What is up with the F301?

[HolyDiver]

Hey all!

I am struggling heavily with the F301 car. I recently installed AMS and did some testing with this car, but it's not really going well so far. The car's rear end snaps away in every corner at every speed the moment I turn in. Now as I recently installed the game and haven't grown accustomed to the physics and the FFB yet, I am not sure whether I have setup my wheel and the FFB right. Should I enable automatic rotation in-game? The car has a default rotation of 360 degrees with 18 steering lock, is the latter too high maybe?

 

[Spinelli]

I just had a look at @Spinelli's video. Correct me if I'm wrong, but it appears the video shows two different kinds of slides. In the first kind the driver is able to hold the slide for a long time while applying lots of opposite lock and modulating the throttle. He seems to be able pull it off with great ease and he is fully in control the whole time. Personally, I have never been able to do that in AMS. I have had thousands of slides, but they never looked like that. Whenever I slide in this game the car is just moments away from snapping in one direction or the other. It feels very on/off. I'm not saying it's difficult because a quick lift of the throttle combined with a very quick countersteer and immediate return to the center usually ends a slide instantly, but it's not progressive at all. On the rare occasion that I am able to hold a slide for a slightly longer time it is usually with the wheel almost centered and with very careful modulation of the throttle.

Thank you. I have been playing this game (physics engine) for 1.5 decades and have raced against hundreds of others and watched thousands of videos of worse and better drivers than me and "aliens" and never see that behaviour.

The second kind of slide is the one where the driver applies full throttle or even increases throttle while countersteering. This is definitely possible in AMS. I have done it countless times. Here is a quick video of me performing such slides. Slides are at 00:32, 01:22 and 01:47.

Yes, the slides you have shown in your video are definitely possible. I've done those tons of times too. Although I do feel even those types are still too aggressive as you go up to cars like F3s, F1s, IndyCar, etc. Usually what happens in those cars is that, while keeping the throttle or even adding throttle, the revs and wheelspin will want to rise. Also, they are usually needing a very quick snap correction. Notice in the video that posted sometimes the revs barely rise (tyre's "claw" effect even when there is wheelspin) and it doesn't always have to be a very quick correction because the rear starts oversteering slowly....If you try the F2000 or F3 in Netkar Pro and even Live For Speed, you can get results like this that are so similar to real life it's almost scary.

I believe if physics engine coding gets seriously taken a look at and really improved, both types of slides will get massively improved because these issues may be all related.

Often my videos are further tweaks or attempts at something physics wise, making them different from the release physics. Having a DD wheel with AMS is really good though, it is 75% the FFB that tells me about the opposite lock, I mostly just let it happen and control the last little bit of input. It feels like that anyway.

The physics engine is really solid and we have not found any bugs. I recently also managed to work out how force combining works (cornering+braking or cornering+accel) and the tire model basically seems to work as you would expect a good tire model to do.

So every single other simulation has it completely wrong? ISI created the first physics simulation software that has no weak areas?

I'm sorry Niels, I love your work, I love Reiza, and I love your company (recently bought a shifter and service was fantastic) but I don't buy that everything in the ISI engine is just spot on behavior. Every sim physics engine has their own peculiar traits, good and not-so-good.

- Apart from what I've been mentioning in previous posts, there's also a tendency to loose grip at very low speeds with very little steering lock. I was driving the F3 and the car's rear wanted to slide out when I had around 20 or 30 degrees of steering lock, no throttle, and going literally around 65 Km/h. It's like the physics engine and/tyre model is in a rush to put a car into a low grip state. Like it's too easy to get a car into a sliding state. Grip is overly sensitive to being lost and, on top of that, grip is overly obedient in staying in a sliding state once grip is lost unlike in real life (and some sims as previously mentioned) where the car/tyres seem as if they hate sliding so they stick out their claws to try and grab the tarmac during a slide.

- Also, there's a tendency for cars to want to continue turning and turning in what seems like an ever decreasing radius by just holding the steering at very little steering angles then holding on very small amount of brakes. It's like an invisible hand comes down from above and grabs the car and makes it turn more and more. Most "aliens" and top drivers really take advantage of this phenomenon and you can really see it if watching them drive. It's slightly improved from rFactor 1 and before sims (eg. F1 2002) but it's largely still present and even in rFactor 2. I remember driving a NASCAR car in RF2 with LITERALLY around 1/4 or 1/5 the steering lock they use in real life and I was driving with an even SLOWER than default steering ratio all this while driving pretty competitive lap times too.

