How force feedback works

Force feedback is the single most important feature in your sim rig. It's the main channel of communication between you and the physics of the car. Without it, you're driving blind. Set it up properly and you can feel grip levels, weight transfer, tyre slip, kerb strikes, surface changes, all through your fingertips with split-second precision. I've spent more time tuning FFB profiles than I care to admit. The difference between a well-tuned and a poorly-tuned wheelbase? It's the difference between fighting the car and reading it.

This guide covers what force feedback actually is, how the signal gets from the game to your hands, why direct drive wheelbases changed everything, and how to set it up for speed rather than just immersion. If you want to skip straight to tuning, jump to Setting it up.

What force feedback actually is

You know the feeling when you turn a real car's steering wheel and the suspension pushes back? That's caster angle doing its job. The front tyres generate lateral force against the tarmac and that force travels up through the rack into your palms. More grip means more pushback. Slide past the limit and it vanishes.

That's force feedback. In a sim, the physics engine calculates those same forces from its tyre model, and the wheelbase motor recreates them on your shaft. Grip loading up through a fast sweeper. Grip vanishing at turn-in because you've asked for too much angle. Weight lurching forward under braking. All through your hands, nowhere else.

Here's what most people miss: FFB isn't about strength. It's about variation. The useful information lives in the changes. Steering loads up as you build front tyre load. Goes light when you blow past the optimal slip angle. Pulls past centre in an oversteer moment because the fronts still want to point where they're going, even though the rears have checked out. Grip the wheel too tightly and you lose all of this. Every fast sim racer I've spoken to, coaches included, uses what they call a light-hands technique. Relaxed grip. Reading the changes, not wrestling the rim.

The FFB loop: game to hands

Six stages between the game's physics and your palms. Any one of them can introduce latency or eat detail.

1. Game telemetry. The sim spits out raw data at its physics tick rate: wheel speed, suspension travel, tyre slip angles, vertical load on each contact patch. iRacing ticks at 60 Hz, so you get a fresh FFB packet every 16.7 ms.
2. Physics engine. Complex tyre models (Pacejka and proprietary variants) crunch self-aligning torque from lateral force and pneumatic trail. How good this model is determines the ceiling for FFB realism, regardless of your hardware.
3. FFB signal. Calculated torque gets converted to a standardised signal (DirectInput constant force effect, usually) and shipped to the wheelbase at up to 1000 Hz over USB.
4. Motor controller. Drive electronics interpret the signal, apply any firmware-level filtering (dampening, friction, slew rate limiting), generate the appropriate current commands via PWM. Each manufacturer's secret sauce lives here.
5. Motor. Electrical energy becomes mechanical torque. The motor's inertia, inductance, and direction-change speed all affect how fast force appears at the shaft.
6. Your hands. You feel it, you react, loop starts again.

In a well-optimised DD system, mechanical latency (stages 5 and 6) is basically zero. No belt stretching. No gear backlash. Nothing eating the detail. Total system latency with the game engine sits low single-digit milliseconds. Belt-driven systems add measurable delay from belt elasticity and pulley inertia, and they chew up high-frequency detail before it reaches your hands.

Motor types: gear, belt, and direct drive

Gear-driven wheelbases run a small motor through a gear train to amplify torque. The gears introduce backlash (mechanical play) and friction, giving you a notchy, cogging feel at low speeds. Peak torque around 2-3 Nm. Fine detail gets chewed up by the gear mesh. Most people start here. It'll teach you the basics but you're missing most of what the sim's trying to tell you.

Belt-driven systems step up with a toothed belt and pulleys. Smoother, less backlash, better detail getting through. Belts stretch over time though, introducing compliance. High-frequency effects, the stuff that tells you about road surface and tyre scrub, get absorbed by that belt elasticity. Peak torque typically 3-8 Nm. Real upgrade from gears. Still filtered.

Direct drive eliminates every intermediary. Steering wheel bolts straight to the motor shaft. No gears. No belts. No backlash. No elasticity. Every nuance the physics engine calculates arrives at your hands unfiltered. DD bases run brushless DC servo motors with closed-loop control, typically 20 to 22-bit encoders giving millions of position counts per revolution. Peak torque goes from 5 Nm on entry-level units up to 35 Nm on the Simucube 3 Ultimate. The dynamic range between the quietest road texture and the hardest kerb strike is vastly wider than anything belt or gear can manage.

