Does gear reduction make a Technic-style slider more accurate?

Gear reduction gives a slider smaller commanded movements for the same input turn, but it does not guarantee better positioning accuracy. Our new positioner combines two gear reductions with a captured rack. One quarter input turn commands about 0.70 millimetres of ideal travel. Real tooth clearance, shaft movement, bending and the way you approach a target can still change where the slider stops.

Build a machine with a useful output

The Two-Stage Rack Positioner is a level-four, hand-operated machine with seventy-six parts and twenty-eight assembly steps. It has three interacting sections: a compound reduction drive, a shared output shaft that turns a rack pinion, and a captured channel that supports and guides the sliding carrier. The purpose is to approach a small position target and investigate reversal, rather than merely watch unrelated gears turn.

Trace the drive from the red input handle. Its eight-tooth gear turns a twenty-four-tooth gear on the blue middle shaft. That middle shaft also carries an eight-tooth gear on a different height layer. This second small gear turns another twenty-four-tooth gear on the yellow output shaft. The output shaft’s lower eight-tooth pinion engages the rack. Each layer has its own job; the smaller gears on shared cross axles cannot spin independently of the larger gears beside them.

Tools & Machines Lab’s independent Compound gear train 125:1 model with Lego Technic inspired the serial-stage principle. Its inspected demonstration shows a hand-driven compound train, and its description explains three five-to-one stages. Our original positioner uses two three-to-one stages and adds a supported linear output. Its layout, parts, dimensions and instructions are different. We do not claim to reproduce the creator’s model, copy its assembly, or measure its performance.

Follow turns all the way into millimetres

The first reduction gives one reverse middle turn for three input turns. The second gives one output turn for three middle turns. Multiply the reductions: three times three is nine. Two external gear contacts reverse direction twice, so the input and output shafts turn in the same sense when viewed consistently from above. The rack then moves sideways according to which side of its pinion it touches.

An eight-tooth pinion has a four-millimetre pitch radius in this model. One pinion turn commands a pitch-circle distance of about 25.13 millimetres. Divide that by nine: one input turn commands about 2.793 millimetres. Divide again by four for a quarter input turn: about 0.698 millimetres. These are ideal geometric calculations, not readings from a ruler. The short rack channel permits only eight millimetres either side of centre, so even one complete pinion turn would exceed its reviewed stroke.

Predict first, then use four positive quarter turns to approach the 2.79-millimetre model target. Read the input, middle and pinion turns separately. A single three-to-one stage would produce a different answer; the point is to trace the whole connected drive. Compare the Compound Counting Bench if you want to isolate the multiplication before building this integrated machine.

Finer movement is not the same as accuracy

A directly driven pinion would command about 6.283 millimetres per quarter turn. This positioner commands one ninth of that distance for the same hand-angle increment. That is finer commanded movement. Accuracy asks how closely the real carrier reaches a reference position. Repeatability asks how tightly several attempts group together. Neither follows automatically from the reduction ratio.

The optional take-up setting in the lab illustrates one reason. Select a forty-five-degree input-equivalent reversal gap, which resets the simulated mechanism preloaded for positive input. Move forward one input turn, then reverse a quarter turn. Some of that reverse command is consumed before the model output responds. The ideal rack position is about 2.094 millimetres; the illustrative take-up model is about 2.443 millimetres. That discrepancy comes from a chosen simplified deadband, not a measurement of our real gears.

The real machine has several contacts and flexible parts, so its play need not behave like this aggregate model. The tool does not say which mesh contains the clearance, predict force, or establish a tolerance. Returning the input marker to its old angle can leave a different output position after reversal. Approach direction belongs in the experiment record.

Test a target fairly

Complete the free positioner instructions, check every cap and hold the stationary frame level. Mark the carrier’s centre. Place a ruler or paper scale beside it, clear of the moving parts, and use the same viewpoint for each reading. Turn only the red input slowly. Keep travel within the reviewed eight-millimetre limit either side; the model has no automatic end stops.

Choose a small target such as about 2.8 millimetres left of centre. First approach it from the same direction on three trials. Return to your starting reference between trials, taking up any observed play consistently. Then repeat after a deliberate reversal. Keep the build, reference, input increment and viewing position fixed.

Record the spread and reading uncertainty as well as the average. A scale too coarse for a sub-millimetre change cannot establish that change accurately. If you cannot resolve a small movement, record it as unresolved rather than substituting the calculator’s value. We have not physically commissioned the model or measured its backlash, accuracy, efficiency or load capacity.

Support is part of the transmission

Each moving shaft has a lower frame or crossbeam bearing and an upper bridge bearing, with retained keyed hardware. Two stationary poles support the upper bridge. The rack’s lower crosspieces carry it, side rails prevent sideways escape, and retained top crosspieces prevent lifting. A gear mesh transmits force; it cannot replace those guides. The simpler Rack Travel Bench explains that capture on its own.

If turning the handle lifts a shaft or shifts the bridge, fix the stops and supports before interpreting a target miss as backlash. If one gear spins without its shaft partner, check the correct cross axle and gear holes. If teeth bind, check the separate height layers, centre distances and whole-shaft phase; do not force the input. Review compound gears, rack and pinion, bushes and friction.

A successful experiment explains the complete motion path and distinguishes commanded travel, observed position and uncertainty. More reduction does not create free energy, guarantee a brake or justify a lifting load. Keep this model unloaded, stop before the rack tabs, and use the target/reversal comparison to decide what the evidence actually supports.