CSL DD, my humble review

[manu68]

Hi.
Just some comment on my récent purchase of a CSL DD + MC laren V2 rim.

I put my order in june, and have been delivered about two weeks ago.
As about always with UPS, it hs been a mess, but that is not really the point here.

My simracing background is simple.
I am 54, and i practise simracing since Internet exists. First race with an 33 kb modem on Graand prix legends.
I have raced some years without ffb on basic wheel . Then purchased my old good logitech G25, which i have use until my recent purchase.

I always thought that it is a wheel which has allowed the simracing to be what it is today, and never considered to replace it , just because i have been satisfied all these years.

Then fanatec has put the CSL DD on the market, which i think is about the same revolution as the G25 .

It has a perfect size, it looks great, it is easy to install and configure, the shifter has a great feeling, the pedals are nice (i have them for about two years already).
And the price is very fair for what i got and needs.

The difference is huge with the G25, in many aspects. We could discuss long, but i think that the smoothness is key, more than torque or strenght. It feels more like driving a real car.
The simracing experience is the same as with the G25 or another wheel, just more enjoyable.

G25 was and remains a fantastic wheel. During many years, all wheels have been designed to clone it. CSL DD sets just the reference a big step beyond in all domains.
In about two years, i am sure that all wheels will be direct drive on the market, and even cheaper.

I recommand it more than 100%.
It is the perfect wheelbase for my needs and my rig.
Thanks to fanatec to have understood the needs of people like me

 

[Cote Dazur]

We could discuss long, but i think that the smoothness is key, more than torque or strenght. It feels more like driving a real car.

Totally agree here, CSL DD 5nm, the smoothness is just amazing, the speed too. I find the strength, with optimized FFB in games, maxed out without clipping, is very adequate. I have never felt a more realistic feeling of driving cars in my favourites SIM. Everything is here, particularly in VR. Not too loud, not too tiring, just right.

 

[RasmusP]

My 2 cents, copied from a pm. Maybe interesting for you two or someone else

Before reading:
Don't get me wrong. For the price it's amazing and if I wouldn't have a csw 2.5 already, the situation would be different!
But since I have one and it's only about getting X amount on the used market vs gaining some nice little extras for Y amount of money... :

I'm gonna send the CSL DD back to Fanatec and I'll wait for either the PS5 version next year with the new quickrelease or the probably at some point arriving CSW DD with hopefully 10-12 Nm.
The current QR on the CSL DD has some play with my plastic QR on the McLaren V2 rim. Lots of reports about this and it's fixable but... Not a great feeling when it has play straight out of the box.

BUT: I'm very satisfied that the new DD settings that the CSW doesn't have are doing exactly what I hoped for. "Friction" and "Inertia" are giving me a weighty and thick feel when driving in a straight line with very slight turns but they don't really slow down any quick movements.

The "natural damping" gives barely any additional resistance when turning just a little bit but when you really flick the wheel, it's like steering against pudding or quicksand that becomes heavier to overcome, the faster you turn.

In Dirt Rally 2 that was super annoying because I couldn't drive precisely in a straight line with a twitchy car, without a high natural damper setting.
But when driving straight was fine, I couldn't flick the car into hairpins anymore without being massively annoyed by the "Pudding-Resistance"

Now with the CSL DD I can set the natural damping very low and the natural fiction + intertia quite high. Resulting in a beefy straight line feel and once I overcome the initial resistance, flicking the wheel is quite easy.

BUT: I only ordered the 5 Nm version without the artificially expensive PSU that makes it 8 Nm and the FFB is just too weak. It becomes "okay" when I ramp up the gain but it's too much clipping then.

I'm using 60% gain with 100% in-car-gain with the Alfa GTA in Assetto Corsa with my CSW and only 80% max torque in the CSW 2.5 base. Currently doing a little championship with that car.

Which in theory should be (60*100*80)% = 0.48 * 8 Nm = about 4 Nm torque in my hands. (but 6.4 nm peak torque)

With the CSL DD I kept the 60% gain in the menu but to have decent FFB I had to ramp up the in-car-gain to 180%, having the base at 100% (5 Nm) which means 0.6*1.8*1.0 = 108% gain. So it's definitely clipping through turns that are fast and bumpy which means I didn't feel the wonky Alfa as detailed as I do with the CSW 2.5

Summary: DD friction + inertia settings are exactly what I'm missing right now but I'm not willing to pay for the 8Nm PSU, which costs 135€!
I'm also not convinced that 8Nm are really enough when I'm using higher friction/inertia and maybe damper settings. A lot of knowledgeable physics guys are constantly posting about high torque + high dampening (including friction and inertia) would be a lot more realistic with a very calm, but strong wheel during peaks.

Also I'm not going for a sidegrade when there's a PS5 version coming. I have a few friends who play Grand Turismo, even doing league racing and I'm tempted to get a PS5 in the future and do some racing with them. With the current CSL DD that would never be possible. My McLaren V2 rim is PS5 ready though.
If it would be a nice upgrade, I would sacrifice the ps5 compatibility though.

