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

[sukerkin]

Hi Greyout

Nice to see that you haven't changed your tune my friend. If it works for you to think of the rates as reversed then that's all fine.

A team of us have worked on this for pretty much as long as the threads been open and I can only report that the overwhelming conclusion we've reached, based on who knows how many hundreds of man-hours of testing, is that the spring rates are not reversed.

As I noted above, the physics engine has much more sensitivity to the whole raft of suspension adjustments than GT3 had. I have to admit that I find it hard to comprehend some of the statements made by others, as I have found such things as the Dampers to be extremely responsive to adjustment and the Stabalisers were of great utility in getting the Elise to handle properly on corner exit.

However, driving style obviously has an overwhelming impact on perceived changes so, as with the Dampers are Reversed school of thought from GT3, I reckon that we'll just have to agree that we observe different things on this matter.

This is not to say that I think that PD should get away with calling this game a "Driving Simulator" tho'.

There is something rotten at the heart of the physics engine, as you can get some very contrary and un-fathomable results from tuning adjustments and can find that the same car, at the same speed, on the same corner will have very different handling characteristics. It's just that it's not as simple as having a single programming transcription error reversing the spring rate values.

Altho' I feel a little treasonous for saying so, I'm hoping that one of the 'contenders' for the GT sim crown will knock PD's hat off this year or next as the Gran Turismo series has lost it's way.

Or maybe I'm just expect too much from a 'game' on a console that claims to be a 'simulator'?

 

[DrBiggly]

I agree; it's a game and not a "simulator." Claiming that it's a simulator is quite a stretch.

I wish I could still enjoy the game but as of the past month I've had a workload of triple or quadruple my norm and thus I've been unable to spend any time to get GT4 back up and running. It locked up in the middle of an endurance race (the last one I had left I believe) and hasn't worked since. The disc is mildly scratched and the console no longer recognizes it; the day I received the game it didn't recognize it save for on the 20th attempt of reinserting the game.

So my game is currently still stuck at 90% or so and has been for the past month. I'm shortly acquiring (well, basically it's mine now) an XBOX and will pick up Forza Motorsports and go with that as all of my racing friends have moved on to it with a greater fun factor and greater realism factor as well. I'll reserve judgment until I get the chance to play with it on my own.

My personal feeling on the "suspension tuning backwards" is not that the springrates are completely backwards, but that the algorithm that changes over and understeer characteristics is backwards with respect to springrate. *All other facets* of the springrate changes for front and rear operate as they are supposed to, save the general over/understeer (or basically rotational) characteristics.

Dangit, I really wish my game would work again. I could use some stress relief these days.

 

[sukerkin]

Ooh, that's tough luck DrB. Getting the 'disk not recognised' error after moving so far into the game must be irritating beyond belief. I sympathise.

I like your expression of how you see the 'suspension gremlin'. That gels somewhat with what we've been observing in our testing in terms of the handling effects.

For myself, I've started to feel that where the 'problem' lies is that there is a basic FF, MR, FR, 4WD set of models in the physics with an overriding weight distribution tag stitched to them.

What brought me down this path was trying to set up a Lotus Elise. It wasn't until I assumed a rough 40/60 weight distribtuion and put the spring rates to that ratio that the car started to behave. Similarly, I find most FR's/4WD's favour a near 50/50 split and most FF's 60/40.

Other than that, like yourself, I feel that most other areas work okay(ish) but that once you stray higher up the HP foodchain things become more obviously odd.

 

[rk]

well you can also call it the "greyout effect" because I get it on several cars

Would you be kind enough to list the cars and the conditions under which your spring behavior reverses? I would like to set it for my cars and experience it.

Perhaps considering the fact that uncounted others could benefit from the action of your being more specific and the results might help you overcome your diffidence.

 

[v BONES v]

phew! finally finished reading this thread.

First off I'll just say that I'm a complete noob at this - only just started GT4 and starting on the long road to understanding. This site is required reading, there is more valuable information here than anywhere, particularly than the manual or official GT websites - keep up the good work guys.

