly higher limits before traction loss.

NEXT: Sport Mediums


M

 

[Greyout]

another thing that annoys me is that the FWD cars, or at least the prelude, understeer MORE in HIGHER gears when at WOT, and OVERSTEER at WOT in lower gears.

in 3rd gear, in the last sweeper before the finish line at tsukuba, the car pushes at WOT, and can be tucked in with a slight reduction in throttle.

in 2nd gear, in turn one, the car is neutral until WOT is applied, then the car OVERSTEERS. As long as the front tires are spinning under acceleration, the steering response is INCREASED, and understeer is ELIMINATED. Infact, to prevent oversteer, I sometimes had luck leaving it in 3rd gear and powering out of the turn, with a slight push, rather then dropping to second, delivering more power to the front wheels, and somehow oversteering.

The only thing I can think of is that when the tires are spinning, acceleration is reduced, and PD's engine simply assumes that less acceleration = more weight on the front = more turning athority from front tires, even if those front tires should have LESS traction because of more power application, or a loss of traction from power.

not sure how I feel about GT4 anymore...

edit: who was the member that is a writer for IGN or something?

and sorry to jump in the middle of your posts

 

[///M-Spec]

Part Three - Sport Mediums

Yep. Just like everyone was telling me all week. Putting the Sport Mediums back on the car completely transformed the way it behaved. Stiffing the FRONT springs seemed to improve steering response, quickened the turn-in BUT had little effect on cornering attitude. It STILL understeered where it used to and in some cases MORE. However, I could get PLENTY of oversteer --but ONLY of the lift-throttle variety. If I kept my foot in it, the car just pushed.

EDIT: I need you guys to verify something for me. When you stiffen the front and soften the rear, does your car BEGIN to get loose only when you lift in mid-turn or can you do it simply by going WOT at turn in?

Stiffening the rear seemed to do nothing. I mean NOTHING, except make everything happen quicker. The car pushed where it did before and was NOT ANYMORE loose in places where it was before.

So this test left me bewildered and somewhat annoyed. I finished my beer and went to bed. I am wondering if PD simply gave you a different physics model DEPENDING ON TIRE SELECTION. ie- N2s are the new "Simulation" tire and everything else in the game is just bunk.

I'm open to suggestions.


M

 

[sukerkin]

I shall still be interested in trying this experimental line out for myself {Sukerkin interested in spannering ... surely not! } but it's starting to sound like the groundworks been solidly explored already and the evidence is pointing towards there being a 'problem' with the handling physics if you use stickier tyres ... .

 

[Greyout]

those are much lower damper settings then I ever used - I always thought that the damper settign had to match the spring rates - but, as I read above, apparently they scale themselves to match the spring rate in previous versions of GT...

This is great news, but kinda weird about the tires...

 

[Greyout]

great news as in at least its ok with normal tires... what about race tires?

 

[Mr Whippy]

In the sim I work with, and have produced an M3 for (Racer Free Car Simulation), even altering the rear ARB rate by 10% is very noticeable when driving, changes the whole attitude of the car, and the way you drive entirely.

It's weird, is GT4 kinda tweaking the ranges of these "1-10" sliders for us. When we add huge rear to front arb ratios, is it working like the brake balance controller?

Is 10f/10r arb the same as 5/5? or is it semi-intelligent, and maybe 1f/10r works out what your after, but knows you don't want 10 units of rear arb for every front 1 unit?

I dunno, just seems weird how it all works, not at all clearly consistent... A usual good sign of how good a physics engine is how well it reacts to REAL LIFE tuning methods, and unfortunately GT4 isn't doing too well.

The only 100% clear things we can modify in GT4 right now are the "spring rate", which may well be the "effective wheel rate", not the actual rate of the spring.
The second is the ride height. The rest is just numbers with no units of measure, and could for all we know mean anything.
In a way, I'm finding the tuning suspension from GT3 better. The MINIMUM stiffness I can get from race suspension on the CSL is 52,000N/m rate, which is double what the real car runs at the front in normal M3 spec... Thats quite stiff.