Comparison about the smoothness and details between CSW 2.5 and CSL DD:
They both have some very mild "stepping" when turning slowly and without ffb.
As soon as the FFB is there, it's perfectly smooth with both wheelbases. The amount of details is the same. Both can feel like sandpaper or grinding teeth across the tarmac so I need some smoothing with both and I wouldn't know what could be improved in that regard.
The CSL has stronger "steps" but is smoother when disabling any damping/friction/inertia. I can spin it more freely.
The CSW 2.5 has more basic resistance and doesn't spin as freely but it doesn't matter since you need more artificial damping etc to make both wheels feel "normal" than the resistance of the wheelbases anyway.

The CSL DD feels a bit nicer due to replacing damping with friction+inertia, making the wheel feel beefier but making it "flick" more swiftly when correcting a slide.
IMO Fanatec could patch friction+inertia into the CSW 2.5 firmware since it's just "fake" by controlling a single shaft motor in a different way.
But it won't happen.

So I'm gonna stick to the CSW for another 6-12 months and then probably get the 8Nm PS5 CSL DD with the new QR that hopefully won't have any play in it and the D-shaped round wheel with lots of buttons.

As I mentioned: Great product! Just annoying with hefty price point for the bigger PSU and currently not worth to sell my csw 2.5 for it.
But with the leaked gt7 rim, 8nm and ps5 compatibility it will be worth the additional price.

Additional info:
Fanalab is quite a new extra software that is mainly used for game specific profiles and to feed the little displays of the wheels with game telemetry.
But there's also a new setting:
Dynamic damping!
You can set a starting and maximum speed, the damper strength and the degrees around the center position where the damper should be active.
I have it set to 50/150 km/h ; 100% strength and 80° (40 to each side) right now.
It's a compromise with the csw 2.5 to get a thicker/beefier straight line feel. Of course it's not always accurate since with the csl DD friction+inertia the wheel feels thick and beefy no matter if the wheel is centered or not.
But in 90% of my driving I was only missing the resistance around the center so waiting for the ps5 version or the csw DD won't be that hard

 

[manu68]

Such a fantastic post again from you.
Always a pleasure to read from you.
Thanks a lot.

I have read about the play with the basic quick release, but i must adit that it is absolutely not the case with mine.

Regarding the 8 Nm torque, this is plenty sufficient for my needs. I am far away of using the all power of the unit, just because my shoulders could not handle that (i have currently some problems)
I have for sure no clipping problems, and the detail get from the wheel are just fantastic.

I need for sure some more tweaking with the different simulations, but i am so far really impressed.

 

[GeekyDeaks]

If it would be a nice upgrade, I would sacrifice the ps5 compatibility though.

There have been reports that drivehub works with it. Not that it's a good option if you want it purely for PS, but I guess it could be an option if you main PC or have non-fanatec pedals. I just started playing with gimx to see if I can make a config to use my alpha mini on my trusty old PS4. https://github.com/matlo/GIMX - main dev is waiting for his PS5 to get it to work on the latest and greatest though https://github.com/matlo/GIMX/issues/673

 

[RasmusP]

Such a fantastic post again from you.
Always a pleasure to read from you.
Thanks a lot.

Thanks!

I have read about the play with the basic quick release, but i must adit that it is absolutely not the case with mine.

Now that I know how a lot of play feels, I can feel it on the csw 2.5 too.. It's only about 10% as much, but I feel the slight clicking when the FFB direction changes..
Gonna write to fanatec and ask for a new QR...
Just difficult since they always want a video and I have no idea how I should hold tight the shaft of the csw 2.5... It's completely smooth and only sticks out about 1.5cm between wheel and base.
With the csl DD I could simply grab the clamp that holds the shaft on the base and then turn the steering wheel.

Regarding the 8 Nm torque, this is plenty sufficient for my needs. I am far away of using the all power of the unit, just because my shoulders could not handle that (i have currently some problems)
I have for sure no clipping problems, and the detail get from the wheel are just fantastic.

One guy on YouTube said "I tested what's the minimum ffb level needed to start to feel anything at all. It's about 3 Nm.
So with the 5 Nm variant, all" feelable" ffb happens within 2 Nm.
With the 8 Nm variant, that's 5 Nm so it's more than double the range for my brain and hands.
That's why the 8 Nm version feels absolutely strong enough and awesome, while the 5 Nm version is a great upgrade for Logitech users but for me, having a dd1 and csw 2.5 as comparison, the 5 Nm version is just "meh".
A smooth "meh" though. "

I'm not sure how technically correct this statement is, but it makes a lot of sense for my gut feel

I'm also not sure how fanatec controls the FFB strength.
Can the 8 Nm version at 63% strength in the base (=5 Nm) use more than 5 Nm to overcome inertia?
Is the FFB level limited as overall power input to the motor or can the motor accelerate with full power but won't give you more than 5 Nm steady torque when there's a human giving resistance?