I have had a go at testing the suspension physics myself using the limited resources I currently have available - I haven't been too scientific about it so you can take it as anecdotal or leave it, probably worth mentioning that I am using the PAL version of the game. I read and fully understand Mspecs tuning guide and won't be arguing against the real world principals outlined there.

I choose two cars to test. I have a Subaru Impreza Sti (4WD) and a Nissan 300ZX (FR), both are pretty much stock with few upgrades, both output around 380 BHP.

Firstly I ran the Subaru on the Special condition 2 course (2 reasons for choosing this track, firstly I get to pick up over 250,000 for research funds every two races, and secondly it consists mainly of straights and hairpins which emphasise understeer). The Subaru performs very well and feels well balanced, cornering is good and I can get slight oversteer under power if needed - I win the races easily. The value of this run is that it gives a good feel comparison when I change to the 300ZX.

I use the 300SX and run the exact same course. Immediately I am getting very noticeable understeer (it is particularly notable as I have literally just "got out" of the Impreza). I still win easily but the margin is much reduced. Okay I have established that the 300SX has a defined understeer problem and I want to get the handling closer to the Impreza.

I go straight to the suspension settings to do the text book thing - stiffen the rear compared to the front. Now I only have the sports suspension, not the full racing suspension, but I notice immediately that the car is set up with more stiffness at the front than the rear - okay, then thats why its understeering right? I also notice that the limits available here actually allow more front stiffness (max 5.5) than rear stiffness (max 4.5), this seems a little odd unless this is reflecting the tendancy to make a production car understeer for safer braking.

Anyway, I alter the settings within the relatively small range available and make the rear stiffer than the front - not drastic, but opposite to how it started out. I take the 300SX back onto the course and get even more understeer - the first hairpin I go into the wall. After a lap I settle into it and it seems less noticeable, but there is definatley more understeer than there was.

I go back into settings and set the springs evenly front and back, this improves the handling but still understeering quite badly and not as good as when it started. I play with the set-ups some more and find that there is no suspension setting which will fully eliminate the understeer let alone promote any oversteer.

Back to the drawing board. For the purposes of this discussion it would perhaps have been helpful to buy the full suspension set up and try the more extreme settings which that would allow. But I started thinking that what I ultimately need is more grip at the front of the car...maybe I could get it from downforce?

Now this shouldn't really work because the downforce gained from adding wings and spoilers should only affect the car when air is passing over the body - not when braking to an almost standstill in a hairpin. But I try it anyway. I get a simple wing and set more downforce in the front than the rear - the result is dramatic - I finally get oversteer (almost a little too much) all over the course and most notably in the hairpins.

So I get a couple of conclusions from this:

1. The suspension geometry physics look wrong

2. Changes in suspension geometry have quite subtle effects on the car (at least this car anyway)

3. Downforce out trumps suspension geometry for effective car balance

4. Downforce physics are not quite correct either!

 

[rk]

vBONESv, hello and welcome. Most participants have already drawn their conclusions at this point, it is however understandable that people still have observations to include. Forgive me if I seem overly inquisitive, somewhere along the trail I picked up the pointy little plastic pitchfork of "Devil's Advocate."

I go straight to the suspension settings to do the text book thing - stiffen the rear compared to the front.

Now you have already read the thread, so we know YOU know the importance of "incremental" changes. Incremental being one or two clicks, of course...

Anyway, I alter the settings within the relatively small range available and make the rear STIFFER THAN THE FRONT - not drastic, but opposite to how it started out.

Not drastic!? What do you consider drastic? Unbolting your transmission and putting it on backwards so you can drive around in 5 reverses and have tons of OVERSTEER?
So to understand correctly, you REVERSED your spring values and are attempting to draw some sort of sensible conclusion from the results...

