Skip to content

Pedal calibration: deadzone, pressure curves, and brake feel

A load cell measures the force you push with; a potentiometer or hall sensor measures pedal travel. That single distinction drives everything you calibrate. A load cell reads the same force lap after lap regardless of where the pedal sits, so you brake to a pressure target instead of a position your foot has to remember. The Fanatec CSL Pedals Load Cell Kit and ClubSport Pedals V3 use a 90 kg load cell; the Heusinkveld Sprint uses a 120 kg sensor. A travel-based brake can’t tell 70 kg of stomp from 80 kg, which is why those pedals feel identical from 80% to 100% travel.

Calibrate in this order, or your numbers will fight each other: Windows or the vendor app first, then the sim. Open joy.cpl (or Fanatec Control Panel / Moza Pit House / Simagic SimPro) and confirm the OS sees full pedal range. Then calibrate inside the game.

In iRacing, set 100% a touch above your hardest realistic corner braking, not a one-time max effort, and with headroom left so a genuine panic stop still has somewhere to go (more on that below). iRacing stores the result in joyCalib.ini; you can edit and lock those values so a bad recalibration on session join doesn’t undo your work.

Set 100% at a force you can repeat, not the cell’s rating. The 90 kg rating is the sensor ceiling, not a target. Most racers put 100% at roughly 60-70% of cell capacity (around 55-65 kg on a Fanatec). Holding full 90 kg every corner is too taxing for anything past a sprint and nets no faster lap times. On a Fanatec, the BRF (Brake Force) setting in the Tuning Menu sets what force equals 100%; CSL LC bottoms out around 10 kg minimum sensitivity.

These are different controls and people conflate them constantly:

  • Deadzone. Travel at the start of the pedal that’s ignored, used to kill creep, sensor noise, or a resting offset.
  • Endpoint (max). The point where the pedal reads 100%.

A “90 degree deadzone” on a Logitech G920 or G29 is almost always an endpoint/calibration problem, not a real deadzone: the OS or sim never saw the pedal’s full travel, so the first chunk of motion does nothing. Recalibrate from Windows first to fix it.

If pressing the brake also shows throttle (the “100% brake, 25% throttle” complaint on a Thrustmaster T3PA set), that’s axis crosstalk or a wrong axis mapping. The brake is feeding the throttle axis. Clear the bindings and remap each pedal to its own axis, then recalibrate per game. If recalibration alone doesn’t fix it, work through pedal troubleshooting.

The brake input is linear by default. The curve reshapes how force maps to in-game brake:

  • Concave / S-curve. More resolution in the light-to-mid range. This is what helps trail braking and threshold modulation, where you’re feathering off the last 10-20%.
  • Convex. The brake bites early, useful if your pedal is too stiff to reach high force.

Set it in the vendor app (Fanatec Control Panel, Moza Pit House, Simagic SimPro) or per-sim where supported. Start linear, then add a mild S-curve only if you can’t modulate the top of the brake smoothly.

Changing brake feel: springs, elastomers, hydraulics

Section titled “Changing brake feel: springs, elastomers, hydraulics”

Feel comes from the force-to-travel relationship, and you tune it with the spring and elastomer stack.

Elastomer relaxation is real. Rubber elastomers soften with heat and time, shifting the force-to-travel curve within a session and over weeks. Your brake point quietly moves. Die springs (e.g. 20mm OD, 25mm or 35mm length) are far more stable. A common fix is stacking a spring with an elastomer (a “sandwich” on the CSL LC) to keep progressive feel without the drift. Heusinkveld’s elastomers are regarded as high quality and hold up well, but the principle stands.

A staged stack does more than tune drift. It gives you distinct feel at the braking zones that matter. Pairing springs and elastomers so resistance steps up around light, mid, and near-maximum braking lets your foot feel roughly where 30%, 60%, and threshold braking sit, instead of one smooth ramp you have to guess at. That tactile staging is often what people are chasing when they say a brake “feels real,” more than any single material.