My csw 2.5 set to 63% feels a lot stronger than the 5 Nm csl dd at 100%.
Sadly I have no possibility to measure things...
At about 50% in my csw, it starts to feel similar, although the csw still feels a bit punchier.

All the best for your shoulder!

I need for sure some more tweaking with the different simulations, but i am so far really impressed.

In theory you shouldn't really need anything apart from degrees for some games and ofc the max torque to mix the dynamic of the FFB with game-gain and base max ffb.
(although dirt rally and wreckfest have a "saturation" setting for the degree range)

There have been reports that drivehub works with it. Not that it's a good option if you want it purely for PS, but I guess it could be an option if you main PC or have non-fanatec pedals. I just started playing with gimx to see if I can make a config to use my alpha mini on my trusty old PS4. https://github.com/matlo/GIMX - main dev is waiting for his PS5 to get it to work on the latest and greatest though https://github.com/matlo/GIMX/issues/673

Oh that gimx looks nice! Thanks for the info!

I don't have a PS (yet) but a few friends have one. I'm not sure driving to them with wheelstand + drivehub and then getting things working would be a pleasant scenario hehe.
I'm mainly hoping for the new QR becoming a thing with the ps5 version and I also really want that leaked gt7 rim! Looks like a plastic, toyish rim but with lots of buttons/rotaries and ergonomic grip.
I want a round rim but don't want to sacrifice the functionality of the McLaren rim.. but the podium hub is way too expensive and the clubsport uni hub doesn't have enough functions.

So it would be a triple:
- driving at friend's PS5's
- getting a nice and cheap rim for my needs
- hopefully getting a better QR

 

[Cote Dazur]

The amount of details is the same

Interesting post, you seem concerned by the strength of the wheel a lot.

Of course a wheel has to be strong enough, but the main reason it has to be strong is to be able to communicate loud enough what the car is doing to the driver.
The more important aspect is which wheel communicate what the car is doing to the driver better.

Is it the CSL DD or is it the CSW 2.5? Keep the wheel that is the more informative.
You wrote, the details are the same, but I find it strange that the smoothest and fastest wheel, is not more informative.

All things being equal, the strongest wheel has a better chance to feel more real, but is it the case between the CSL DD and the CSW 2.5, is everything equal except for the strength?

 

[RasmusP]

Interesting post, you seem concerned by the strength of the wheel a lot.

That's quite a funny statement in this case because I'm always stating that I will never need a stronger wheelbase as the csw 2.5.
I never use it at 100%!
For AC, I mostly use 50-70% gain and 65-75% base strength.
That means, averaged, 60% gain * 70% strength = 0.6*0.7*8Nm = 3.36 Nm with peak torque 5.6 Nm (in theory).

I tested if 63% strength would still be fine, which would equal the 5 Nm of the csl DD. And it is!
But sadly, the csl DD feels significantly weaker so either the csw is stronger or the csl DD doesn't really reach 5 Nm...

I also think that the 8 Nm csl DD would be plenty strong enough for me!
But without ps5 compatibility and/or the new QR2, the "sidegrade" isn't worth it for me personally.

Of course a wheel has to be strong enough, but the main reason it has to be strong is to be able to communicate loud enough what the car is doing to the driver.
The more important aspect is which wheel communicate what the car is doing to the driver better.

Definitely, 100% agree. I always only use as much gain/strength as I need to feel the level of grip changing depending on load and aero effects.
And even with 100% gain and 100% strength, I didn't feel what I needed.
And I don't really understand this because with my lowered strength csw 2.5, I do feel everything although it shouldn't be stronger (at 63% base strength).

You wrote, the details are the same, but I find it strange that the smoothest and fastest wheel, is not more informative.

That's because you don't take the damping (and friction+inertia for DD) into consideration for.
If I put all damping to 0 and put friction+inertia for the DD to 0 too, the csl DD is smoother and faster and yeah, I get "more details".
But it feels completely ridiculous with both wheelbases.
Like my car would have zero mass and the tyres would be rigged up springs instead of rubber on a road.
I wouldn't say that you get "more details" either if cobblestones feel like grinding my face across loose gravel.
That's just noise.

So when I tune both wheels to feel good (natural damping (drift mode) to -1 or -2 (-5 being default and maximum) on the csw 2.5 and for the csl DD natural damping to 20% (50% being default and comparable to -5 on the csw), alongside natural friction+inertia to 50% (0% being default):
Both wheels deliver the same amount of useful details and feel great.

Due to friction and inertia, the csl DD feels a bit more real since it feels like it would have more mass without slowing down quick flicks or swift countersteer.

So in the end, the 8 Nm version might feel quite a bit better than the csw 2.5 and keep more useful information while feeling like it would have more mass, which I like and find to be more realistic.

But it's not worth the extra cost right now.
With QR2, new round rim with lots of functionality and ps5 compatibility? Hell yeah! 100% gonna buy it!

CSL DD or is it the CSW 2.5, is everything equal except for the strength?