What you actually feel through the wheel

Set up a DD base properly and the steering wheel becomes a live instrument. I'll break down what each sensation is actually telling you:

Understeer: steering goes light. Fronts have blown past their optimal slip angle. Less lateral force, less self-centring torque. Wheel goes dead. You see the car not rotating visually and you feel it through the loss of resistance at the same moment.

Oversteer: wheel stays heavy, might even pull past centre. Fronts are still biting while the rears have gone. Relax your hands and the self-centring force counter-steers for you. That's exactly why the light-hands technique matters. Let the physics help.

Weight transfer: lift the throttle mid-corner, front loads up. Steering gets heavier. Brake and heavier still. Roll onto the throttle, weight shifts rearward, fronts unload, steering lightens. All happening without you changing steering angle. What you do with your feet changes what you feel in your hands.

Surface changes: camber shifts, kerb strikes, grass transitions, compression zones. A DD base with enough dynamic range reproduces these as distinct textures. Not vibrations. You feel a kerb as a kerb.

Pro tip from coaching circles: chase the peak force through corners. Keep the fronts laterally loaded at their optimal slip angle, feel for maximum self-centring resistance, hold it there. Steering goes light? You've pushed too far. This only works if you can feel the variation, which brings us to the most important setting.

Clipping: the enemy of detail

Clipping happens when the force the sim wants to send exceeds what your wheelbase can actually deliver. Everything above the limit gets flattened. Heavy, dead steering with no variation. No information. Just resistance.

Some clipping's fine. Half a second over a big kerb at Monza, who cares. But if you're clipping through the first half of every corner entry, you're missing all the grip information that would actually make you faster.

This is where people with low-torque bases (2-6 Nm) make a classic mistake: they crank the in-game gain until the wheel feels heavy. Wheel clips constantly, FFB signal's dead. You'd be better off running lower gain, feeling lighter forces, and preserving the variation. That's where the lap time lives.

Why more torque matters, even if you do not use it all

Most competitive sim racers run between 10-15 Nm of actual force. So why buy a 25 Nm base?

Headroom. I ran a Simucube 2 Pro at 15 Nm for months, then switched to the same 15 Nm on a borrowed Moza R16 V2 maxed out. Night and day. The SC2 Pro's bigger motor delivers force faster, holds it with more precision, and never thermally clips during a long stint. When a motor's cruising at 60% instead of 100%, the response curve stays more linear and you get sharper detail in the quiet bits.

Same reason a 200-watt speaker amp playing at 50 watts sounds cleaner than a 50-watt amp cranked to its limit.

Real-world force comparisons

To put the numbers in context, here is what actual steering forces feel like across different vehicles. These come from coaching data and vehicle dynamics measurements:

The 10-15 Nm range is where most serious drivers land. Strong enough to feel everything with precision, not so strong that it'll damage your wrists over a four-hour endurance stint.

Encoder resolution and motor control

Two specs that matter more than most people realise:

Encoder resolution tells the motor controller where the shaft is. A 16-bit encoder gives 65,536 positions per revolution. A 22-bit one gives over 4 million. That's the difference between knowing your shaft position to 0.005 degrees and knowing it to 0.00009 degrees. In practice: smoother torque at low speeds, finer texture reproduction, less cogging. Most modern DD bases sit at 20-22 bit. You won't notice much between 20 and 22, but both are a massive step up from 16.

Slew rate is how quickly the motor can swing its torque output, measured in Nm/ms. High slew rate means the base snaps to sudden changes: kerb strikes, snap oversteer, rapid weight transfers. The SC2 Pro does 8.0 Nm/ms. SC2 Ultimate manages 9.5 Nm/ms. Low slew rate? FFB feels blurry and laggy, like the motor's always a beat behind what the sim's asking for.

PWM frequency (pulse-width modulation, how the controller delivers power to motor windings) affects feel too. Below about 20 kHz you can hear the motor whine and sometimes feel a graininess in the torque. Higher PWM frequencies push that noise into ultrasonic range and deliver smoother current to the coils.

Setting it up: the practical approach

FFB setup isn't about making the wheel feel strong. It's about maximising the useful information reaching your hands. Here's what works for me, after years of fiddling and talking to coaches who race on these things daily:

Step 1: Baseline

Zero everything. No damping, no spring effects, no filtering in the wheelbase software. Raw signal to your hands. You want to hear what the game actually sends and what the base can reproduce before you start processing anything.

