Yes,

I was known as the ISI hater, but I converted, ish. At least there was more possible than I thought; even though its not perfect.

I was also a big hater of force feedback. I still am ish. But much like the ISI story, I now went out and bought a G25 to see what it is made off.

I never liked force feedback because:
- wheels of dodgy build quality
- lots of internal resistance, can't turn quickly if you want it
- big gearing meaning it can't turn fast by itself either
- gears give notchy feel
- jittery unstable jerking forces
- low fidelity, not many distinguishable levels (steps) of force

The G25 is better in most areas, quality is good, internal resistance is much lower, spikes and stability are a lot nicer and it has a higher turning speed. But is it good enough?

What should FF be for me?
I only really want force feedback if its going to help me. In most cars, the suspension geometry and tire alligning torque mean that the front tires will want to keep rolling in the direction the car is moving. So when the back steps out, the tires keep pointing ahead and the wheel turns by itself. It also should straighten once a slide comes back and goes straight. Force feedback should be a driving aid that helps you keeping the car on the road at extreme conditions.

Force feedback should also give some fidelity of information regarding what is happening. Steering should be heavier when the car is pushed down on the road by road camber or track undulations. Steering should get lighter or heavier depending on suspension geometry and slip angle etc etc.. For this you would need a fair amount of 'dynamic range'; you should be able to feel small forces and large forces.

See the video:
http://www.youtube.com/watch?v=-XgQm_fITok

The only thing I change is the rotation angle used in the logitech driver.

Here's the deal:
I bring the car in an oversteer situation and let go of the steering, then see how the wheel and game fight it out. Ideally you'd see that the front tires are kept in the direction the car is going, so it should straighten out. Or at least, a big nod in that direction. Then I do a donut, and see if it gets out of it by itself or not, in general taking notice of what happens.

900 mode
At 900 degrees in both LFS and rFactor you can see the car spins out every time I let go of the wheel. You can see it *tries* to keep the tires rolling in the right direction but it can't keep up and the car spins out. Doing donuts, the same happens, both sims just stay spinning unless you really go easy on the throttle.

600 mode
Here you can see, especially in rFactor, that its closer. Yet still in both sims there is not a quick enough response and the car is out of control most of the time. It sill does not respond and the game is held back by the slow wheel.

300 mode
Again the similarity between sims is great: Now the wheel is fast enough! It keeps the front tires aiming where they should and now, when the first slide is anything smaller than a spin out, it willl catch the following ones until the car goes straight! Thats what I'm talking about! Donuts too, it easily straightens out and drives off. Now, when I let go of the wheel I can trust it will do what it is supposed to do; trying to keep the tires pointed in the direction the car is going.

Summary of modes..
At 600 and 900, the problem is that even though you let go of the wheel, there is still an unrealistic amount of ''friction'' and slowness. What happens when you let go is simply NOT what the sim wants to happen! Its a bigger handicap than it is an aid then! At 300 it works, but of course 300 is often 780 degrees less than these road cars would use in real life. It does however proof that the sims and software are ok, but that the hardware seems to be lacking still.

What about this fidelity and dynamic range thing?
How many steps of force feedback do you feel? In ISI, the default is 11500. In LFS I think 128, which can be increased to 256. HOWEVER!! that does NOT mean your wheel copies these fine steps. I wish I had a spare loadcell laying around so I could measure the actual fidelity of the wheel, which I will now do by feel. Wheel electronics, transmission 'drag' and resistance, these things compromise how many fine steps are actually left over at your fingertips. Using a force 'test' program, I can drag a slider per pixel each way and every pixel it moves a greater force is outputted to the wheel.

Now my hands can't measure 11500 or 128 steps of force. However, when slooowly moving this slider and keeping the wheel in one hand, I distinghuished 6 big steps of force.. six.. Most likely 'notches' where the cogwheels went to the next teeth. I'm not saying there are only 6 steps of force feedback. However, when there is a notchy feel, one tends to only feel these notches and eventual small steps in between these notches don't really register. I tried rFactor at a fidelity of 8 instead of 11500 and the experience was mostly the same.

So even the G25 is still very notchy and does NOT give this dynamic range of smooth FF that comes in plenty of small steps.

So I still dislike it?
Yes.. but at 300 degrees, for the first time ever, I actually saw that force feedback can work, which is quite a feat!

Interesting.

LfS went to 256 steps as default in one of the latest patches (not sure which one), doesn't invalidate your (very interesting) review in any way though...

I would have thought you'd get that in real cars with large steering angles too.