1. The suspension geometry physics look wrong inconclusive, need much more data

2. Changes in suspension geometry have quite subtle effects on the car (at least this car anyway)subtle CHANGES have subtle effects, pronounced changes have unpredictible effects

3. Downforce out trumps suspension geometry for effective car balance or possibly puts more downforce over the rear wheels which were overstiffened, lessening their tendancy to slide off the track which can cause the sensation of understeer

4. Downforce physics are not quite correct either!

backatcha

 

[v BONES v]

yikes! Guess I'm in the mix now...

First off I'd like to point out that this is just a simple one off experiment which draws a particular conclusion, I never said I had it figured out - so inconclusive is probably the right word.

What I meant by saying that settings were not drastic was that I didn't go to extremes on the settings (ie. one end opposite to the other) I just put them into a sensible looking configuration (rear stiffer than front) with a small difference between them (one or two clicks) - I was aiming to go to that set-up anyway, just happens that it started out the other way around which appeared wrong. Ultimately, the "real world" setting (stiffer rear to promote oversteer) worked the wrong way around and I found no setting either way which sucessfully addressed the understeer.

Since I didn't have full race suspension the range of movement possible was much reduced anyway, so any change would be quite subtle...so, fair enough, I haven't yet seen what a more extreme race setting might do to the car yet.

Not sure I understood your reasoning on the third point, what I got was oversteer - which was exactly what I was looking for, and the configuration of the downforce (more front than back) worked exactly the way I assumed it would. It just shouldn't really have been as effective as it was, especially in the hairpins where there should have been no extra downforce at such low speed?

 

[rk]

Not sure I understood your reasoning on the third point, what I got was oversteer - which was exactly what I was looking for, and the configuration of the downforce (more front than back) worked exactly the way I assumed it would. It just shouldn't really have been as effective as it was, especially in the hairpins where there should have been no extra downforce at such low speed?

A valid observation about the low speed downforce, I have found similar results in regard to camber changes as relates to stopping distances (and acceleration patches!) Apparently it is difficult for the physics engine to determine the transitive state between driving straight and cornering and since cornering is "the essence" of the Real Driving Simulator, the physics engine, when in doubt, defaults to turning dynamics. Bottom line is that good cornering camber also gives the best acceleration and braking.
What I was implying with my final point is the possibility that you overstiffened the rear, causing it to lose traction- which can resemble understeer as the rear slides out of a corner and you steer the front to "follow"- then you would or could have compounded the confusion by counteracting this with added rear downforce...
I am always considering the uncounted masses who just read these threads without contributing, so I try to "clarify" when I can. Subtlties and tact are great for statesmen, but they tend to not elicit information; thanks for being a sport about it.

 

[v BONES v]

hmmmm...interesting (makes note about camber settings).

Forgive my ignorance on camber, but I assume that good cornering camber is when the tyre is angle into the car at the top (is that positive or negative?) and the tyre will effectively try to "right itself" under load when cornering?...and this would then imply that you would lose traction when not cornering due to tyre patch size and scrubbing?

Still not sure about overstiffening the rear (sorry). If I lose traction at the rear when cornering, how can it lead to anything other than oversteer? Steering the front to follow is just the corrective action for overstreer isn't it? (God, I hope there's not another huge swathe of information I'm missing!)

 

[Madd Dog]

Hi, couldn't help but join in. I'm one of the uncounted masses as referred to earlier and find suspension settings etc really bewildering I know its not supposed to be easy but I'm a Gran Tourismo veteran and still havent mastered the art of fine tuning. I tend to make small ride height adjustments and stiffen or soften the stabilisers to my preferance, only recently have I started using different camber settings. Somtimes I only use a sports suspension upgrade on stock cars so as not to get myself totally lost.

 

[v BONES v]

Right with you there Madd Dog, welcome to the fray!

I also played all the previous incarnations of GT without tuning (well, just a few abortive attempts) but I feel inspired - particularly by the wealth of information presented here - to try a little harder this time around.

The 300ZX I was messing with last night is now a much better car, very stable and a lot of fun to drive - although I suspect I got a little lucky, GT4 appears to have a few unexpected surprises in store for determined tuners!

