Start here, before you read a word of the rest: turn your in-game Car Specific FFB Multiplier down, not up. The stock Le Mans Ultimate profile ships hot, and almost every “my FFB feels like mud” post is someone running 100% multiplier on a wheelbase that’s already clipping before the car has finished the pit exit. Set your driver strength to 100%, set the in-game multiplier to somewhere between 0.55 and 0.75 depending on the car, and turn on the clipping meter so you can actually see what you’re doing. That’s the fix. Everything after this is why.

The second thing: LMU is rFactor 2 underneath, and rF2’s force feedback is genuinely one of the best physics-derived FFB models in sim racing. It’s not synthesized. There’s no “canned effect” slider bank like ACC’s road effects or iRacing’s damping. What you feel is what the front tire contact patch is doing, filtered through the steering rack. Which means when it feels wrong, the problem is almost never “I need more effects.” It’s that you’re crushing the signal.

The clipping meter is the whole game

In the pause menu, under the FFB tab, there’s a real-time meter. Use it. Most people never open it, which is how you end up with three years of forum threads full of people swapping magic numbers.

Here’s what you’re looking for: on your fastest lap, at the highest-load corner on the track, that meter should be nudging the top of its range and basically nothing else should. Porsche curves at Le Mans in a Hypercar, say, or the entry to Turn 1 at Sebring where the car is loaded and the surface is trying to shake your fillings out. If the bar is pinned across a whole lap, you’re clipping, and clipping is not “strong FFB.” Clipping is your wheel outputting maximum force through a range of inputs where the car is doing three different things. It’s information deletion. The wheel physically cannot tell you the front is going light because it’s already at 8 Nm and it can’t go to 8.5.

The counterintuitive part is that a lower multiplier usually feels stronger subjectively, because the peaks stop merging into a wall and you start feeling the transitions again. I’ve walked several people down from 1.0 to 0.6 who were convinced I was sabotaging them, and every one of them came back within a session.

Actual numbers, by wheelbase

These are starting points I’ve settled on after a lot of laps, not gospel. Driver strength stays at 100% in all of these because you want the in-game multiplier doing the scaling, not two attenuators fighting each other.

The Logitech and Thrustmaster rows are the interesting ones. On a belt or gear base you’re fighting a totally different problem: not enough force, and a deadzone in the middle where the gears mesh. Cranking the multiplier to 1.0 on a G29 is correct and it will still feel thin at Indianapolis in the rain. That’s the hardware, not the settings, and no config file fixes it.

The three settings that actually matter beyond gain

Steering Torque Minimum (in the JSON, or under FFB in-game on newer builds). This one only exists to paper over deadzone in cheap bases. On a G29 or a T150, dial in 0.03 to 0.05 and the on-center numbness goes away. On any direct drive base, leave it at zero. Setting it above zero on a DD wheel adds a constant fake force that makes the car feel like it’s driving on a slightly cambered road forever. People do this by accident and then complain LMU has “notchy center.”

Steering Torque Sensitivity and Steering Torque Filter. The filter is a smoothing pass. Default is 0 in LMU and I’d leave it there on anything decent. If you’re on a Simucube 2 Pro with a stiff wheel and you’re getting cogging noise on kerbs at Monza, try 1 or 2. Not higher. At 4+ you’ve turned the FFB into a suggestion box and the tire information arrives about 40 milliseconds after you needed it.

The one people miss: Steering Wheel Range. LMU respects per-car steering lock properly, so set your base to 900+ degrees and let the game do the mapping. If your base is locked at 540 and you’re driving a GTE Corvette that wants ~720, your steering ratio is wrong and every input feels twitchy in a way that no FFB tweak will fix. Check your base software. Set it high. Move on.

Per-car, because LMU is not one car

The Hypercars and the GTE/LMGT3 cars need different multipliers and anyone telling you to run one global number hasn’t driven both back-to-back.

The Porsche 963 and the Toyota GR010 have hybrid front axles and huge downforce. They load up hard, they clip easily, and they want a lower multiplier than you think. I run 0.58 on a 9 Nm base in the 963. The Ferrari 499P is similar but slightly more talkative on the front.

