Mouse, Keyboard & Controller Faults — Test, Diagnose, Decide What to Do
Every input device fails in one of three ways: it reports movement you did not make, it repeats a press you made once, or it swallows a press you did make. None of them look like hardware at first — they look like a laggy game, a sticky setting, or a dirty desk. This hub walks the whole path: confirm which device is actually at fault, read what the verdict means, then decide whether to repair it, replace it or return it. Every test runs in your browser and nothing leaves your device.
The three faults, and what they get mistaken for
These problems drag on for months because the symptom points somewhere other than the cause. A creeping aim gets blamed on the game, a doubled letter on autocorrect, a double-click on the mouse settings panel. Match what you actually saw against this table before you touch a single setting.
| What you see | What it usually gets blamed on | What it usually is | Confirm with |
|---|---|---|---|
| Your aim creeps while nobody is touching the controller | The game deadzone, or unsteady hands | Stick drift: the potentiometer no longer returns to centre | controller drift test |
| One letter appears twice from a single press | Keyboard repeat rate, autocorrect, a sticky OS setting | Chatter: the switch contacts bounce and close twice | keyboard chatter test |
| One click opens, selects or fires twice | The double-click speed slider, or the app itself | A worn microswitch under the button | mouse double-click test |
| A key does nothing, but only during fast typing or gaming | The game, the USB port, the operating system | Rollover limits: the keyboard cannot report that many keys at once | keyboard rollover test |
Step 1 — Confirm which device is at fault
Run the matching tool before you change any setting, because changing settings destroys the evidence: once you raise a deadzone or slow the repeat rate, you can no longer tell whether the hardware was ever at fault. Each of these samples raw events in the browser — no install, no account, no upload.
Controller drift test
Sample both sticks every frame and read the resting value a warranty agent will ask for.
Keyboard chatter & rollover test
Find the one failing switch, and see how many keys your board can report at once.
Mouse double-click test
Measure the gap between clicks per button and catch a worn switch before it gets worse.
If you have a second mouse, keyboard or pad, test both on the same machine. Two devices showing the same problem points at the port, the cable, the OS or the app; only one showing it points at the hardware. That single comparison settles more arguments than any threshold ever will.
Step 2 — Read the verdict, not a single frame
Every tool here ends in one of five states, and the one most people misread is "not enough data yet". That state is neither a pass nor a fail — it means the sample is too small to know, and the honest answer at that point is to keep going. We deliberately suppress both conclusions until the evidence exists: a handful of frames that happened to spike used to report a fault, and a handful that happened not to used to report a healthy device. Both were the same overconfidence in opposite directions. The full verdict model, every threshold and where each number comes from are on the methodology page.
Step 3 — Decide: repair, replace or return
This is the step most testers stop before, and it is the one that actually costs money. A confirmed fault is not automatically a reason to buy a replacement: cleaning is free, a switch is a few dollars, and a warranty claim is worth more than both if the device is still inside the window.
| What the tests keep showing | Do this first | Then this | Stop and replace when |
|---|---|---|---|
| A single warn, then ok on a re-run | Nothing — that was noise | Re-test in a week if it comes back | It never comes back |
| warn on two separate runs at the same threshold | Clean it: contact cleaner or compressed air | Swap the switch or the stick module | The part costs more than the device |
| bad on one device, ok on another | Save the result — that is your evidence | Open a warranty or return claim | The claim is refused and repair exceeds replacement |
| insufficient, again and again | Give it more samples; the tool is saying it does not know | Try another port, cable or machine | A second device passes where this one never will |
Controller faults
Stick drift is the one input fault that worsens slowly enough to be denied for months. How to test controller drift covers connecting the pad and reading the numbers like a repair tech; once the fault is confirmed, the stick drift fix guide walks the cleaning-versus-replacement decision with the warranty maths included. Buying second-hand? Run the used controller check before you pay, not after. Aim feels floaty but the drift test reads clean? The controller dead zone test splits the problem into resting drift and true dead travel, which is the difference between a module repair and a settings change.
Keyboard faults
Chatter is a switch problem: one contact bounces and the board registers two presses. Ghosting is a different problem entirely — the key matrix cannot report that combination — and no amount of cleaning will fix it. How to test keyboard chatter isolates the failing key, how to fix keyboard chatter covers clean, swap and claim, and rollover and ghosting explained tells you which limits are worth replacing hardware over.
Mouse faults
The double-click bug is almost always a worn microswitch, not a setting — raising the double-click speed only hides it until the switch stops registering altogether. How to test your mouse for the double-click bug shows which gap values matter; how to fix the double-click bug orders the options by cost and risk. If the problem is not the switch but the feel — sensitivity that will not carry over to a new mouse — measure it properly with how to check your mouse DPI.
What this hub deliberately leaves out
Webcams, microphones, speakers, screens and phone sensors have their own tests on this site, and they are not listed here on purpose. Input devices fail in ways you can prove with numbers; most of the rest fail in ways you prove by substitution. Folding everything into one "test your gear" list is exactly how a site ends up looking like it is about nothing in particular. If your problem is not a mouse, keyboard or controller, start from the full guide index instead.
Every verdict here comes from one person with a documented method. Who he is and how the thresholds were chosen: About GearLint.
Frequently asked questions
My controller drifts in one game but not another. Is it still drift?
Yes, and the difference is the deadzone, not the hardware. Every game swallows a different amount of stick movement near centre: a wide deadzone hides early drift completely, a tight one exposes it as aim that creeps. The test reads the raw resting value so you can predict which of your games will feel it — drift that only shows up in one title is still drift, it is simply drift that one title stopped hiding.
How many times should I run a test before I believe the result?
Twice, on separate runs, at the same threshold. One flagged event is noise: a hand brushing a stick, a key pressed at an odd angle, a click released slightly early. Two runs agreeing is the point at which a fault stops being a coincidence. This is also why the tools show "not enough data yet" instead of a verdict when the sample is small — a conclusion drawn from a handful of events is a guess either way, and a guess dressed as a measurement is the thing this site exists to avoid.
Can I fix a double-clicking mouse without soldering?
Honestly, no — not permanently. Software debounce and the double-click speed slider mask the symptom by making the driver ignore a second press inside a short window, which also adds latency and eventually stops working as the switch wears further. Cleaning the switch with contact cleaner can buy weeks. A permanent fix means replacing the microswitch, which means soldering, or claiming the warranty if the mouse is still inside it.
A key types two letters — is the switch definitely broken?
Not definitely, and the order of checks matters. Rule out the operating system first: keyboard repeat rate and delay settings, accessibility filters, and any macro utility. Then reproduce it in a different app and, if you can, on another machine. If the doubling follows the keyboard and not the computer, run the chatter test on that one key — the gap between the two events is what separates a bouncing contact from a software repeat, and software repeats arrive at a perfectly regular interval while chatter does not.
Is any of my test data uploaded?
No. Every tool reads raw events in your browser tab, computes the verdict there and keeps it there. There is no account, no telemetry, and no upload endpoint for test results. The only network request a tool page makes is the anonymous community baseline, and that only happens if you explicitly press the share button.
I have no idea which device is at fault. Where do I start?
Start from the symptom, not the device. Movement you did not make points at a stick or a sensor; a press that registered twice points at a switch; a press that vanished points at rollover, a dead switch or the port. Run the matching tool for whichever of those you actually saw, and if two devices could explain it, test both on the same machine — the one that reproduces the fault is the one that has it.