 

[Madd Dog]

Definatly.there are several cars the game that seem to suffer from the most horrendous understeer, no matter what tweaks I make, be they large or small, I can never seem to cure them. My last attempt to solve the problem usually comes in the form of adding downforce to the front to increase the front end grip. No doubt this isn't the best way to cure it.

 

[v BONES v]

It might not be the way you would go about it in real life...but its what works in the game which is important. Wing downforce should be variable dependant on airflow (speed = downforce) and should cause drag, but GT seems to give you a fixed amount of downforce regardless of speed (certainly on the few cars I have tried it on) and that means its a win - win option that you would be crazy not to take - even if it does cost a little drag.

Thats what I'm going with for now anyway - well, until something completely different happens on another car!

 

[Madd Dog]

Yea, but the downforce doesn't seem to affect the straight line speed, well, in my case anyway. I tend to increase the downforce more on the front than on the rear, but with obvious effects to the grip of the rear of the car through fast corners. I really need serious advice on setting the suspension up.

 

[v BONES v]

...even better - no drop off of downforce with slowing and no noticeable drag? ...just leaves improved grip at all times!!!

Have you checked through Greyout and Mspec suspension tuning stickies?

Takes a couple of reads but makes a lot of sense once you get your head round it. Problem is that GT4 doesn't always seem to implement suspension tuning correctly - but at least with the theory covered off you know where you're trying to get to and what sort of settings should work.

 

[Madd Dog]

once I get my ps2 back from my b-spec driver I'll be getting stuck right in. I have a very nice Toyota MR2 fully upgraded with around 420bhp to play with, it is just very difficult to get all that power down, leads to a lot of frustration I can tell you. Haven't even used it in an event yet because I'm not 100% convinced on the tyre wear rate. Em, what are the "stickies" you mention?

 

[DrBiggly]

...
3. Downforce out trumps suspension geometry for effective car balance or possibly puts more downforce over the rear wheels which were overstiffened, lessening their tendancy to slide off the track which can cause the sensation of understeer

RK,

Looking at this statement I have to ask: Do you believe that there is a point where in adjusting a rear springrate that as it goes up, it starts causing less oversteer instead of continuing to cause more? Is that what you are trying to say here?

-Biggly

 

[v BONES v]

try these:

https://www.gtplanet.net/forum/showthread.php?t=56645

https://www.gtplanet.net/forum/showthread.php?t=57625

Haven't got too far yet but I was thinking earlier that there could (should) be a point where adding horespower will do you no good because you won't be able to convert it all - even at best you would still lose grip as the tyres go off faster.

In GT3 I tended to mod everything up to the hilt, lower everything, stiffen everthing and then deal with the consequences. I'm trying to be a little more purist with GT4 (at the moment) and trying more subtle changes whilst keeping everything balanced. I'm quite pleased with the way my cars have turned out so far considering how little I have spent on them!

 

[v BONES v]

3. Downforce out trumps suspension geometry for effective car balance or possibly puts more downforce over the rear wheels which were overstiffened, lessening their tendancy to slide off the track which can cause the sensation of understeer

Ah....I think I see where you're coming from now. The added downforce at the rear gives more grip and therefore works against oversteer.

Probably a good point seeing as how GT4 seems to give permanent extra downforce at all speeds when wing angles are increased. But as I said, what I actually got is the oversteer I was looking for. I didn't experiment too much after that point, but I had the distinct feeling that I could control under and over steer very well by simply changing the downforce setting - no other adjustment necessary - and I could equally get the car to have undriveable under or over steer as well

 

[Madd Dog]

I'd agree there, I too usually use the downforce settings to counter oversteer/understeer as it stops me from having to adjust the suspension settings. Are the default settings any good, or should you steer clear of using them and make adjustments?

 

[rk]

Looking at this statement I have to ask: Do you believe that there is a point where in adjusting a rear springrate that as it goes up, it starts causing less oversteer instead of continuing to cause more? Is that what you are trying to say here?