- Also, if I spin, at least with my OSW, the game makes the wheel snap all the way to the locks and there's a constant force that keeps turning my wheel almost as if there's another person grabbing my wheel and massively resisting me from being able to return the wheel to centre. I can barely hold onto the wheel if I get into a spin. I have to let go as if I got into an accident sometimes. Just spinning wants to snap my wrists. In real life, heck, you can even hold onto the wheel with one hand, easily, if you spin. This unrealistic and overly aggressive FFB behavior does actually remind me of some of the unrealistic and overly aggressive slide behavior. Maybe there's a relation. It would be quite the coincidence if this FFB behavior had no relation to physics, I believe they're related.

- Also, many of the top drivers in the ISI physics engine tend to run very rearward brake bias like 52%F, 53%F, 54%F. These low biases seem to give an overly amount of advantage relative to real life. In real life, I personally see most drivers running like 56 to 59 (generalizing here) from F1 to everything I've heard about other series and also talking with multiple real life racers myself. This is something that has been going on in the RF engine since the early 2000s. I believe this in combination with the phenomenon I explained above (the NASCAR example) are areas many top drivers exploit.

I'm sorry but to say everything is all roses and blissful, I'm sorry, I just don't buy it unless I'm playing a different game than you.

I don't know how any one can drive with all these things going on and not think there are areas of coding that need to be looked at. There's probably some deep-rooted combination of algorithms and/or formulas buried in the core EXE physics engine and/or tyre model core coding that don't get along, or at least don't get along during certain combinations of particular states of vehicle dynamics, kinematics, slip angles, slip percentages, etc.

 

[Marc Collins]

- Also, if I spin, at least with my OSW, the game makes the wheel snap all the way to the locks and there's a constant force that keeps turning my wheel almost as if there's another person grabbing my wheel and massively resisting me from being able to return the wheel to centre. I can barely hold onto the wheel if I get into a spin. I have to let go as if I got into an accident sometimes. Just spinning wants to snap my wrists. In real life, heck, you can even hold onto the wheel with one hand, easily, if you spin. This unrealistic and overly aggressive FFB behavior does actually remind me of some of the unrealistic and overly aggressive slide behavior. Maybe there's a relation. It would be quite the coincidence if this FFB behavior had no relation to physics, I believe they're related.

BINGO. Among other things, you have a seriously misadjusted wheel. The rest of us do not and experience reasonably realistic forces under all circumstances. If you have a proper wheel and settings, the forces lighten as you lose grip, exactly as they should to provide the feedback to gently bring the car back under control. In a slide situation, you do feel the tires clawing for grip. In a spin situation, depending on the car and tires, it can take you by surprise, or be completely predictable--just like in real life.

 

[Niels_at_home]

I'm not saying everything is spot on. I'm saying that during all the years of testing and analysis I have not found serious bugs. In rF1 I found an inertia bug which got fixed by ISI in our GSC / AMS version (probably first fixed in the Superleague game). This is WAY different from saying that there ARE no bugs and that all the models are perfect.

I've done some weird ass tests like driving the car on a big flat road, that halfway through turns into perfect ice, then lock the tires, and apply a hacked memory edited side force to the front wing. Then check MOTEC to see if the rotational acceleration matches what you would expect this force and inertia to do

Sofar all the time I suspected the engine to have some sort of bug, it seemed most likely that it didn't.

The problem is going from subjective 'this feels like that' to something physical like 'this force with this inertia results in incorrect acceleration'.

If rF2 does some of the same things, well that is a completely new rigid body system and tire model. It would surely be possible that Terrence made the same mistake twice but he is a really clever chap so there would be a good chance he would've spotted errors in the rF1 era of system and fixed it in rF2. From what I have heard, the rF2 rigid body system is clever with chassis flex etc, but when not using it, virtually identical to the rF1 system in results.

As far as I have the ability to know, all the forces on the car work as expected and generate the movement and acceleration that is expected. The tire model can do a decent job compared to other theoretical models, although that is parameter dependent and there isn't a lot of tire data available to be sure.