After everything I wrote, yes, more or less.
Csl DD 5Nm feels better in theory but is a bit too weak to feel as real.
The differences in details/information vanish when dialing in realistic damping/friction/inertia.

Kyuubey who you know from the extended physics and csp gyro talk stated multiple times that almost all DD users are using way too low damping to be anywhere close to reality (which includes friction/inertia in this case), while using "real" torque values.
Which results in a fake, hyperreactive wheel that goes completely nuts when let go

He actually measured this kind of stuff from real cars. Torque, inertia, friction from real steering racks so I trust him.
Mclarenf1papa agrees with that and he works with a real lmp team for simulation stuff.

 

[Cote Dazur]

Thank you for the detailed answer, you are correct in all your statements, as they relate to you. Different people will make different choices.

I never use it at 100%!
For AC, I mostly use 50-70% gain and 65-75% base strength.
That means, averaged, 60% gain * 70% strength = 0.6*0.7*8Nm = 3.36 Nm with peak torque 5.6 Nm (in theory).

We are on the same page, I changed from my (old) CSR Elite modified with Buhler motors because it was having a few issues, strength was not one of them. I could have gone with a "normal" DD, but they always seem to me outrageously strong for no good reason.
When Fanatec introduced the CSL DD, it felt to me that they had done it just for me . I went with 5nm, because more strength has never been on top of my list.
I may or may not upgrade to 8nm, the jury is still out, as I am still getting to know and understand my 5nm CSL DD, but if I do, I already know, it is not going to be for more strength.

That's because you don't take the damping (and friction+inertia for DD) into consideration for.
If I put all damping to 0 and put friction+inertia for the DD to 0 too, the csl DD is smoother and faster and yeah, I get "more details".

I have played a little, I should do more testing, with friction&inertia. For now, I do not like the "noise" it adds.
When I started, the first few hours, using my CSL DD, I was shocked, in comparison to my CSR, at how "light" the wheel was feeling. I am not sure if it is because the wheel aged a little or because I got used to it, but that "light" feeling is not there anymore. That said, neither my CSR or the CSL DD, ever had or have the weight feeling I have in my real car. If I could find a setting I like that would make it closer to the IRL "weight" without adding "noise", I would be ready to sacrifice a little smoothness and agility.

To me FFB is a compromise, a balancing act, between being real and informative.
I totally agree with Kyuubey, Mclarenf1papa and anyone with the same opinion, that most DD user felt in love with the DD "effect", ultra strong and ultra reactive, almost as a justification of the expense, but distracting them from the balance approach.

And to finish, yes, some very limited number of cars, can take advantage of the extra strength of a 20nm wheel base, but it is the exception.

 

[RasmusP]

We are on the same page, I changed from my (old) CSR Elite modified with Buhler motors

Oh the CSR.. I looked up the tuning menu explanations for that wheel as I remembered something got changed regarding the drift mode:
"Drift Mode (DRI) When "Drift Mode" is enabled, the wheel has less dampening which allows for quicker turning. This setting is useful for "drifting" or cars that require quick inputs. This can be adjust in increments of 1, from 1 to 5, 5 being quickest, or least amount of dampening. Or you can set to off to have full dampening."
That is different to the CSW 2.5... I have -5 to +5 with 0 being "no damping and no artificial accelerating".

Here's the blog entry from Fanatec about this:

Fanatec

fanatec.com

So with the CSR and the CSW 2.0 you couldn't reduce the damping. You could only choose between in my opinion way too much damping (-5 feels radicicolous with the CSW 2.5 and was the default for the other wheels!) or a combination of reduced damping+artificial acceleration (positive drift mode values on my csw 2.5).
These positive drift mode values feel even more weird than -5.

So there lies a massive difference between our old wheels in comparison to the CSL DD.

Do you remember what you had your DRI (drift mode) set to?

When Fanatec introduced the CSL DD, it felt to me that they had done it just for me

Haha yeah same for me! I'm a bit annoyed by the fans of the csw 2.5 as they are louder than my PC fans (only silent wings 3).
And since my wheelstand is placed on my side while doing office stuff, I'd like a smaller wheelbase. I sometimes hit the base with my chair. Never the rim though for some reason...

I have played a little, I should do more testing, with friction&inertia. For now, I do not like the "noise" it adds.

Mh.. yeah I know what you mean but I don't think it was an issue for me. What are your other settings?
In the advanced menu you get added:
FFS: LIN vs PEA - I used peak
damper, NDP: I used 20-30
friction, NFR: I used 50
interpolation, INT: I used 5
intensity, FEI: I lowered that to 80
FOR, SPR, DPR default at 100

And then 75% gain in AC, 100% car gain, 100% wheelbase strength. 900°, 0.5% min force, csp gyro, no ffb post processing.
And:

BTW I liked the straight line feel of the TS-PC better than the CSW 2.5. The TS-PC has a really high "initial resistance" so it feels super beefy and weighty when driving straight. But as soon as you overcome the initial resistance, you can flick the wheel around without issues.
The rest of the wheel+base is no comparison to the CSW 2.5 though. Slower, louder, coil whine, cracking sounds, louder fan, no on/off etc.

