A gear rotation puzzle is solved by placing every candidate gear into a fixed socket so its contacts form one connected, conflict-free train from the clockwise driver to the target. In Gear Rotation Puzzle, that task unfolds across five deterministic levels: each board has a fixed clockwise driver on the left, a target on the right, and a small set of candidates and empty sockets between them. You assign every candidate to exactly one socket, and every socket must be filled before the train can be tested. The board does not compare your arrangement with a hidden answer key; it calculates which gears physically touch from their coordinates and visible radii. An external contact reverses rotation, while the target must also be reachable from the driver without a contradictory direction or speed. A valid train awards 200 points and opens the next gearbox. Five valid layouts produce exactly 1,000 points. If a full arrangement fails, the game explains whether the target is disconnected, a gear is isolated, or a direction conflict exists. You can recover from one distinct failed layout, but a second different failure closes that run, so checking geometry before submitting matters.
Prepare the fixed gearbox board
Begin by locating the fixed clockwise driver on the left and the target on the right. The rack contains three or four candidates with different visible sizes, and each size corresponds to the gear's underlying tooth count. The empty locations are fixed sockets. This is a complete placement assignment rather than a one-gear direction question: every candidate goes into exactly one socket, no socket remains empty, and every placed gear must belong to the driver's powered component.
The five levels are deterministic. Candidate radii, tooth counts, socket coordinates, driver dimensions, and target dimensions stay unchanged between runs, so a failed arrangement can be revised deliberately. The game runs locally in your browser and requires no account, download, upload, API, remote model, random service, or hardware permission.
Solve a gearbox step by step
- Read the board. Locate the driver, target, candidate rack, and empty sockets. The keyboard addresses sockets from left to right with Q, W, E, and T, so note which positions are available on the current level.
- Hold one candidate. Select a gear from the rack with a pointer, touch control, or number key from 1 through 4. Once that gear is placed, its rack control becomes disabled, preventing the same candidate from being used twice.
- Place it in a fixed socket. Click or tap an empty socket, or press its corresponding Q, W, E, or T key while the candidate is held. Each candidate must occupy exactly one socket.
- Complete the assignment. Fill every remaining socket. To move a placed gear, hold no candidate, clear its occupied socket, then select the replacement and place it. Selecting, removing, or replacing a candidate does not count as an error by itself.
- Trace the powered route. Follow the visible gear sizes and contact gaps from the driver. Make sure every placed gear belongs to the powered chain and that the target is actually reached by that chain.
- Check the reversals. Every external mesh flips rotation. Use contact parity as an early warning, especially around closed loops, but rely on the final connectivity and conflict checks.
- Test the completed train. Press Enter when every socket is occupied. A partially filled board cannot be submitted. A valid train turns the target, reports its clockwise or counterclockwise direction, and displays the ideal target-to-driver speed magnitude.
- Repair or advance. If the test fails, use its diagnostic feedback to revise the arrangement. One distinct failed full layout is recoverable. If the train is valid, it awards exactly 200 points and opens the next fixed gearbox.
Separate visual proximity from actual contact
The solver does not rely on a hidden list of accepted arrangements. Two external gears create a contact edge when the distance between their centers matches the sum of their radii within the game's declared tolerance of 0.75 game units. A gear that merely looks close to its neighbor can therefore remain disconnected. Visible size affects both the possible reach of a gear and its tooth-count ratio, but filling every socket only proves that the board is occupied. It does not prove that the gears form a powered train.
Follow the chain from the driver rather than judging the layout as a whole. Each permitted contact reverses direction, and an idler can bridge a physical gap while adding another reversal. A chain that merely ends near the target is not enough: the target must touch a gear in the driver's powered component, and every other placed gear must join that component too.
See what the solver checks when you test
Validation begins at the driver with signed speed +1 and proceeds through geometric neighbors using breadth-first propagation. The board visits every gear that actually touches another gear, assigns the expected opposite direction, and adjusts speed according to tooth counts. Its rule is ω_neighbor = −ω_current × teeth_current ÷ teeth_neighbor. If another route reaches an already assigned gear with a contradictory direction or speed, the arrangement has a conflict. Success requires every placed gear and the target to be connected with no such contradiction.
Along a simple train, external meshes alternate clockwise and counterclockwise, so the number of contacts determines whether the final direction matches or opposes the driver. An idler changes that parity, but its tooth count appears once in a numerator and once in a denominator, causing it to cancel from the end-to-end speed ratio. The final speed magnitude depends on the driver and target tooth counts. Read why an idler changes rotation direction without changing the simple train's final speed ratio for the underlying relationship.
Contact parity also explains why loops can fail. An odd closed loop asks one gear to rotate in both directions, producing a conflict, while an even loop can remain consistent. The target label reports clockwise or counterclockwise rotation, and the side ledger reports the ideal speed magnitude. Reduced-motion users receive the same direction and ratio information without relying on animation.
Turn failed feedback into a repair
Submit only a complete placement. The result identifies the specific structural problem that needs attention:
| Test result | What it means | Next move |
|---|---|---|
| Partially filled board | No complete assignment is available to test. | Fill every socket with one distinct candidate. |
| Target disconnected | No powered path reaches the target. | Rearrange the train to create a continuous route. |
| Gear isolated | A placed gear is outside the driver's powered component. | Move or bridge the isolated gear. |
| Direction conflict | The contact graph demands contradictory direction or speed. | Change the contact arrangement, especially around a loop. |
| Valid train | Every placed gear and the target are connected without conflict. | Receive 200 points and open the next fixed gearbox. |
One distinct failed full layout can be repaired. Submitting the identical placement assignment again is deduplicated and does not consume the remaining recovery chance. A second different failed assignment closes the run, so use the structural feedback before trying another complete arrangement.
Control the entire construction loop
Pointer, touch, and keyboard input all support construction. Keyboard play covers selecting candidates, placing or clearing gears, testing, and restarting.
| Input | Action |
|---|---|
| Pointer or touch | Select a rack candidate, choose a socket, and tap an occupied socket to return its gear when needed. |
| 1 through 4 | Hold an available candidate gear. |
| Q, W, E, and T | Address sockets from left to right. With a candidate held, place it; with none held, clear an occupied socket. |
| Enter | Test a completely filled train. |
| R | Restart the game. |
Every pointer target is at least 44 pixels, and the interface provides visible focus and held states so the current selection can be identified during construction.
Follow the fixed 1,000-point route
Each valid train awards exactly 200 points and opens the next of five fixed levels. Completing all five produces exactly 1,000 points. The final completed board remains visible and rotating while interactions are frozen, so the finished result is preserved rather than replaced by an empty screen. There is no time bonus, random level, adaptive difficulty, hidden multiplier, or score penalty for exploring placements. Restart returns to level one with an empty board, zero score, no failed signatures, and no held candidate.
Keep the engineering model in perspective
The game deliberately models ideal same-module external spur gears for educational play. It is not engineering design software and must not be used to size a real transmission. The model omits torque, power loss, tooth stress, backlash, profile interference, shaft alignment, lubrication, noise, wear, materials, safety factors, real manufacturing tolerances, and dynamic loads. Its educational scope appears beside the board and in the methodology rather than being hidden from the player.
For general mechanical background, MIT OpenCourseWare's Gears and Linkages lecture and Tec-science's explanation of gear stages provide useful context. The puzzle's disclosed ideal model nevertheless governs every contact, direction, ratio, and scoring decision inside the game.