In a car with 2 and a half turns it would also have a fair bit more resistance if purely the friction of the wheels had to whack the wheel round half an entire lock in a second or two - which is why sports cars have much smaller steering wheel mranges lock to lock.

Quote from Crashgate3 :

In a car with 2 and a half turns it would also have a fair bit more resistance if purely the friction of the wheels had to whack the wheel round half an entire lock in a second or two - which is why sports cars have much smaller steering wheel mranges lock to lock.

Well the resistance in the steering is nearly all down to the tires and the road surface, which act irrespective of the amount of lock. The reason racing cars have very small steering angles is so that you get extremely responsive, but heavy, steering with very little movement required so changes in direction can be very quick. IRL in cars with lots of turns of lock it can be advisable in certain situations to let the steering wheel correct itself because just as with the motors in FF wheels you can only turn the wheel so fast.

interesting thread and to add my 2 cents, i found that 540° is the best setting for me and this for every isi game (gtl, gtr2, rfactor in particular). For lfs i always use the turn degrees of the cars (270° for the mrt, 720° for road cars and gt2 class, 450° for the fox and fo8, 400° for the bf1 and 540° for the gtr class)

Im also using 105% ff in the software while dumper and rebound is disabled and center spring is enabled but set to zero. In lfs i use 30% for the bf1 and for all the others between 30-40%. This way (in lfs) im almost always in the position to feel at which point i start losing grip or how the tires behave under load (weight transfer ect). Its not too much to fight against the wheel and not too less to have no feeling at all.

So in this respect, having more degrees enlarges the range of this response while having the disadvantage that its getting harder to catch the car at over- or understeer situations. As you can imagine, with less degrees youre able to catch the vehicle at almost every situation but the range of where youre able to "feel the grip" is getting very small which can produce even more situation where you get either over- or understeer.

Did you take in account the “wheel turn compensation” slider at LFS’s options?
The force feedback is ok for me. I can understand the limitations of the hardware and compensate with them. At least it is much better than nothing…

My first thought when I saw your video was 'That G25 looks very very slow'. Don't know why, but I feel my wheel is faster than that. Gonna go test it right now and see. I know in the LX6 I can let the wheel go and it sorts itself out. Also, be aware that not all cars will right themselves in all situations if you let the wheel go. The RX-7 I drive will often do exactly like in the vid when driven poorly. So don't blame the wheel for everything. Some cars are just too out of shape to be brought back with the wheel alone, they need weight transfer (throttle use).

I think force feedback helps more than it makes me slower.
What is faster than a wheel then?

You can easily steer while guided with the wheel.
G25 is far better than what I had earlier, a MOMO, and before that a non-FF steering wheel.

btw: the force feedback can change if you change the caster (BIIIG influence)

EDIT: #1100

Quote from BlakjeKaas :

btw: the force feedback can change if you change the caster (BIIIG influence)

EDIT: #1100

Front ARBs also affects how 'dead' the FFB feels.

Atle, I measured full left to full right, with a microphone.. (its a bit noisy ) and that was 1.5 seconds; which is 1.66 revolutions per second. Its unlikely that I have a 'slow g25' I think; you'd say production quality will be pretty equal from wheel to wheel..

Wheel turn compensation, I think, is some sort of non linearity in the steering. I wanted to keep the sims setup exactly the same and only change the range of the Logitech.

What should have happened was 900, 600 and 300 giving exactly the same results. The same thing is being done to the cars, the sim wants the same thing from the wheel... What exactly happens is less important to the fact that there are *big* differences. This is proof of the wheel not being up to what the sim asks from it.

Quote from Niels Heusinkveld :

Wheel turn compensation, I think, is some sort of non linearity in the steering. I wanted to keep the sims setup exactly the same and only change the range of the Logitech.

BTW, are you absolutely sure your Logitech profiler's FFB lock stop matched the car/setup lock exactly in correct degrees? And you were sure to disable all steering lock compensation and mumbo jumbo in-game too?

Just asking because no cars in LFS use 900-degree steering lock... or 600-degree... or 300-degree.

The game settings are exactly the same; all I varied was the logitech profiler steering lock. Nothing was changed in the game, just the wheel having either 900, 600 or 300 degrees of turning to try and do.

Yes, yes... but since the car inside the simulator has a certain steering lock depending on the setup and chassis or suspension shouldn't you match the virtual and hardware steering lock values to get accurate results?

No car in LFS has a steering lock of 900, 600 or 300 degrees unless limited by steering lock setting. Road cars max out on 720 degrees, GTRs at 540 and open wheelers were something like half of that if I remember correctly.

