Match the check to your symptom
Start with what you can reproduce. A wheel that turns, a wheel that clicks and a page that moves are separate observations.
Scroll jumps backward during a steady turn
Compare received direction with the direction you intended. Keep the expected sign fixed for that run.
Open the wheel direction test →The wheel moves but nothing is recorded
Keep the pointer inside the capture area and try again. Compare another app before interpreting a zero.
Check the capture area →Only one page jumps or stutters
Compare the same action on a plain page, in another browser and in another app. Keep those observations separate.
Follow the comparison sequence →The movement is sideways or diagonal
Inspect horizontal and vertical input separately. A zero vertical value may accompany a horizontal event.
Inspect the horizontal event counters →Pressing the wheel fails, but turning works
Use the mouse button tool for the middle press. Its result answers a different question from wheel rotation.
Check the middle mouse button →Record a clean single-direction run
Keep deliberate reversals out of the measurement. The browser receives input; it cannot know which direction you meant to turn.
- Open the wheel test. Keep the pointer inside its capture area, turn once and read the deltaY sign in the log. The capture area intentionally stays still.
- Select that expected sign, then reset. Turn slowly in the same direction several times without deliberately reversing. Note your approximate pace and whether you used ratchet or free-spin mode.
- Stop and inspect opposite-sign events alongside the raw log. Keep deltaMode with each value: 0 means pixels, 1 lines and 2 pages. One browser event is not one physical notch.
- Save or download that run before resetting. For the other direction, read its sign, select it and reset again. Changing the sign selector reclassifies the current run; it does not start a new one.
Interpret the result before changing settings
Compare the recorded sequence with your action and the symptom in ordinary use. A clean short run does not exclude an intermittent problem.
- Opposite-sign events: repeat after a reset and check that you did not reverse deliberately. Repetition is worth documenting, but does not identify the faulty part.
- Stable input here, jumping in one app: compare that app and its scroll behavior. A wheel event and actual content movement are not the same measurement.
- Zero events: this run contains no wheel measurement. Check pointer placement and repeat elsewhere. The deliberately stationary capture area is not evidence of a failure to scroll.
- Vertical zero with horizontal input: inspect the axis counters. Compare counts within the same protocol; do not add amplitudes expressed in different units or treat them as page travel.
Compare one condition at a time
Save a baseline before changing anything. Repeat the same direction and approximate pace after each change, and keep the other conditions as similar as possible.
- App or browser: repeat in a plain document and a second browser. Note whether smooth scrolling, natural scrolling, extensions or button remapping were enabled. Change one setting only, repeat, and restore it if it does not help.
- Connection: if applicable, compare a direct connection and another USB port. For a wireless device, record its charge level and connection mode before repeating.
- Device or computer: try another mouse here, then this mouse on another computer if available. Record which combination reproduces the symptom; a difference narrows the comparison without proving a cause.
When to use model-specific support
If the symptom persists, look up the exact model and operating system on the manufacturer’s support site. Follow its cleaning and update instructions; an online wheel log cannot establish whether the encoder needs replacement.
Check model compatibility before installing software or firmware. If cleaning or opening the mouse is being considered, consult the manufacturer’s procedure and warranty terms first. Keep the support request tied to your repeatable symptom.
Read Logitech’s scroll troubleshooting reference →
This reference is for Logitech devices. Other brands and models can require different steps.
Keep evidence someone else can reproduce
Save separate reports for each condition in Device Lab. Download the observations if you want a portable copy. Saving is optional and stays in this browser; nothing is uploaded by these tools.
- Conditions: model, operating system, browser, connection, direction, approximate pace, scroll mode and the one variable changed. Use report labels or notes; the tool does not detect all these conditions automatically.
- Observations: the expected sign, total and opposite-sign events, axis groups and units. The JSON summary covers the full run; only the latest 250 raw wheel events are retained. Keep runs short if you need the complete raw sequence.
- Outcome: where the symptom repeated and where it did not. A recording of the page symptom can complement the input report; neither is a hardware certification.