Could be worse though, its not iRacing at the limit.. ;-) ;-) ;-)

 

[Marc Collins]

I'm not saying everything is spot on. I'm saying that during all the years of testing and analysis I have not found serious bugs. In rF1 I found an inertia bug which got fixed by ISI in our GSC / AMS version (probably first fixed in the Superleague game). This is WAY different from saying that there ARE no bugs and that all the models are perfect.

I've done some weird ass tests like driving the car on a big flat road, that halfway through turns into perfect ice, then lock the tires, and apply a hacked memory edited side force to the front wing. Then check MOTEC to see if the rotational acceleration matches what you would expect this force and inertia to do

Sofar all the time I suspected the engine to have some sort of bug, it seemed most likely that it didn't.

The problem is going from subjective 'this feels like that' to something physical like 'this force with this inertia results in incorrect acceleration'.

If rF2 does some of the same things, well that is a completely new rigid body system and tire model. It would surely be possible that Terrence made the same mistake twice but he is a really clever chap so there would be a good chance he would've spotted errors in the rF1 era of system and fixed it in rF2. From what I have heard, the rF2 rigid body system is clever with chassis flex etc, but when not using it, virtually identical to the rF1 system in results.

As far as I have the ability to know, all the forces on the car work as expected and generate the movement and acceleration that is expected. The tire model can do a decent job compared to other theoretical models, although that is parameter dependent and there isn't a lot of tire data available to be sure.

Could be worse though, its not iRacing at the limit.. ;-) ;-) ;-)

@Niels_at_home, being a wheel and FFB expert, what could explain @Spinelli getting what seems to be reverse or greatly exaggerated forces in some circumstances?

 

[Niels_at_home]

I don't know for sure, I never spin.... ;-)

We've had some suspension geometries that didn't like full lock steering as much, which could result in larger FFB forces near the lock stops. I'm also not 100% sure how many cars ended up using the new PneumaticTrailGripFractPower parameter, which makes self aligning torque drop off more realistically.

The downside of a product that is in development for like 8 years is that there will be some variance between older and newer cars as we got better and implemented more stuff..

 

[Erwin Sch]

Also, on exit under power, be sure to mostly only induce wheelspin and play with the limits when you're almost pointed in a straight line. Inducing wheelspin while still turning usually ends up with a sudden snap which requires a snap-correction and on/off "digital" wheelspin. It's almost impossible to induce wheelspin while maintaining and manipulating the car via wheelspin.

I agree with this part, in regard to most formula cars. Whether it's driver skill, settings, hardware or physics, I don't know, but it seems in real life wheelspin is easier to modulate.

Although it differs per car. The Formula Retro for example is much easier to control with wheelspin on exit. The slides I can do with this car look similar to the slides in that real life video of F3 slide examples.

The Formula V10 (with TC off) on the other hand revs up so quickly that it means your either under the limit, or at red line revs. It's too dangerous to play with wheelspin on exit of a slow corner, risking the car snapping 90 degrees left/right.

Maybe the torque mapping could be adjusted for some cars, so there will be less torque at higher revs: better control of wheelspin? F1 drivers even have a dial for this on their wheel.

Apart from this, formula cars are very nice to race in AMS

 

[Spinelli]

I agree with this part, in regard to most formula cars. Whether it's driver skill, settings, hardware or physics, I don't know, but it seems in real life wheelspin is easier to modulate.

Although it differs per car. The Formula Retro for example is much easier to control with wheelspin on exit. The slides I can do with this car look similar to the slides in that real life video of F3 slide examples.

The Formula V10 (with TC off) on the other hand revs up so quickly that it means your either under the limit, or at red line revs. It's too dangerous to play with wheelspin on exit of a slow corner, risking the car snapping 90 degrees left/right.

Maybe the torque mapping could be adjusted for some cars, so there will be less torque at higher revs: better control of wheelspin? F1 drivers even have a dial for this on their wheel.

I agree with basically every word you said. The instant wheelspin/revs shooting to redline instantly - as if the car lost absolutely 100% of it's grip in combination as if the car suddenly gained a HUGE amount of power is very prevalent especially as you get to more racey cars. Yes, it's often too dangerous to really be manipulating the car and hustling it around the track due to this and some other physics anomalies in the engine.