That said, neither my CSR or the CSL DD, ever had or have the weight feeling I have in my real car.

Yeah... the issue is that in reality you have a combination of these things:
- 2 wheels that have inertia (gyro) against any movement when spinning
- 2 wheels that push towards the center position (or whatever is "straight" for them depending on the road cambering etc) via the caster.

So basically when driving straight, your real wheels will "Hold tight in center position". How do you want to simulate this with one motor? It can't hold tight, only against getting pushed into a direction. In theory, when you've have absolutely zero deadzone and instant reaction + instant torque build up, your brain wouldn't be able to feel the difference.
But sadly we have some latency in everything meaning going too tight, you'll get oscillations. If you dampen the oscillations, you also dampen the rest of the ffb in center position.
In reality that's not damped at all (apart from the gyro effect ofc). If the road pushes the front wheels, that's instant, while the front wheels still want to self align.
Basically you would need the ffb motor to push the steering wheel into one direction, then back to the center and instantly stop, while still giving instant torque if you want to move the wheel.
All that while not making you feel the "ffb shut off points" when reaching the center.

The Logitech wheels use 2 motors that push against one shaft. Sadly that's even worse due to latency from both motors oscillating against each other. so you need a massive deadzone.

My Logitech LUTs for AC try to be a workaround. I set some ffb output for the 0% entry. Which is just noise. It rumbles a little bit when standing still. Dialling this in took a few hours. Just raised it until the wheel felt a bit "thicker" in the middle where no ffb happens.
Then I knew from wheelcheck that the motors only react from about 15% or higher.
If you set min force to 15% though, you will feel the "edge" of the ffb.

So I made a smooth curve up to 20%, very steep, but smooth. Beginning with the "0% ffb noise" to thicken things up.

Here's a graph of my last LUT I created before being happy until the Fanatec upgrade:
Y = ffb output to wheel driver
X = game output

 

[Cote Dazur]

Do you remember what you had your DRI (drift mode) set to

Yes, that one is easy, 0 or off, not sure but was not using it. Way back I got the settings from Aris for my wheel, I figured that would be best, it had a lot of setting, DRI was one of them, I did not liked it, I prefered using everything off and still did to the end. Aris settings made the wheel too busy for me, more informative but too much SIM like.

For the CSL I have been mostly using the recommended Fanatec settings with between 90 to 75 % in AC, everything else off or zero.
I tried the CSP FFB tweaks, very interesting, but not totally convinced.
I am still trying this and that, as the CSL is still recent, about a month now, but my main objective with FFB was and still is today to forget about it, just enjoy the immersion. So I hope to just use it without thinking about settings sooner than later.

You are way deeper in the subject than I am, congratulation on all your research and the extend of you test. Looking forward to read about your findings.

 

[RasmusP]

Hehe yeah I'm quite deep into it after I just couldn't get happy with the clunky piece of logicrap...
I absolutely can't understand how anyone can like that.
For me, having a clunky dead zone while driving straight kills all immersion since no real car would ever be anything like that..

But yeah, buttons, pedals, shifter for a nice price so if you can sacrifice the straight line feel, it's a good package...

But I didn't give up and read into it. But I'm also studying/working mechatronic engineering so a ffb wheel from software to manufacturing covers all I've learnt!
So I liked and enjoyed uncovering layer per layer over a few months until I finally got happy with my custom LUT.
It's basically just a smooth minimum force setting like project cars 1 had.

Yes, that one is easy, 0 or off, not sure but was not using it.

Ah, so it was basically like my csw 2.5 and -5.
In this case the csl DD will feel incredible in fidelity, swiftness and smoothness I guess!

If you want to know what my csw feels like with drift mode disabled:
Put friction, inertia to 0 and natural damper to 10%.
Which is why -1 or -2 are needed to give the wheel some realism hehe.

I tried the CSP FFB tweaks, very interesting, but not totally convinced.

They are a little bit different, yep.
But undeniably very "real".
I'm using 0.1.60 though which only has range compression (always 100% (inactive)) and the gyro with a strength slider.
I'm using the default 25%.

 

[manu68]

I am so happy to see that my humble review had turned into an amazing conversation !

I have made ome trials yesterday with some friction and inertia (porsche 993 , nordschleife). It is interesting.
I also gave a try at the ffb tweak "real force" options, with a 8 Nm peak. I should try again but in my opinion it feels really good. More stronger peaks by driving on curbs , better feel on high speed curves,...
Not 100% sure but it seems interesting to me.

A youtuber in France has also made some trials using the LUT with his direct drive wheel
Of course his wheelcheck is pretty linear, but it changes things on the high forces domain (lut straight line , then some kind of curve at the end).

 

[RasmusP]

I am so happy to see that my humble review had turned into an amazing conversation !

Me too! I wrote most of my first post here in a PM to a racing buddy and he said "You should post this publicly" but I was hesitating about the interest of "Put friction and inertia to 50%, like no one on the planet does apart from you"
After seeing the replies coming in after a few days, I'm very happy I got involved!