Because we are testing exclusively with road cars with very stiff suspensions, it's worth noting the new "body stiffening" option in the suspension area. Clearly cars might be wobbly, and having very stiff suspension on a non-rigid body might be making the results for super-stiff suspension (or even super soft at 52,000N/m is very stiff), not come out as we assume.

Perhaps test on a racing car these handling traits? Having 100,000N/m,10.0kg/mm springs on a stock RX8 is pretty damn stiff, and tyre camber change is more likely to be through body deflection if it's been modelled.

Anyway, some of what Greyout has been saying about fwd cars is a bit odd. It's funny to see him noticing understeer being more of a function of acceleration rather than real tyre grip. Spinning fronts cause terminal understeer, unless using an LSD or similar. Weird to see that spinning fronts in a lower gear is causing oversteer indeed.

Worries me, delay after delay, a new physics engine... hmmm, wonder if they fudged it all.

Afterall, this "new" physics seems worse than GT Concept, which imho was the best last game. Automatic cars like SL55 AMG slurred as we changed gear, the Skyline concept had 0 shift times etc.
In GT4, the CSL M3 shifts up like a manual, not an 80ms robotised shifter. I wonder if the TT DSG Audi shifts in 0 time, or has now mysteriously got a manual gearbox!?

So many flaws and faults, just doesn't feel finished in the physics department. Rest of the game is awesome, but B-spec and photomode would have been better OUT and decent physics polish IN...

I hope the online version has these problems fixed, if it ever comes out. Otherwise I may not be bothering getting it, and wait for GT6...

Dave

 

[///M-Spec]

those are much lower damper settings then I ever used - I always thought that the damper settign had to match the spring rates - but, as I read above, apparently they scale themselves to match the spring rate in previous versions of GT...

This is great news, but kinda weird about the tires...

Yes, my settings were pretty low. I wanted to simulate a typical "street/track day" spring set you might find on a typical lightly modded car. 5 kg/mm is like a 300 lb/inch spring... so I just went with that. There seemed to be plenty of dampening and body control when I drove the car.

Great news and bad news... It just seems ridiculous to me that if you want the suspension to work the way people thinks it should, you have to put N tires on.

I have not yet tried this thing with R tires. I'm broke in the game right now and need to go make some $$


M

 

[Greyout]

is it possible to be in practice mode with tire wear on?

I wonder what happens as the sport tires heat up & wear? Could it be PD's goofy way of simulating cold tires? (rather then just reducing their grip...)

 

[///M-Spec]

^ I don't think so. (getting tire wear in practice mode. I'm pretty sure they are set to "indestructible")

Speaking of tires... I was driving to work this morning thinking how nice it would be to get a pyrometer reading from the tires which would be a HUGE help (in GT4, not my daily driver.. though that may be useful too ). But I may as well start wishing for money to fall from the sky and Cate Blanchett appearing at my door for a cocoa butter rub-down.


M

 

[awdrifter2]

Deleted.

 

[///M-Spec]

Sure thing. And awdrifter2 (and anyone else trying to replicate my test results), when you go from Sport Mediums to N2, just remember the Ns loose traction much sooner than the Sports, so even just braking in a straight line the car will plow like mad. You'll want to use 80-90% braking force or else the ABS will kick in and greatly lengthen your stopping distances. At first, it seems like crazy terminal understeer all over the place, but if you back off the brakes and steering a little, the car is very "real" feeling and will respond to conventional tuning techniques.