I don't know if it's as simple as just a torque map problem. That would be too obvious and too easy a fix and I'm guessing that would have been fixed many, many years ago if the problem stemmed from something that simple (this physics engine has suffered from this since I first started playing one of it's earliest versions in F1 2002 - almost 20 years ago). On top of all that, you can watch F1, F3000, F3, GP2, IndyCars, etc. from before they used all this torque mapping trickery (especially to the extent it's used today). However, until the problem is truly fixed, maybe manipulating torque at certain revs as you mentioned can at least be a temporary "band-aid" fix to at-least reduce the severity of the issue?

BINGO. Among other things, you have a seriously misadjusted wheel. The rest of us do not and experience reasonably realistic forces under all circumstances.

I have an OSW. It's done this regardless of OSW firmware, regardless of old OSW style, the newer simucube "mode" (for non-true simucube users), and true simucube (simucube hardware) mode (what I currently have). Regardless of 5,000 PPR encoder or 10,000 PPR encoder.

If you have a proper wheel and settings, the forces lighten as you lose grip, exactly as they should to provide the feedback to gently bring the car back under control.

They do lighten as I loose grip. In this spin scenario I'm describing, the force lightens so much that it actually starts applying a TON of force in the opposite direction (opposite lock).

In a slide situation, you do feel the tires clawing for grip.

There is no claw affect and, also, I'm not even talking about feel, I'm talking about pure physics regardless of feel. In real life and some other sims, you can get wheelspin and often times the naturally clawing affect from the tyres prevents the wheelspin from increasing and sometimes even ends the wheelspin on it's own. This is easily observable by watching videos all over the web and it's done quite well in NKP and LFS but in RF/ISI/AMS physics, the majority of the time, the rear wheelspin will just keep on increasing (unless you, yourself, lifts off the throttle) as if every last tiny ounce of grip was lost or as if your car suddenly gained tons more Hp. There is no "clawing" effect which makes the natural grip of the tyres/car want to stop or at least slow down the wheelspin.

In a spin situation, depending on the car and tires, it can take you by surprise, or be completely predictable--just like in real life.

Yes, I know that some spins can take you by surprise. I myself have been caught in a couple snap-spins in real life F2000 racing. One where I was leaning hard on the right-side tyres during a fast left...my right-rear touched a bit of grass on the outside and it snapped very suddenly because I was already laterally leaning on the right tyres very hard. And I had another one in the wet that was similar but involved hitting a piece of track that was much more wet than the other parts I was on before. In bot those examples, I was already leaning on the tyres very hard and already extracting close to 100% grip from the tyre prior to suddenly driving over a much more slippery surface.

Here's a video I made showing the overpowered opposite lock. Note, I am purposely making the car spin but, even so, notice how every one is an insanely sudden snap almost as if at the very moment the car looses grip the game says "I don't care what inputs the driver makes, I've already decided I will put the car into an instantaneous spin and there's nothing the driver can do about it. I will ignore any and all driver inputs." Anyways, I made this video to show the over-powering opposite lock effect, not actual physics stuff (but AMS' FFB is almost purely based on physics so I believe there's got to be some relationship here).

 

[Marc Collins]

I agree with basically every word you said. The instant wheelspin/revs shooting to redline instantly - as if the car lost absolutely 100% of it's grip in combination as if the car suddenly gained a HUGE amount of power is very prevalent especially as you get to more racey cars. Yes, it's often too dangerous to really be manipulating the car and hustling it around the track due to this and some other physics anomalies in the engine.

I don't know if it's as simple as just a torque map problem. That would be too obvious and too easy a fix and I'm guessing that would have been fixed many, many years ago if the problem stemmed from something that simple (this physics engine has suffered from this since I first started playing one of it's earliest versions in F1 2002 - almost 20 years ago). On top of all that, you can watch F1, F3000, F3, GP2, IndyCars, etc. from before they used all this torque mapping trickery (especially to the extent it's used today). However, until the problem is truly fixed, maybe manipulating torque at certain revs as you mentioned can at least be a temporary "band-aid" fix to at-least reduce the severity of the issue?

I have an OSW. It's done this regardless of OSW firmware, regardless of old OSW style, the newer simucube "mode" (for non-true simucube users), and true simucube (simucube hardware) mode (what I currently have). Regardless of 5,000 PPR encoder or 10,000 PPR encoder.

They do lighten as I loose grip. In this spin scenario I'm describing, the force lightens so much that it actually starts applying a TON of force in the opposite direction (opposite lock).