I have made some trials yesterday with some friction and inertia (porsche 993 , nordschleife). It is interesting.

"Interesting" haha. Can you try to find words for what you like and don't like about it?

I also gave a try at the ffb tweak "real force" options, with a 8 Nm peak. I should try again but in my opinion it feels really good. More stronger peaks by driving on curbs , better feel on high speed curves,...

I've read about this, I think it was from kyuubey and mclarenf1papa again. Real force basically just takes your gain settings and compares them to the theoretically real force outputs. Then adjusts the output to match.
The issue is that it only really works with cars that do have realistic force outputs, which afaik only some mod cars do.
Also it only really works for cars without power steering, since AC doesn't have a realistic power steering simulation and you therefore can't use the force outputs from the steering rack for cars with power steering.

Overall the two guys suggested to only use real forces with the big DD motors and only for cars without power steering and only when they are accurately modded.

In the end it doesn't change the ffb though, only the gain! So if you disable the real force option and simply lower/raise the gain to the same level, it should feel 100% identical.

Your description sounds like lower gain, so you get more dynamic ffb. Difference between slow corners and fast corners seems to be bigger, which is the effect of lowered gain and stronger peak torque of the motor.
The difference can become too much btw. I'm using around 60% gain with 70% motor strength. Which is 0.6*0.7*8 Nm = 3.36 Nm.
("Nominal torque"? I don't really know how to call it.. It's not an average. It's the max. torque you would get if you would use 100% gain and 42% motor strength (8 Nm * 0.42 = 3.36 Nm). So you get the same ffb output when just hitting the clipping limit but with 60% gain and 70% motor you have additional headroom for the ffb to you "beyond the original clipping level".

Anyway, I could also use 42% gain and 100% motor strength. Resulting in the same 3.36 Nm "nominal torque" but even more max. peak torque.
I call this "ffb dynamic". Difference between "nominal torque" and the peak torque.

You have to dial in the ffb dynamic to your liking. I like to be able to keep the wheel in control no matter what happens. I don't really enjoy getting my fingers broken. So I dial in the gain until I hit the clipping limit when going over a sausage kerb but won't hit the clipping limit when going over normal kerbs. This gives me maximum dynamic for normal cornering and driving an limits anything that is beyond that.

A youtuber in France has also made some trials using the LUT with his direct drive wheel
Of course his wheelcheck is pretty linear, but it changes things on the high forces domain (lut straight line , then some kind of curve at the end).

This seems to be pretty weird... The thing is, I don't trust wheelcheck. Or better said, I don't really think that the wheelbase resistance is linear over "swinging speed".
In numbers:
If you give the wheel a 3 Nm push, it will travel 180°.
Will it really travel 360° if you push it with 6 Nm?
What about inertia? Will the wheel have the same turning speed when the ffb pulse stops or would you need to adjust the length of the ffb pulse to match the "let go" speed before starting to measure the travel?

I have the feeling for example that the "drift mode" natural damping of my csw dampens more, the faster I try to spin the wheel. This would make wheelcheck measuring a curve instead of a straight line, right?

From my short testing, the CSL DD, without any ffb happening but with the power on and all damping disabled, felt like a very very mild "stepping" when turning the wheel really slowly. Like going from one magnet to the next (like PC fans with their only 4 magnets mostly).
Once you overcome this mild resistance, it was super smooth, almost zero resistance. But when you really want to throw the wheel, you'll hear some "neow" sound coming from the motor.
This all indicates that the resistance isn't linear at all, so wheelcheck won't measure a perfectly straight line.

That's why I don't think that this lut in your screenshot makes a lot of sense to be honest... However it might feels quite awesome, when you get way stronger ffb right before the clipping limit. It's basically an "exciter", the opposite of compression.

But if I'm right and wheelcheck is measuring incorrectly, you get non-linear ffb at the top end of the range, which might not really make sense to your brain.
Going 10% faster through a high speed corner will feel like the tyres have 30% more load on them. So you'll think that you have 30% increased grip but when you lose the rear, it will go back to into the linear range.
Tricking you and maybe making you slower or less controlled at the higher end.

Here are some wheelcheck results at 50% and 100% motor strength:
Remember with my wheel, Drift at -5 = max artificial "natural damping", 0 = no damping, just the internal resistance of the base.

(I'm fearing for my wheel at 100%... Wheelcheck throws the wheel to the rotation-stop after the last pulse when setting drift to off. I could set the X offset to 400° though so it doesn't hit the end stop anymore)

50% motor, drift off:

50% motor, drift -3:

50% motor, drift -5 (max):

100% motor, drift -5:

100% motor, drift -3:

100%, drift -1 (For science! Fearing for my wheel...):

100%, drift 0 (hope my wheel survives lol, would be a shame to have to buy the 8 Nm CSL DD...):

The curves aren't that smooth due to only being a "max count = 10" test. Default is 50, LUT generator author suggests using 100 for more precision.