Or am I missing the point completely again?

Did you try changing the FF strength setting? It seems it should influence the result of the experiments with the wheel catching the slides on its own. When the FF is not strong enough the wheel has problems overcoming its internal resistance and turning quickly enough to keep up. Obviously, this is more likely to happen with higher angle modes because the wheel has to rotate quicker in such cases.

Quote from spankmeyer :

Yes, yes... but since the car inside the simulator has a certain steering lock depending on the setup and chassis or suspension shouldn't you match the virtual and hardware steering lock values to get accurate results?

No car in LFS has a steering lock of 900, 600 or 300 degrees unless limited by steering lock setting. Road cars max out on 720 degrees, GTRs at 540 and open wheelers were something like half of that if I remember correctly.

Or am I missing the point completely again?

I think the wheel won't "work properly" if you don't match the locks ingame and on the profiler. No idea if this has anything to do with the experiment though

LFS will and always will work better with 270 degrees or around there. I have noticed that you really cannot drive LFS to its limits when you use a higher steering lock because you'll be so much slower in catching the tiny little snappy slides LFS creates while you drive. You need a faster and more direct response by using a lower steering lock range in order to be the quickest.

Though I personally don't think the force feedback return-levels at a higher steering lock are at fault for why 720 degrees doesn't work well. Using 900 degrees isn't a good idea either, no car has that steering lock in LFS. 720 Degrees can feel fine, however the one reason I notice why rFactor gives better return feedback is because the steering is so rigid and the FF always feels rather fierce in that game. With or without the ReelFeel plugin, it doesn't matter, rFactor has always felt like it has a harder FF bite. I've liked LFS's FF feel simply because it is smooth and a less strain on my hands and arms. But nothing is wrong with rFactor's either, it just feels different. I can be fast in rFactor too, but in all sims, you can never be faster with a higher steering lock, so, sadly your test doesn't mean much . But still interesting to see.

Quote from Tweaker :

LFS will and always will work better with 270 degrees or around there. I have noticed that you really cannot drive LFS to its limits when you use a higher steering lock because you'll be so much slower in catching the tiny little snappy slides LFS creates while you drive.

That is a matter of practice… , not a general truth…
I use always the same degrees of rotation as the car in LFS has… I did that even with the DFP… now with the G25 it much easier… In no way it is slowing me down.

Quote from geeman1 :

Interesting.

I totally agree... I use a MOMO btw, I like it because it's fast, not like G25 with 900 degrees...that's terribly slow as I saw it in the previous video

Some rally cars do have lock less than 500 IRL, but it certainly feels very odd and wrong to use less than 720 with street cars, as I'm more used how cars act IRL.

Of course that wheel does not turn itself is huge problem also, however I prefer to turn as much as I would need to do IRL over that feedback, I don't like to learn to drive car so that I need to turn just 2mm of wheel, that would be troublesome IRL again.

I use 540 degrees. So I get 540 degrees in road cars, something like 420 degrees in GTR's and 360 in Formulas.

Let’s add this video to the review. http://ie.youtube.com/watch?v=8Sx1dlJoVNM
Why for him the wheel is fast enough, even in 900 degrees mode?
(rhetorical question)
(btw this has nothing to do with drift vs race, it's all about driving… don't start a war here.)

Oh it had nothing to do with being able (or not able) to use the 900 degree mode; with LFS's somewhat ''sluggish'' feel it seems to work well enough.

The point was to see if there are any differences in how 'stable' and 'quick' the FF is in 300, 600 and 900 modes. When the sim is kept the same, it will *desire* the same thing from the wheel in these three wheel settings, so if different things happen, the wheel must be the cause.

I think that if you did the same thing in a real car, you would have the same result. The car still needs driver input to keep from spinning out after an intentional oversteer. You can see it in kaynd's video, he is intentionally oversteering, lets the wheel go to auto countersteer itself, then grabs it at the point where he wants to hold the oversteer for the drift.

No car is going to correct it's self perfectly for real. The back end is going to occillate just as it did with your video. That is just my thoughts on it. I've never put my car intentionally into an oversteer and just let go of the wheel. I have had that happen by accident due to snowy/icy roads. The back end went out on me, I let the car countersteer automatically, but grabbing the wheel at the point where it needed to stop. As it started correcting, the back end occillated the opposite way and it countersteered automatically for that. I had to grab the wheel then as well as it countersteered. If I had left it alone, my truck would have spun out, just as your video shows.

I use 210 degrees all around, but hey, I'm a lazy bastard