The LMP2 Oreca 07 is the opposite. No hybrid, less downforce at the front, mechanical-feeling steering. I’ll bump the multiplier to 0.72 for LMP2 and it feels alive instead of numb. It’s also the best car in the game for learning the FFB because it tells you about front lock-up before it happens, which the Hypercars kind of don’t.

GT3 cars sit in the middle, around 0.65, and they’re the cars where you’ll notice the road texture most because they don’t have downforce papering over the surface.

Save these per-car. LMU stores FFB multiplier per vehicle, which is one of the few places the UI is genuinely helpful.

Things people argue about that don’t matter much

The FFB smoothing debate on the LMU Discord goes in circles roughly every two weeks. Someone posts a config with filter at 3 and everyone gets very serious. Try it for an hour. If you’re on a decent DD base and you like it, fine, you’ve traded some detail for less fatigue, that’s a real trade. But nobody is faster because of it.

The other one is people importing rF2 FFB configs wholesale. LMU’s tire model and steering rack simulation have moved on. The old rF2 “car specific FFB mult” values from 2019 spreadsheets will give you nonsense. Start from the game’s defaults and work down.

And the perennial one: whether LMU’s FFB is “better than iRacing’s.” It’s different. iRacing’s is cleaner and more predictable, which is why people can drive it for four hours. LMU’s is noisier and carries more surface detail, and the first time you feel the tire actually let go under braking into Mulsanne corner in the wet, you understand why people put up with the noisier signal. I’d rather have LMU’s. That’s a preference, and I’ll defend it, but I won’t pretend it’s a fact.

The thing that’s probably actually wrong

Desk flex. If you’ve got a direct drive base clamped to an IKEA desk and your FFB feels vague and mushy at exactly the moment the car loads up, that’s not your settings. That’s your desk absorbing the force feedback and giving it back to you late and blurry.

Put your hand flat on the desk surface next to the wheel clamp and take a corner at Le Mans. If you feel the desk move, that’s your problem, and no multiplier value fixes it. A wheel stand like the Next Level Racing Wheel Stand 2.0 (around $150) is a bandaid. An actual 8020 aluminium profile rig starts around $400 for a basic one and is the last frame you’ll buy. I put this off for two years, which was two years of thinking my wheelbase was underwhelming.

The order of upgrades that actually improves your feel, from someone who did it in the wrong order: rig first, then a load cell brake pedal (the Moza SR-P or Fanatec CSL Elite V2 both land under $300 and are a bigger lap-time improvement than any wheelbase upgrade), then more torque. Going from 5 Nm to 15 Nm on a flexing desk with potentiometer pedals is spending £900 to feel worse.

FAQ

Why does my LMU FFB feel weaker than in ACC or iRacing?

Because it probably isn’t, it’s just less compressed. ACC and iRacing both apply more internal shaping to the output, so the average force is higher even when the peaks are the same. LMU gives you a wider dynamic range with quieter mid-corner forces and sharper spikes. If you want it to feel like ACC, raise the multiplier and accept some clipping. I wouldn’t.

Do I need to edit the Controller.JSON file?

No. Everything that matters is exposed in-game now. Editing the JSON was necessary two years ago; it isn’t since the FFB tab got the full set of sliders. If a guide tells you to hand-edit values, check the date on it.

What’s the best FFB filter setting for a Logitech G29?

Zero, and instead set Steering Torque Minimum to about 0.04. The G29’s problem is deadzone and not-enough-force, so smoothing takes away detail you can’t afford to lose. Also turn off Logitech’s own “Centering Spring” in G HUB. It fights the game’s FFB the whole time and it’s on by default, which is a genuinely baffling decision.

Before you change anything else

Open the FFB tab, load an Oreca 07 at Sebring, and drive five laps with the clipping meter visible. Set the multiplier so the meter tops out only in the Turn 17 loading and the Turn 1 entry. Write that number down. Every other setting on this page is a rounding error next to getting that one right.

And if the answer is “it still feels vague at 0.6 with no clipping,” check your desk before you check your wallet.