There is no claw affect and, also, I'm not even talking about feel, I'm talking about pure physics regardless of feel. In real life and some other sims, you can get wheelspin and often times the naturally clawing affect from the tyres prevents the wheelspin from increasing and sometimes even ends the wheelspin on it's own. This is easily observable by watching videos all over the web and it's done quite well in NKP and LFS but in RF/ISI/AMS physics, the majority of the time, the rear wheelspin will just keep on increasing (unless you, yourself, lifts off the throttle) as if every last tiny ounce of grip was lost or as if your car suddenly gained tons more Hp. There is no "clawing" effect which makes the natural grip of the tyres/car want to stop or at least slow down the wheelspin.

Yes, I know that some spins can take you by surprise. I myself have been caught in a couple snap-spins in real life F2000 racing. One where I was leaning hard on the right-side tyres during a fast left...my right-rear touched a bit of grass on the outside and it snapped very suddenly because I was already laterally leaning on the right tyres very hard. And I had another one in the wet that was similar but involved hitting a piece of track that was much more wet than the other parts I was on before. In bot those examples, I was already leaning on the tyres very hard and already extracting close to 100% grip from the tyre prior to suddenly driving over a much more slippery surface.

Here's a video I made showing the overpowered opposite lock. Note, I am purposely making the car spin but, even so, notice how every one is an insanely sudden snap almost as if at the very moment the car looses grip the game says "I don't care what inputs the driver makes, I've already decided I will put the car into an instantaneous spin and there's nothing the driver can do about it. I will ignore any and all driver inputs." Anyways, I made this video to show the over-powering opposite lock effect, not actual physics stuff (but AMS' FFB is almost purely based on physics so I believe there's got to be some relationship here).

Hopefully others will chime-in, but unless you have set your forces for normal steering as an alternate to weight lifting at the gym, you have a seriously misbehaving wheel. Somehow, some forces are reversed or magnified. Can you confirm that your normal steering effort is reasonable and "normal?" A properly configured wheel will not behave in that extreme manner.

 

[Spinelli]

Yes, my steering is set not that high. In fact, I had to lower it from what I wanted because of the insanely over-powerful as well as over-reactive/accelerative opposite lock forces putting massive strains on my wrists (not to mention way overly powered return-to-centre / S.A.T. spring forces but that happens in all racing sims with DD wheels so I think that's got to do with FFB technology and games in general, not specifically AMS).

 

[Marc Collins]

Yes, my steering is set not that high. In fact, I had to lower it from what I wanted because of the insanely over-powerful as well as over-reactive/accelerative opposite lock forces putting massive strains on my wrists (not to mention way overly powered return-to-centre / S.A.T. spring forces but that happens in all racing sims with DD wheels so I think that's got to do with FFB technology and games in general, not specifically AMS).

So you are saying all sims produce exaggerated forces with DD wheels? I can see that separate programming threads would be required for DD versus wheels with mechanical friction (we know how differently the different levels of internal friction feel with the same game settings), but surely everyone with DD wheels is not putting up with these unrealistic forces? Why would anyone spend so much extra to get worse FFB?

I hope @Niels_at_home can advise.

 

[Spinelli]

So you are saying all sims produce exaggerated forces with DD wheels? I can see that separate programming threads would be required for DD versus wheels with mechanical friction (we know how differently the different levels of internal friction feel with the same game settings), but surely everyone with DD wheels is not putting up with these unrealistic forces? Why would anyone spend so much extra to get worse FFB?

I hope @Niels_at_home can advise.

I don't think all games have the insanely over-powered counter steering force which snaps to full lock with such ferocios instant acceleration and torque like in the AMS examples - I think that's maybe more connected to the physics which seem to do something similiar. Having said that, most sims do have general over-powered and over-reactive forces with DD wheels like massive return-to-centre "springs" as well as overly-instant opposite lock forces.

I'll never forget one of the first review videos of a DD wheel I watched. It was a guy driving iRacing just raving about the OSW (as 99% of people do). He was getting a Skip Barber car sideways through the final corner of Sebring and the DD wheel would start correcting the car so instantly that the driver would consistently fully let go of the wheel and let the DD wheel "automatically" save every slide on it's own. He was trying to show how good/realistic the DD wheel was. While every one was wetting their pants about the DD wheel's greatness, I was practically face-palming myself in regards to the DD wheel's unrealism. Ironic how he was using such unrealistic behaviour when trying to showcase the wheel's realism.