Here's the last one with 100 steps (my poor csw...):

And with drift at -5 (max damping):

 

[RasmusP]

Oh, I forgot to write the conclusion about the graphs:
According to wheelcheck, the damping changes the force output linearity very drastically.
But when you're holding the wheel tight and the motor outputs the force, there's no damping happening since the damping only slows down movements and doesn't lower static holding torque.

My arms+hands guarantee you that the driving forces feel exactly the same no matter the damping level. Apart from faster movements during swift turns or when drifting of course.

I'd love to do the same wheelcheck tests with different inertia and friction levels with the csl DD.
Didn't think about this somehow..

 

[Cote Dazur]

@RasmusP i would like your opinion, as you seem much better than me with maths.
Is it accurate to say that a 5nm csl dd will actually be a 4.75 nm wheel if set at 100% and game set a 95% and that a 8nm will be a 4.4nm wheel if set at 100% and game set a 55%?
At those setting both wheel strength will feel very similar.

 

[RasmusP]

@RasmusP i would like your opinion, as you seem much better than me with maths.
Is it accurate to say that a 5nm csl dd will actually be a 4.75 nm wheel if set at 100% and game set a 95% and that a 8nm will be a 4.4nm wheel if set at 100% and game set a 55%?
At those setting both wheel strength will feel very similar.

Almost!
There are 4 important values:

1.
The peak torque, which is the maximum the wheelbase will output no matter the game's output. That's limited by the wheelbase strength. In your example 100% both with 5 Nm and 8 Nm.

2.
Then there's the game gain, which is basically "what was the clipping limit with the default 100%".
When using 50% gain, "what was clipping before" is now only half strength but now you have 50% headroom!
It will still clip at 100%, but these 100% will now be double the strength compared to default.
I always use "driving straight", "cornering", "sausage kerb", "hitting the wall" as example values.
At 100% gain (fantasy numbers) :
Driving straight = max 10% ffb
Heavy Cornering = about 80% ffb
Sausage kerb = 120% ffb (limited at 100%)
Hitting wall = 300% ffb (limited at 100%)

3.
Wheel motor torque (5 Nm, 8Nm)

4.
And then there's the mix of these 3, which I would call "nominal torque".

It becomes easier when you always note down all 3.

The calculation goes like this:
(multiply = *)

100% gain * 100% base * 5 Nm = 5 Nm nominal torque and 5 Nm peak torque

50% gain * 100% base * 5 Nm = 2.5 Nm nominal torque and 5 Nm peak torque

100% gain * 50% base * 5 Nm = 2.5 Nm nominal torque and 2.5 Nm peak torque

75% gain * 75% base * 5 Nm = 2,8 Nm nominal torque and 3,75 Nm peak torque
(0,75*0,75*5 Nm = nominal torque)
(0,75*5 Nm = peak torque)


So far so good. Now when using the 8 Nm psu, you can do 2 things:

1. Making it exactly like the 5 Nm by lowering the wheelbase strength.
That would be 8 Nm divided by 5 Nm = 0.63.
So putting the 8 Nm motor to 63% wheelbase will limit the max torque to 5 Nm.

100% game * 63% base * 8 Nm = 5,04 Nm nominal torque and 5,04 Nm peak torque

Or you could lower the game gain to 63% to have the same nominal torque but a higher peak torque:

63% game * 100% base * 8 Nm = 5,04 Nm nominal torque and 8 Nm peak torque!

Now most people would say you should "use the possibilities!" -
and use 100% in the wheelbase to get the full 8 Nm peak torque.

Comparing my 4 example situations:
1. Driving straight = max 10% ffb
2. Heavy Cornering = about 80% ffb
3. Sausage kerb = 120% ffb (limited at 100%)
4. Hitting wall = 200% ffb (limited at 100%)

With:
5 Nm psu, 100% game, 100% wheelbase compared to:
8 Nm psu, 63% game, 100% base,
it would be like this:

1. Driving straight;
10% game * 100% base * 5 Nm = 0,5 Nm
(63%*10%) game * 100% base * 8 Nm = 0,5 Nm

(I'll leave out the 100% base now since it's always the same for these 4 examples)

2. Heavy Cornering:
80% game * 5 Nm = 4 Nm
(0,63*80%) game * 8 Nm = 4 Nm

3. Sausage kerb:
120% game (clipping) * 5 Nm = 5 Nm
(0,63*120%) game (no clipping) * 8 Nm = 6 Nm

4. Hitting a tree:
300% game (clipping) * 5 Nm = 5 Nm
(0,63*300%) game (clipping) * 8 Nm = 8 Nm

Now the sausage kerb example shows the important difference. You now have the "correct relation in strength" between heavy cornering and hitting a sausage kerb

While the "hitting the wall" example clips with both. But hey, you can now successfully snap a finger when hitting a tree thanks to the 8 Nm psu

Now imagine the same examples with 20 Nm DD wheels.
Making it the same nominal torque as the csl DDs:
(20/5 Nm = 1/4th or 25% game gain).