I apologize for my ambiguity, it is so difficult for me to be clear AND concise. For the record, I currently do not believe that there is a point while adjusting a rear springrate upwqards, that it will start causing less oversteer instead of continuing to cause more. BONES actually made the most sense from my statement:

Ah....I think I see where you're coming from now. The added downforce at the rear gives more grip and therefore works against oversteer.

Thereby masking the oversteer caused by too stiff rear springs.
BONES: right about camber, so far as I know, all camber is considered negative (top in), and when suspension loads while cornering the "tilted" tire becomes prependicular to its contact patch, so it should grip better only then, but I can share verifiable experements that show otherwise.

 

[Mr Whippy]

Wow, this thread still going

All I know is, GT4 feels great.

Being doing lots of testing in fwd cars and leaving them stock with the N1 tyres (as sim as possible).

No driving aids. Really realistic feeling 90% of the time.

However, longitudinal grip at low wheels speeds. Hmmmm, any fwd car seems to have NO grip. My real car can be planted in 1st gear even in a slight damp and won't spin until 5000rpm ish, even when rolling it won't spin up.

Do the same in the car in GT4 and it just spins all over, doing exactly the same thing. Even S2 tyres do the same.... really poor.

Brakes, do we have ABS or not. GT4 is so iffy it's unbelieveable... Some cars like the Ford GT understeer at the slightest hint of brakes while turning, yet other cars can be driven right into the apex with full brakes almost!

Then there is this oversteer understeer problem. After LOTS of testing myself I find the cars very unlikely to react to real tuning alterations. The main problem is that when we add racing suspension, we get damper rates in unit-less 0-10 notches. These mean NOTHING, and for all I care have little real effect on anything.
Even the lowest spring rates are TOO stiff with racing suspension, sending most cars flying off to the outside as you hit the rumble strips on the apex.

To be honest, having NO telemetry readouts of camber and suspension deflection, damper shaft speeds and wheel loadings, as well as roll-centre adjustment, the whole game is impossible to really fathom.

I mean, 206 RC, N1 tyres, standstill, apply full right lock, floor the throttle, oversteer ensues!!! What?

Do this in my 306 Gti-6 (same kinda car and suspension setup and power etc), and the car understeers until I lift off.

Now, GT4 is flawed. It clearly has a really poor low-speed tyre model, and has major problems with weight transfers. I think spring stiffness and roll-bars are ignored, and GT4 uses a combined roll-stiffness per end of the car and leaves it at that.

Also, I've never seen the Clio 172 or Focus RS or 206 RC lift an inside rear wheel when turning hard.

GT4 feels authentic but is so far from it, you can't honestly expect real tuning changes to do what they do in real life in GT4.
They might to a degree, but I think what tuning there is is bodged in...


Telemetry would prove everything, but notice how they don't put it in! No real car tuner would alter toe and camber and anti-roll bars without telemetry to go on, especially damper tuning for race cars!

Such a nice load of artwork and content ruined by a simplified physics engine. No better than PGR2 in alot of ways, just bodged to feel more real than arcadey, but still hugely arcadey and flawed compared to say GTR for the PC etc...

Dave

 

[sukerkin]

I take your points, Dave. I do agree that there is something amiss in the physics engine when it comes to trying to turn 'road' cars into 'race' cars.

In general terms tho', 'real world' tuning approaches seem to follow the rules, it's just that the emphasis given to some adjustments seems underplayed or non-existent.

My personal theory is that the physics engine is actually coded to be more complicated than it is currently implemented in GT4 i.e. there are parameters that have been made fixed that should be variable or have been left as null and so give rise to the "Huh?!" factor .

For now, I've stopped putting the Racing Suspension on my cars (unless I'm also putting on a Roll Cage and Wing) as I've found that the Semi-Race works quite well enough with cars in the 300-500 HP range (plus I don't have to spend two hours spannering it to get it to behave ).

 

[rk]

My personal theory is that the physics engine is actually coded to be more complicated than it is currently implemented in GT4 i.e. there are parameters that have been made fixed that should be variable or have been left as null and so give rise to the "Huh?!" factor .