M

 

[Nelsy]

hey guys, i've read the whole thread, it's taken me about an hour...
now i'm getting all confuesed.. but hey thats not too hard to do...
i have GT4 and GT3... and GT2 and GT1... in fact i have them all....

but here's the thing with GT4...
i only started playing a few days ago..
and i've been tuning my G4.. coz it's a cool little car...
now after nearly fully modding it...
i had real bad understeer...
so i put on a standard GT3 FR set up on it...
this didn't help me at all.. in fact it got worse...
so i defaulted it.. went back to the drawing board...
and now it has great handling, although a little oversteer exiting the turns.. but nothing that can't be controlled..
then i noticed my breaking distances were way to long...
so i tweaked to BC and it helped a little, but i was still not happy with it, as this car is really lite..

so i went back to the drawing board..
i raised the height a lil and dropped the spring stiff...
and now the car is nearly perfect.. i just need to tune out a little oversteer during corner exits...
i'm very happy with my results in tuning this car...

basically all the techniques i have learned about tuning from the other GT series seemed to work here in GT4, although i had to be alot more delicate with what i was doing...

next i'm working on something with a lil more power, 3000GT or something.

thanks guys for your time.

ps.. anyone played online yet?

 

[awdrifter2]

Deleted.

 

[Maturin]

I disagree with M-Specs test results. I tried the same test with RX-8 S, with N2 and N3, and have the same results as my initial post. More stiffness = more grip.

Could people PLEASE put aside preconceived notions on how the game SHOULD work, and just test how it DOES work? Get a family member or friend to adjust the game's slider for you, blind, and then do a blind test.

 

[Maturin]

great news as in at least its ok with normal tires... what about race tires?

It's not OK with ANY tires. Try it yourself.

 

[Eric.]

is it possible to be in practice mode with tire wear on?

I wonder what happens as the sport tires heat up & wear? Could it be PD's goofy way of simulating cold tires? (rather then just reducing their grip...)

You can turn tire damage on in two player mode. Just raise the number of laps so you can go longer. (I assume you can change the number, I haven't attempted.)

 

[///M-Spec]

I disagree with M-Specs test results. I tried the same test with RX-8 S, with N2 and N3, and have the same results as my initial post. More stiffness = more grip.

I don't know why you would dismiss my results out of hand without being more specific about where the behavior was occuring on the track. To say a car ALWAYS understeers or ALWAYS oversteers is a gross oversimplification and indicates to me inadequate testing.

Where did it understeer? What were you doing to the steering? The brakes? The throttle? Do you use DS2 or a wheel? What steering mode? These are all variables that are not supplied.

In real life, a car can understeer or oversteer, sometimes in the a single corner, depending on how you drive it. I can make any car understeer if I wanted to. I can also make most cars oversteer even if it weren't particularly inclined to.

GT4 is quite accurate in that regard when N tires are used. I stand by the results of my testing.

Could people PLEASE put aside preconceived notions on how the game SHOULD work, and just test how it DOES work? Get a family member or friend to adjust the game's slider for you, blind, and then do a blind test.

The implication that somehow you are being objective and "people" are not (I can only assume me at this point) is highly unwarranted.


M

 

[Greyout]

My tests last night with a Prelude & S2000 disagree with your results also.

The N tires simply reduced grip, but as I adjusted the settings, the car still responded incorrectly. Stiff front, soft rear = oversteer.

Was everyone agreeing that the sway bars were backwards also, or no? Their effects are too small for me to tell for sure.

My rear wheel drive cars (be them mid or front engined) cars are all suffering from initial turn-in oversteer, and on-throttle understeer. They only have about 300 hp, which isn't enough to light up the tires in any gear, but the push when on the throttle is a giving me some trouble... any suggestions?

My initial thought was to stiffen up the rear shocks, but god knows what that will do...

 

[///M-Spec]

Just went back and double-checked my results. Sorry guys. I stand by my statements on the N tires. Stiffer rear = generally more oversteer in situations where the stock oversteered. Less understeer in situations where the stock car understeered. Vice-versa when settings were reversed.

That doesn't mean I don't get understeer when driving with the stiff rear settings. It means I get LESS, when all other things are equal. Same track. Same car. Same approach to a corner. Same speed at turn-in, etc. As I careful state in my guide such tendancies are just that. Tendancies.

Until people begin supplying the details I requested elsewhere in this thread regard controllers, steering settings and the exact circimstances under which the behavior occured, I'm going to conclude there is insufficient reason for me to doubt my own findings and we'll simply have to agree to disagree.