Step 2: Set your base force

Find a force level you're comfortable with for a two-hour stint. For most people that lands around 10-12 Nm. Don't max out the base. A Moza R5 or R9 at 100% will overheat during a long race, and if something bugs out at full power it'll wrench your hands. Not worth the risk.

Step 3: Match the gain in-game

iRacing makes this easy. Set Wheel Force to match your base output (say 11 Nm), click Auto. It'll adjust per-car gain so you're using 80-100% of your range without clipping badly. In ACC, Le Mans Ultimate, and Assetto Corsa you'll need to dial in-game gain manually until clipping only triggers on kerb strikes.

Step 4: Address signal quality

Think of FFB like an audio signal. Feels sharp and peaky, metallic reverb through the rig? Peaks are too aggressive. Add a small amount of filtering (Fanatec calls it interpolation, Simagic calls it filter level, same concept). Go just high enough that sharp edges become comfortable but you can still distinguish kerbs from tyre scrub from surface changes. Over-filter and everything turns to mush.

Step 5: Minimal effects

Understeer indicators, enhanced road effects, spring effects, all that stuff? Distracting. Every esports driver I've talked to runs the least effects possible. The raw sim signal is the closest thing you'll get to what the tyre model's actually doing. If you need something, 10-15% dampening can smooth nervousness without killing detail.

Step 6: Keep it consistent

Pick a force level and stick with it across all cars. Real life, switching from a GT3 to a Formula car feels completely different because your body, posture, g-forces all change. In a sim, your physical environment stays the same. Consistent force level builds muscle memory. Adjust per-car gain to fill the same force window on your base rather than changing base strength between cars.

Why direct drive changed everything

DD didn't just bolt on more torque. It changed the fundamental quality of what reaches your hands. Five things happened at once:

Zero mechanical filtering. No belt absorbing the high-frequency stuff. No gear mesh chewing up the signal. Physics engine calculates it, you feel it. Nothing in between.

Latency effectively gone. Motor shaft is the steering shaft. Between the electrical signal and mechanical torque there's only the motor's own response time, which sits in single-digit milliseconds.

Dynamic range. A 25 Nm DD base reproduces everything from a subtle road crown to a full ABS intervention without running out of headroom. My old Fanatec CSL DD clipped on half the corners at Spa. The SC2 Pro at the same force level barely touches the ceiling.

Sustained torque. Industrial servo motors are built to run all day. No belts wearing. No gears overheating. FFB at hour four of an endurance race feels the same as it did on lap one.

Bidirectional precision. Same motor pushes the wheel, holds it, and pulls it back. With a 20+ bit encoder providing millions of position readings per revolution, the controller applies torque with a precision belt systems can't get near.

Common mistakes

Running too much gain. Wheel feels heavy and impressive, but it's clipping constantly. You lose all the subtle grip information that makes the difference between a good lap and a great one. Dial it back until clipping only happens on kerb impacts.

Fighting the wheel. Death-gripping the rim with locked arms. The FFB's trying to communicate. Fight it and you can't hear what it's saying. Relax your hands. Let the wheel move slightly. Feel the changes.

Over-filtering. Adding 30% damping, 20% friction, maximum smoothing. The FFB now feels "nice" but you've destroyed the signal. Can't tell a kerb from a gear change from a surface transition. Start with zero filtering and add the minimum needed to remove any harshness.

Changing settings constantly. Every time you adjust your FFB settings, you reset your muscle memory. Find a profile that works, commit to it, let your body learn the language. You'll be faster in a week than if you keep tweaking every session.

Ignoring per-car gain. Different sim cars produce wildly different levels of FFB output. A GT3 car might produce 15 Nm of virtual steering torque, while a road car produces 5 Nm. If your base is set to 12 Nm and the road car only sends 5 Nm, you're using less than half your range. Adjust per-car gain so every car fills the same force window on your base.

Where to go from here

If you are looking for a direct drive wheelbase, the buyer's guide covers every base on the market with live pricing, normalised specs, and our 7-axis rubric scoring. Every base page includes community-sourced FFB settings for iRacing, ACC, Assetto Corsa, Le Mans Ultimate and more.

For head-to-head comparisons, the comparison tool puts any two bases side by side with specs, prices, and a comparative verdict.