.

That is a very shrewd guess, sukerkin. The fact that the game won't even run without a memory card to suppliment the onboard RAM, that they gutted the online implementation and the reluctance to support any "real" open cockpit racing indicates there are resource issues. I think they use some sort of dynamic allocation of resources and that is why my screen jitters at the start of open cockpit races and my cars drive so differently in 2 player mode, because in those circumstances my resources are all used up. (Would it run better if I got a bigger memory card? j/k)
It would be interesting if our theories dovetail, when PS3 is released, all they would have to do is write a new cover and call GT4 a vastly improved release.



...

 

[Zardoz]

...It would be interesting if our theories dovetail when the PS3 is released...

Bet the house on it.

I'm convinced that most of the problems we have with the game are caused by the fact that we're actually running it on the wrong platform. We're running a disabled version of what it will be when it is revised and opened up for the PS3.

 

[Tornado]

I have a interesting story to tell. I took a Honda CRX (stock) and lowered it fully and put the spring rates highest settings on both front and rear. I went to twin ring Motegi West course and went into the first turn. It handled very neutral (actually alot like a NSX) into the turn (I did all of my braking before I entered the turn) and when I applied throttle coming out of the turn (right after the apex), the right front wheel started spinning and the car went into a full spin towards the right. Before I knew it the car was in a drift that I would say was 30 degrees or so and even after I lifted off of the throttle the car would not straighten out. I had to hit the brake and turn left for it to stop spinning. The LSD was set to 3 so I know it wasn't why it happened.

 

[djaft3rb3ats]

great, off topic? kinda? lol

 

[nos2]

I tried out the down force theory, it worked pretty good with the Chevy SSR. I have race suspension, front spring rate at default, rear to full soft, and front downforce 30 and rear df 0, it works well. I think SCEI definitely fkkked up the suspension modeling.

 

[Gedscho]

this is the thread the made me read for hours and hours... this is the thread that made me sign up on this forums... this is the thread that probably contains the most useful information about gt4, or at least so i think.

i am wondering... why did it die?
is it resolved? did people get bored?
(if you want to hear my guess: i think its like some sort of knowledge of black magic. dont talk about it. dont think about it. let it disappear in the unknown. thats not really bright or scientific, but it might just keep us all sane and alive )

above all the questions this discussion raised inside my (already smoking) brain i have ONE, that needs to be answered, before i can go any further...

DS2 gets some additional driving aids you cant control in any option menu, compared to the ForceFeedbackWheel especially at low speeds.
this may throw a lot of test results outta the window.

correct?

 

[sukerkin]

Hello Gedscho and welcome aboard the sometimes stormy ride that is GTP .

As to the thread, altho' no formal conclusion was ever written within it's scope I believe that if you asked those of us who enjoy delving into the Suspension Arts, the majority would say that the suspension tuning in the game was not backwards but that it does contain some biases that you have to be aware of.

For example, it appears that the game is very front-end biased (by which I mean that changes done to the front suspension have a greater effect than those done to the rear).

Also, camber changes have a much reduced effect in the game than they would do out on the real world track.

Further, because the PS2 does not have the computational wherewithal to actually work out suspension motions in real time, then everything you see is a realistic feeling approximation rather than a true simulation. That's what can lead to people concluding something is wrong. As an example, such an occurence may be when you wind up the spring strength on the rear of a MR car and it appears to reduce oversteer rather than causing more. That's because spring rates influence lateral and longitudinal roll stiffness as well as compression resistance and, by helping keep the car flat in all dimensions, smooth out the oversteer rather than having it hit in one unpredicatable lump (or at least that's one possible explaination).

As to the effects you mention with the DS2, one thing I have read elsewhere ios that it is believed that there is a smoothing algorithm in the steering. Also, not just restricted to the DS2, there are 'limiters' on lift and roll that stop the cars turning over and these can have an intefering effect on suspension turning, giving false impressions, particularly when extreme settings are used.