As I've already said: There is no such thing as a car that ALWAYS understeers or ALWAYS oversteers. The same car can exhibt either behavior depending on what the driver does and I reject such oversimplifications.


M

 

[Greyout]

I am using the first 180 degree turn, and the last right hand sweeper (of Tsukuba), as my testing point. The car is a Honda Prelude.

I am braking well before the turn, turning in smoothly, and applying just enough throttle to maintain speed. I use the regular controller.

I started off with 10.0 front / 10.0 rear. Car oversteered slightly when the throttle was closed, and pushed when power as applied. with 8 front / 16 rear, the car pushed more, and the speeds at which I took the turns before resulted in a understeerage off the outside into the grass. With the settings reversed, 16 front / 8 rear, the car oversteered slightly at turn in (requiring the steering to be brought nearly back to the center for a second or to). in mid turn, the car maintained a slight drift angle that was controlled by throttle.

The other problem: in 3rd gear, application of more throttle resulted in understeer. in second gear, application of more throttle resulted in wheel spin, which tucked the front tires in more, requring countersteer to prevent the car from oversteering into a spin. again, this a FWD honda...

 

[Maturin]

I don't know why you would dismiss my results out of hand without being more specific about where the behavior was occuring on the track. To say a car ALWAYS understeers or ALWAYS oversteers is a gross oversimplification and indicates to me inadequate testing.

I don't need a detailed rundown of testing to know that when I start at springs f8/r12 shocks 6/8, the car understeers in all situations (entry, midturn, exit, WOT), and then OVERSTEERS GREATLY in all situations (entry, midturn, exit, WOT). On all tracks, in all weather conditions, with any car I choose.

I don't know how you came to the results you did, other than you can't look beyond your own guide. Even Greyout can't replicate your results.

Ask yourself this: What is more plausible? That the physics work, as they do, on all tire types, or that there are THREE tires types that have the opposite physics from the REST of the game. Which would require that there are TWOs physics engines -- one for most of the game, and one different (to figure out changes in camber, geometry, weight transfer), for JUST THREE TIRES, which are NOT default, which you have to BUY, and which perhaps 3% of the populace is even aware of. That's a lot of wasted programming for THREE tires that maybe 10,000 hardcore sim addicts will use. And if you were going to have a "real" physics engine, why wouldn't you just use that for all of the tires, since reversing the sliding direction on the sliders wouldn't seem to cause casual gamer's heads to explode. If I was the CEO of Sony, I would FIRE them for wasting the shareholder's money.

"Uh...yeah, Mr. CEO, we doubled the manhours on the physics engine, which is most of the game, so that there could be three obscure tire types that match M-Specs theory at GTPLANET.NET. Uh...yeah, my office will be empty by the end of the day."

 

[Maturin]

Until people begin supplying the details I requested elsewhere in this thread regard controllers, steering settings and the exact circimstances under which the behavior occured, I'm going to conclude there is insufficient reason for me to doubt my own findings and we'll simply have to agree to disagree.

Oh boy....

The implication that somehow you are being objective and "people" are not (I can only assume me at this point) is highly unwarranted.

Er...you were saying? You're the only one who isn't seeing what the rest of the board is seeing, what our own eyes see. And it's for only THREE tires types, as well. How odd is that? So...yeah....I would say its DEFINATELY WARRANTED.


M-Spec's gone bye-bye, what do you have Greyout?

 

[Maturin]

controllers, steering settings and the exact circimstances under which the behavior occured,

I'm using Dual Shock for all tests (unless you think they used a third physics engine for Dual Shocks). Should eliminate TWO of those requirements right off the bat. As for the third, I did that in my first post, and most recently ("all circumstances" physics are reversed). However, because I never changed my three tuning SETS for the RX-8, I went back and bought N1s to do the SAME test as the original post. Here it goes:

TEST

Car: RX-8 Type S(J) - FR 285HP (I assume this car shares the same physics engine as the rest of the game)
Tires: N1 (According to M-Spec, does not share the same physics engine as rest of the game)
Track: Trial Mountain (I assume this shares the same physics engine as the rest of the game)