1. Driving straight:
(25%*10%) game * 20 Nm = 0,5 Nm

2. Cornering:
(0,25*80%) game * 20 Nm = 4 Nm

3. Sausage kerb:
(0,25*120%) game = 6 Nm

4. Hitting wall:
(0,25*300%) game = 15 Nm

Congratulations, you didn't get anything useful from using the full 20 Nm since cornering and sausage kerbs won't clip anyway.
But when you crash and hit a wall, you can now break your wrists. Yay!


To close on this and posting my personal and technically logical opinion:
You should go on track, use only 20% gain or whatever will make you never reach the clipping limit (the gray bar in the pedals app in AC).
Then record telemetry with something like acti + motec.
Do a full lap, use the kerbs, hit some sausage kerbs etc.
But don't crash, we only want "useful" ffb, not "sadly it's that strong in real life when something bad happens" ffb levels.

Then look at the FFB trace in the telemetry. Search for the maximum. Let's that at 20% gain that will be 35% ffb output.

Now raise the game gain so that the 35% will be 100%. This means all useful ffb will never clip but when you crash, you won't get unnecessarily strong jolts into your hands!

In this case 100/35 = 2,857
So the 20% gain multiplied by 2.857 = about 57% "theoretically perfect gain".

That's how I do it!

If a car is too strong for my hands, I then lower the max torque (wheelbase %).

In games that don't have a clipping meter I mostly start with around 75% gain and 75% base, which means 4,5 Nm nominal torque and 6 Nm peak torque with my csw 2.5.

If I feel that the average ffb is too low but the peaks are too strong, I raise the gain and lower the base strength.
If I feel there's constantly too much force and I can't really differentiate the grip levels, I lower the gain and raise the base strength.

When I want better differentiation but also no stronger peaks, I lower both.
And if I want more average torque and stronger peaks to have an exciting ride, I raise both

 

[Ole Marius Myrvold]

Can I just say... A future plan of mine, after getting a bigger apartment so it's possible to have a "rig" (more heavy table and maybe a chair without wheels), I do want a DD wheel.

Reading this thread though, makes me appreciate my G29

 

[RasmusP]

Can I just say... A future plan of mine, after getting a bigger apartment so it's possible to have a "rig" (more heavy table and maybe a chair without wheels), I do want a DD wheel.

Reading this thread though, makes me appreciate my G29

I have a wheelstand pro v2 for about 120€. For heel & toe the pole in the middle is a bit annoying but for left foot braking it's totally fine. I put my pedals slightly to the left to the pole is in the middle of throttle and brake.
Totally rigid enough for 8 Nm! Sure, it's wobbling a bit around and if you'd use 100% gain and 100% base strength, the wheelstand might start wandering around it's fine since 4 years for me!
I have done little drawer on wheels below my desk. Pushing it against the wall, then placing the wheelstand against the drawer fits perfectly with the 80cm depth of my desk.
The wheelstand plate is hovering slightly higher than the table, with about 5cm overlap.
So basically like the wheel would be desk mounted.

Then my chair on wheels as close as possible, folded newspaper behind the 3 rear/side wheels and that's it.
V3 pedals with the stiffer brake performance kit, using the load cell with 50% max load = 100% brake input and about 6.5 Nm peak torque

My recommendation: wait for the ps5 version, then you'll get a very functional (but slightly toyish looking) round rim + pedals!

 

[Ole Marius Myrvold]

I have a wheelstand pro v2 for about 120€. For heel & toe the pole in the middle is a bit annoying but for left foot braking it's totally fine. I put my pedals slightly to the left to the pole is in the middle of throttle and brake.
Totally rigid enough for 8 Nm! Sure, it's wobbling a bit around and if you'd use 100% gain and 100% base strength, the wheelstand might start wandering around it's fine since 4 years for me!
I have done little drawer on wheels below my desk. Pushing it against the wall, then placing the wheelstand against the drawer fits perfectly with the 80cm depth of my desk.
The wheelstand plate is hovering slightly higher than the table, with about 5cm overlap.
So basically like the wheel would be desk mounted.

Then my chair on wheels as close as possible, folded newspaper behind the 3 rear/side wheels and that's it.
V3 pedals with the stiffer brake performance kit, using the load cell with 50% max load = 100% brake input and about 6.5 Nm peak torque

My recommendation: wait for the ps5 version, then you'll get a very functional (but slightly toyish looking) round rim + pedals!

This is all good. But genuinely. We are two people living in a 54m^2 apartment. The choice was between a proper dinner table, or a 2m IKEA table as a "desk" for two PC's. There's no room for any wheelstand or anything in the apartment. There's just no room My PC on the left side of the IKEA-table is right in between the table and the sofa. On the right side, the other PC is right alongside the small kitchen table So that's it when it comes to space!

You could say that with a 2m table, I could mount a wheel to the table. But it's IKEA. It's thin "wood" with lots of air in it. I feel bad enough for the one on my right that I use the G29 and have that shaking

Bigger apartment, then more gaming stuff. Woop Woop