Stick Drift Test – Check Controller & Joystick Drift Online

This stick drift test measures how far each analog stick moves when nobody is touching it. Connect a controller, press any button so the browser can see it, set it down and start the test: for five seconds the page records both sticks and reports the peak deflection, resting offset, jitter and the smallest deadzone that hides the drift. A run history, a deadzone preview and a snap-back test help you act on the result. It works with DualSense, DualShock 4, Xbox, Switch Pro, Joy-Con and most PC gamepads, and nothing is uploaded: readings stay in this tab and disappear when you close it.

How to run the controller drift test

Connect and wake the controller. USB or Bluetooth both work. Browsers hide gamepads until a button is pressed, so press one.
Put the controller down flat and take your thumbs off both sticks. Grip pressure in your hands looks exactly like drift.
Press Start drift test. After a three-second countdown both sticks are sampled for five seconds, and a dashed coloured ring on each plot marks the furthest the stick moved on its own.
Run it two or three times. The run history keeps the last five runs with the peak per stick and the change from the run before.
Set the deadzone slider to your game's value and read the verdict per stick.
Flick each stick to the edge in all eight directions and let go for the snap-back test.

To check buttons, triggers and vibration too, use the full gamepad tester.

What stick drift is and why it happens

Stick drift is input the controller reports while the stick is at rest: the character creeps forward, the camera pans, a menu scrolls by itself. Most analog sticks measure position with two potentiometers, one per axis: a metal wiper slides along a resistive track, and the voltage at the wiper tells the controller where the stick is. Drift is anything that moves the reading away from the stored centre without the stick moving. Three causes produce it, and they need different fixes.

Potentiometer wear

Every movement rubs the wiper over the track and removes a little material, most of it around the centre. As the track wears, the resistance at rest changes and no longer matches the stored centre. Wear drift grows slowly over months, points in a consistent direction and often comes with a jumpy reading. Cleaning can hide it for a while; it cannot rebuild the track.

Dust and debris

Dust, skin oil and dried drink residue get between wiper and track. Debris drift comes and goes, changes direction, and often improves right after the stick is rotated hard or cleaned, which is why compressed air sometimes works and why the effect can fade.

Calibration offset

Sometimes the hardware is fine but the stored centre is wrong, after a firmware update, a module swap or from the factory. Calibration drift is a steady offset with low jitter, the same on every run, and recalibrating fixes it for free.

Reading your stick drift test results

Values are on the Gamepad API scale, where each axis runs from −1 to +1 and full deflection is 1.000.

Peak deflection is the largest distance from centre recorded in the five seconds. It sets the grade because it is what a game has to hide: up to 0.030 is no drift, up to 0.070 minimal, up to 0.150 mild drift, above that drift.
Resting offset is the average x and y position. A large offset with small jitter points to a calibration or wear shift in one direction; a near-zero offset with a high peak points to noise or intermittent contact.
Jitter (RMS) is how much the reading wanders around that average. Healthy sticks sit in the low thousandths; jitter that jumps between runs is typical of debris.
Suggested deadzone is the peak plus a 0.02 margin, rounded up to the next whole percent.

A reading of exactly 0.000 does not always mean a perfect stick: some drivers apply a small deadzone before the browser receives the data. Compare runs on the same computer and browser.

The snap-back test: does the stick return to centre?

A hands-off test only shows where the stick rests once it has settled. In a fight you notice something else: you let go and the aim keeps sliding for a moment, or stops slightly off centre. Push a stick to the rim and lift your thumb away. When the reading falls from the edge to below 0.2 within 150 ms, the page counts a release (a slow return guided by your thumb is ignored) and records:

Offset at 300 ms: the average distance from centre 300–350 ms after release. Below 0.05 is clean; above 0.10 the stick is stopping noticeably off centre.
Settle time: how long until the reading drops under 0.05 and stays there for 100 ms. A stick that never gets there within 1.5 s is marked as not settling.
Overshoot: how far the stick swings past centre to the other side.

The eight compass cells show which directions you have tested and the worst result in each. One bad direction usually means a weak spring or a worn track on that side. Timing resolution is limited by how often the browser receives new controller data, shown under the test.

Choosing a deadzone: axial vs radial

A deadzone is the area around centre where a game ignores stick input, and the cheapest drift fix there is: if it is larger than your peak deflection, the drift is invisible. The cost is precision, because small intentional movements inside it are ignored too. In the deadzone preview the grey ring shows the slider value next to your measured peak. The verdict reads yes when the deadzone covers the peak with at least 0.02 to spare, barely when the margin is smaller, and no when the drift reaches past it.

Radial deadzones ignore the stick while its distance from centre is below the threshold, keeping diagonals intact. The preview draws this kind.
Axial deadzones test each axis separately, creating a cross-shaped dead area that makes the stick snap to straight up, down, left and right near centre. Since neither axis can exceed the distance from centre, an axial deadzone of the same size hides at least as much drift.
Scaled radial deadzones rescale everything outside the threshold back to the full range, so input starts smoothly instead of jumping.