SETUP 1 --- SETUP2 --- SETUP3

Springs f10/r10 --- f10/r15 --- f10/r6
Height 120/120 --- 120/120 --- 120/120
Bound f8/r8 --- f6/r8 --- f8/r6
Rebound f8/r8 --- f6/r8 --- f8/r6
Camber 2.0/1.0 --- 2.0/1.0 --- 2.0/1.0
Toe 0/0 --- 0/0 --- 0/0
Stability 4/4 --- 4/5 --- 5/4

RESULTS:

SETUP1 (Default): neutral, some real minor throttle oversteer
SETUP2 (rear stiffer): Very noticeable Understeer, in all turns.
SETUP3 (rear softer): Moderate to serious oversteer, in all turns. Braking, mid-turn, exit.


AS ALWAYS, I ENCOURAGE OTHER BOARD MEMBERS TO REPLICATE THESE RESULTS (particularly M-Spec, and then post his results), OR DEBUNK THEM. SHOULD BE RELATIVELY STRAIGHTFORWARD TO SEE WHETHER I'M LYING

 

[Kolyana]

AS ALWAYS, I ENCOURAGE OTHER BOARD MEMBERS TO REPLICATE THESE RESULTS (particularly M-Spec, and then post his results), OR DEBUNK THEM. SHOULD BE RELATIVELY STRAIGHTFORWARD TO SEE WHETHER I'M LYING[/B]

I don't think anyone has accused you of lying...

This has been a most informative and interesting thread. Everytime I come in this forum I look for it and the latest findings and I will probably add my own at some point this weekend.

However, you do appear to be somewhat inflammatory and aggressive towards M-Spec who has, to me, at least attempted to thoroughly provide a detailed explanation of his findings and he has not - anywhere - been aggressive or derogatory of you or anyone else, even if their findings seem to be contrary to his.

I would like to see more findings added to the pile and I - for one - will be attempting to correlate engine position and drive type to suspension settings. I would hope ... and I do not know what these will be yet ... that if my findings differ from yours, even in part, that you will not see this as a reason to dismiss me as some form of lying M-Spec-wannabe and fangirl.

I am genuinely interested in how this study works out - one way or the other - because the net result is that - either way - we'll have a much greater understanding of how GT4 works .... whether it vindicates your position, M-Specs or even somewhere inbetween.

Personally, I am grateful that you brought this subject forward as I think it has sparked one of the more interesting discussions anywhere in the forums, but - please - realize that it's not a personal attack upon you if someone concludes findings slightly different from your own. Heck, I could sit in front of GT4 tonight and find that what you're saying is exactly what I find, but then I will not take that as some form of M-Spec ploy to deceive the community and take over the world, but rather simply one more piece in an ever revealing study ... a study I will still be following with open-minded interest even AFTER I've done my own testing.

 

[ford_racer]

I just asked a friend of mine to give it a shot.

He softened up the rear of the prelude, and stiffened up the front. He said it is the "oversteer king"

This is very bad.

Until I can go home and try this myself, I take back anything good I ever said about GT4 physics.

i am not an expert tuner, mearly a novice, but maybe if there is alot of the car's weight (maybe over 50%) and because there is alot of body roll, the rear tires start to "roll over" and the rear end slides out, once again, because of the body roll. just my opinion

 

[Greyout]

i am not an expert tuner, mearly a novice, but maybe if there is alot of the car's weight (maybe over 50%) and because there is alot of body roll, the rear tires start to "roll over" and the rear end slides out, once again, because of the body roll. just my opinion

it would cause the inside front to lift off the ground, with a vast majority of the transfered weight being suppported by the outside front. The outside front would be severely overloaded, the inside front would be off the ground, and you'd have a one-way ticket straight to the wall

 

[Greyout]

has anyone played with the shocks yet, to find out how they effect transition balance?

I am pretty sure the damper rate does not have to match the spring rate (as in the range adjusts automatically with the spring rate)