Such a simple question, yet answer is extremely complex and quite unknown to us. I don't think even Scawen himself would know how to answer it at this point
What I'm trying to say is that it's not about gear or belt system, It's about the motor which makes the wheel feel smooth or not. Brushless motor is allways much smoother, because there are no brushess that cause friction, plus motor poles on DC brusshed motor are much stronger than on any brushless.
Why do you think all car transmission systems use helical gears and not belts and pulleys? Helical gears are much smoother than a timing belt and pulley system. The only downside with any gears is a tiny bit of a backlash they have, compared with belt system where there isn't any backlash.
I would have to dissagree with you guys. I like the feel of XFG and FOX more on G29, but everything else on T300 felt like a XRG, heavy and dumped due to belts. Overall T300 has better ffb, due to it's brushless motor which gives it smothness and less friction and somewhat better resolution of motor driver circuit, but it's still quite far from what DD wheels bring to the table.
On G29, generaly stiff cars felt better due to more quick reaction of helical gear system. If G29 had a brushless motor and a 3 phase motor driver from T300 it would be comparable, but like this it's not even fair to compare them. Both have pros and cons and it's about what suits different people more.
Yes, digital resolution of ffb is just a part of the story. The actual motor and mechanism are the other part. They all play together to result in the final ffb feel. If the motor is crapy and meshanism is very heavy and with quite some friction, it will eat away and dampen ffb details. This is the limitation in all non-direct drive wheels.
Bottom line, both G29, T300 and similair are good enough ffb wheels for the guys who want to feel ffb for the first time and who can aford it (by money and free time).
On my G29, it's a bit smoother. In old lfs ffb was 25Hz with 200 steps, once I set 100Hz and anything => 400 steps it was noticeable and better. In my T300, up to about 1000 steps you can feel that ffb is smoother, above that there is no difference, probably they have somewhat higher pwm resolution, but the frequency is lower, one can even hear the squicky noises comming out at 8kHz.
You shouldn't be worried about how 10000 steps will be scaled down to 400 or what ever is pwm signal resolution in logi wheels. Those are a very simple operations with integer numbers, no overflow or truncating will occur. Their firmware will handle it just fine.
All good, except I would set FF steps to 10000 and let logi wheel firmware deal with it and use as much as PWM resolution available. I believe that the motor is powered by an H-bridge driven with a PWM signal at 20kHz with 400 steps of resolution, but I didn't measure it, I just saw someone mentioned maybe Scawen himself that no more than 400 steps is available, due to low res PWM signal.
Logitech's trueforce has 4kHz ffb refresh rate, which is revolutionary in it self. This opens up many possibilites, but unfortunately not many games utilize it. With such a high update rate the bandwith of ffb signal is increased from tens of Hz to 2kHz. Vibrations mainly fall under the 100Hz range and everything above that falls into the audible range, so what games do is just relay part of the audio signal into ffb signal. This is in my opinion useless as we already use our ears to sense it, but it makes a great marketing.
Another thing is, to be able to reproduce 2kHz signals with an electric motor (bare in mind it is not a speaker), it has to be extremely sensitive and with a tiny rotor inertia. This can be done only with a powerful 3 phase AC servo motor, like the ones found in direct drive wheels. Normal brushed DC motor like the RS570 found in logi G25, G27, G29, G920, G923 has no chance at reproducing such high frequencies.
DirectInput API is old indeed (developed in 1999) but it's widely used. It consists of 12 ffb effects and some more flexibility like fade in, fade out, durration, delay and so on (borrowed from MIDI). Again, rarely any game uses more than 1 or 2 ffb effects. It is always the easiest for developer to use the game engine and calculate all the relevant forces internally, sum them up and send to ffb device through 1 ffb effect - constant force, a single value of force with additional details like magnitude and direction in case device supports more than 1 ffb axis. This effect is then called many times per second like in LFS 100Hz, to give an illusion of dynamic forces. This rate is crucial for correctly reproducing all relevant details. The remaining 11 ffb effects usually refered as canned effects are the periodic effects like sine, square, triangle, sawtoothup and sawtoothdown. In principle with a linear combination of these one can produce an arbitrary wave, but it's a hard way and games simply just use a triangle for curb effects. This is quite robotic as amplitude and frequency are fixed throughout effect durration. There are also effects that depend on (steering) axis position like spring, on axis speed or 1st derivative of position like damper and friction and finaly on 2nd derivative or acceleration of axis like inertia. There are also some more time dependent effects like ramp and custom force where one can play a predefined force sample. It is quite a versatile API but games mainly use only constant force effect.
Ok, not to make long story even longer, I think this gives some overall idea of how ffb works to anyone who was wondering about it Once Scawen release updated LFS with 1000Hz physics and ffb refresh rate, it's gonna be better than trueforce. At the moment, ACC holds a record for highest ffb update rate of up to 400Hz.
I have a laptop with integrated gfx inside an Intel i7 CPU, it is HD4600 and this is the previous generation compared to your HD520 I beleive. So even this integrated one is fine for LFS, but your dedicated nvidia one should be much better for it.
Can't be done at the moment. When Scawen include gyro forces from front wheel and with new tyre physics it will be possible. There is a plan for it, but it's in somewhat distant future.
We're happy to announce the results of our event. Thank you all for participating it was quite a blast.
The results from the qualifying session are: https://imgur.com/QScslxv
Last edited by rane_nbg, .
Reason : updating images
I see. In fact the nickname of the guy who did post in my github is "Notriatv", it does resemble your nick, but it was just my brain playing tricks on me during sleep time
It's not just steering wheel axis input, but all axis of all players, including pedals and so on. Doing it would increase multiplayer server load dramatically. I'm not sure if there is enough resources for this and this is what keeps LFS online experience at a very good level. Recently car coordinated and vector speed update rate in miltiplayer had been increased from 4 to 12 times per second. Still, ping of many players doesn't even allow it to be utilized fully.
It's generaly a nice suggestion, but I'm afraid LFS is not ready for it yet.
Please if you post a question with some issue and you manage to solve it, also post what you did to solve it. A lot of times people have the same problems and it may save them the trouble if they see answer right away.