For reference, Microsoft's XInput headers define default thumbstick deadzones of 7849 and 8689 out of 32767, about 24% for the left stick and 27% for the right. That is why many PC games hide drift the preview flags at 5–10%, and why players who lower the deadzone notice drift first.

How to fix stick drift, from free to paid

Rule out software. Test in a second browser or on another computer. Unplug other controllers, wheels and flight sticks: games can read input from a device you forgot about.
Raise the in-game deadzone to the value the preview recommends. It costs nothing and is fully reversible.
Recalibrate. On Switch, go to System Settings → Controllers and Sensors → Calibrate Control Sticks. Steam's controller settings include calibration and deadzone options for many pads. For DualShock 4 and DualSense, open-source tools can rewrite the stored centre over USB; they fix a calibration offset, not a worn track.
Update the firmware. Xbox controllers update through the Xbox Accessories app; a DualSense updates from a PS5 or with Sony's firmware updater for Windows.
Clean the stick. With the controller off, blow compressed air around the base of the stick while rotating it, or work a little electronic contact cleaner or isopropyl alcohol into the gap and rotate the stick to spread it. Let it dry fully before use.
Use the warranty. A drifting stick on a controller still under warranty is a claim to make before you open anything, because opening the shell usually ends it.
Replace the stick module or the controller, as below.

Platform-specific steps are on the PS5 controller test, the Xbox controller test and the Joy-Con drift test pages.

Hall-effect and TMR sticks

Hall-effect sticks replace the wiper and track with a magnet and a sensor that reads the magnetic field, so nothing rubs and nothing wears in the way a potentiometer does. TMR (tunnel magnetoresistance) sticks are also contactless and read the field with a different sensor type, which manufacturers promote for lower power draw on wireless pads. Both remove the most common cause of drift, but they still have springs, a gimbal and a stored calibration. They can show a small resting offset, a sloppy return to centre or a poor calibration, and this test measures those the same way.

When to replace the stick module

Replace, rather than keep cleaning, when the test shows any of these:

Peak deflection keeps climbing across runs over days or weeks, even after cleaning.
The resting offset points the same way every time and sits above the deadzone you are willing to use.
Snap-back fails in one direction every time, which points to a worn spring or track rather than dirt.
Jitter stays high after cleaning, which means the track itself is damaged.

On most controllers, including the standard DualSense and the Xbox Wireless Controller, the stick module is soldered to the main board, so replacing it means desoldering, a skill-intensive repair. Joy-Con sticks connect with a ribbon cable and swap without soldering, and the DualSense Edge takes replacement stick modules sold by Sony that swap in without tools. If a replacement still drifts, check the solder joints and the board traces around the new module before blaming the part. For flight sticks and other non-standard devices, the joystick tester shows every axis individually.

Frequently asked questions

Can stick drift be fixed permanently, or will it keep coming back?

It depends on the cause. A calibration offset is fixed for good by recalibrating, and debris drift can stay away after a thorough clean. Drift from a worn potentiometer track comes back, because cleaning does not restore the track; the permanent fix is a new stick module, ideally Hall-effect or TMR. Rerun this test in the weeks after a fix: a peak that creeps back up means wear.

My controller only drifts in one game. Is the controller broken?

Probably a small drift that most games hide and this one does not, because each game picks its own deadzone. Run the drift test and set the deadzone preview to that game's value: if the verdict is no or barely, raise the game's deadzone to the suggested value or set one in Steam Input. Also unplug other USB input devices.

My Xbox controller drifts on Windows but not on my Xbox. Why?

The console and the PC game are applying different deadzones, or the PC is reading a second input device. Run this test on the PC: if both sticks read close to zero at rest, the fix is in software. If the peak is above a few percent, the controller does drift and the console's default deadzone is hiding it.

I calibrated my controller and it still drifts. What should I try next?

Then the drift is not a stored-centre problem but debris or wear. Clean the stick, compare runs in the history, and raise the in-game deadzone to the suggested value meanwhile. If the peak stays high after cleaning, the module is worn: claim the warranty or replace the module.

Should I raise or lower the deadzone to stop drift?

Raise it. The deadzone is where input is ignored, so a larger one hides more drift and a smaller one exposes it. Set it just above the measured peak, to the suggested value, rather than guessing high, so you give up as little precision as possible.

Does blowing compressed air into the stick actually fix drift?

Sometimes, when the cause is dust between the wiper and the track. Blow air around the base while rotating the stick, then rerun the drift test and compare the peak in the run history. If it stays down for days, it was debris; if it returns quickly, the track is worn.

I replaced the joystick module and the drift is still there. What went wrong?

Check the board, not the new part. Desoldering can lift pads or break the thin traces around the module, and a damaged trace or weak joint gives the same symptoms as a worn stick. Inspect and reflow the joints, repair broken traces, and recalibrate, since the new module does not match the old stored centre.

A few minutes after plugging in, my new controller's stick stops reaching the edge in one direction. Is that drift?

Not in the usual sense: drift is movement at rest, while this is lost range, which shows up as a flat or dented section when you rotate the stick along the edge of the plot. Watch the trace and circularity readout, then run the snap-back test in that direction. If it only happens on one computer and clears on reconnect, suspect the driver or USB connection first.