Hydraulics (Asetek Invicta, mBooster-style) run a piston and fluid chain to a pressure sensor. The engineering critique is blunt: they add a different way to transmit force to a system that doesn’t need it. A good load cell with a spring-plus-elastomer stack gets most of the way there, and you can replicate the rest with input curves. Hydraulics make sense if you specifically want that progressive fluid feel and don’t mind the complexity.

  • joy.cpl (Windows Game Controllers). Verify full travel and OS-level calibration first.
  • DiView (Heusinkveld). Sets deadzone and endpoints on any analog HID device, not just Heusinkveld pedals.
  • DXTweak2. Manual axis calibration and deadzone editing for any DirectInput device.
  • Leo Bodnar USB load cell interface. For DIY pedals, gives a clean HID axis from a raw load cell.

Calibrate at the OS level with one of these, then let the sim’s calibration ride on top of a known-good range.

Frequently asked questions

What force should I calibrate 100% brake to?

Set 100% at a force you can repeat, not the cell's rating. Most racers put 100% at roughly 60-70% of cell capacity, around 55-65 kg on a 90 kg Fanatec cell. Holding the full rating every corner is too taxing past a sprint and nets no faster laps. On Fanatec, the BRF (Brake Force) setting in the Tuning Menu sets what force equals 100%.

In what order should I calibrate, Windows or the sim first?

Windows or the vendor app first, then the sim. Open joy.cpl (or Fanatec Control Panel / Moza Pit House / Simagic SimPro) and confirm the OS sees full pedal range, then calibrate in-game. In iRacing, calibrate the brake to 100% with your maximum realistic stomp; iRacing stores it in joyCalib.ini, which you can edit and lock so a bad recalibration on session join doesn't undo your work.

Should I use a brake pressure curve, and which way?

Start linear. Add a mild concave/S-curve only if you can't modulate the top of the brake smoothly. It gives more resolution in the light-to-mid range, which helps trail braking. A convex curve makes the brake bite early, useful if the pedal is too stiff to reach high force. Set it in the vendor app (Fanatec Control Panel, Moza Pit House, Simagic SimPro) or per-sim where supported.

Why does my brake point keep moving over a session?

Rubber elastomers relax with heat and time, shifting the force-to-travel curve within a session and over weeks. Die springs (e.g. 20mm OD) are far more stable; a common fix is sandwiching a spring with an elastomer to keep progressive feel without the drift.

Should I calibrate 100% to the hardest I can ever push?

No. Leave headroom for panic braking. Calibrate 100% to your repeatable threshold-braking force, then keep some force in reserve above it. One common approach is a 50 kg max so normal threshold braking lands around 75-80% and a genuine stomp still has somewhere to go. If 100% is set at your absolute maximum, every emergency brake just saturates at 100% and you lose the top of the modulation range exactly when you need it.

Should I run a lighter brake on a wheel stand or for long stints?

Yes. A wheel stand or folding rig flexes under a stiff brake, and a hard pedal that's fine for a 20-minute sprint becomes fatiguing over an endurance stint. Calibrate 100% to a lower force and run a softer spring/elastomer stack so the rig isn't fighting you and your leg lasts the race. You lose a little high-force resolution and gain a pedal you can drive for an hour straight.

My brake feels wrong after calibrating. How do I tell if it's the hardware or the sim?

Verify it in the vendor app first (Fanatec Control Panel, Moza Pit House, Simagic SimPro), then load a *different* sim. If the raw force bar in the vendor app moves cleanly and the brake behaves correctly in a second sim, the hardware is fine and your first sim's calibration or curve is the problem. Reset and recalibrate there. If it's wrong everywhere, including the vendor app, it's a hardware, mounting, or driver issue. Isolating the layer this way saves you from re-tuning a curve that was never the cause.