Zxmoto Ping Pong Robot Machine Parts: How They Work

• 1
  Key takeaways
• 2
  What a ping pong robot actually is
• 3
  Launch wheels, hubs, and the bore that sets ball speed
• 4
  Ball feed, separation, and why roller gaps matter
• 5
  Oscillation, pivot housings, and backlash you can feel
• 6
  Material choices and surface treatment for robot parts
• 7
  When CNC machining is the wrong process
• 8
  Part families, critical features, and process fit

Strip away the housing and a table tennis robot is a feed mechanism, a launch mechanism, and one or two positioning axes. Balls drop from a hopper into a channel, get separated one at a time, and pass between two spinning wheels. The wheels grip the ball and throw it. A separate motor swings the head left and right so the ball lands in different spots.

Every one of those functions depends on a small group of machined or formed parts. The launch wheels sit on hubs that must run true. The feed roller turns on a shaft that must not wobble, or two balls arrive at once. The swing arm pivots in a bearing housing where the bore diameter controls backlash. None of this is exotic, but the tolerances are tighter than they look.

This is why Zxmoto ping pong robot machine parts are usually specified as machined components rather than stamped or molded ones. A stamped hub with a 0.1 mm bore variation will produce a visible speed spread between two identical units. A machined hub held to ±0.005 mm will not.

The rest of this page walks through the five part families that decide how a robot performs, then explains the fits and the materials that make them work. Skip to the table if you already know the mechanism and only need the selection criteria.

The launch wheels are the parts most people notice, but the machined hub behind them does the real work. Each hub carries a wheel on one end and a motor coupling on the other. If the wheel seat and the coupling bore are not concentric, the wheel runs out and the ball leaves with a wobble that no amount of speed tuning will fix.

A practical concentricity target for the wheel seat relative to the motor shaft is 0.02 mm total runout. That sounds loose next to a ±0.005 mm machining tolerance, but the point is that the machining tolerance makes the assembly target achievable without hand fitting. If the hub comes in at 0.05 mm runout, the robot will throw a ball that curves differently on every shot.

Wheel speed is the other half of the problem. Two wheels at 4,000 rpm with a 1% speed difference produce a spin difference large enough to change where the ball lands. That difference usually comes from friction in the hub bearing seat, not from the motor driver. A bore that is 0.03 mm undersized will pinch the bearing and add drag.

For spin-heavy shots, the two wheels run at different speeds. The slower wheel grips and drags, which is what generates topspin or backspin. That makes bearing seat diameter and surface finish more important than raw wheel power. Ra 0.8–1.6 μm in the bearing seat is a reasonable target for a press fit.

• 1
  Wheel seat runoutKeep total runout under 0.02 mm or the ball flight becomes inconsistent.
• 2
  Bearing seat fitA press fit held to ±0.005 mm avoids both pinch drag and looseness.
• 3
  Coupling boreMatch the motor shaft with a light press or clamped split hub, never a loose slip fit.

The feed section looks simple: a hopper, a channel, and a rotating roller that lets one ball through at a time. In practice this is where most field failures start. A ball is 40 mm in diameter and weighs about 2.7 g. It deforms slightly under roller pressure, which means the gap that works for a new ball may jam on a worn one.

The separation roller gap is normally set between 38 mm and 42 mm depending on ball hardness. Too tight and the roller crushes or stalls. Too loose and two balls pass together, which the launch wheels then throw as a double hit. The machined parts involved are the roller shaft, the roller end caps, and the side plates that hold the shaft parallel.

Parallelism between the two side plate bores matters more than their absolute size. If the shaft sits 0.1 mm out of parallel over its length, the gap tapers and balls jam on one side of the channel. Boring both plates in one setup, or pinning them together during machining, removes that error before it reaches assembly.

Hopper and channel parts are usually not machined. Sheet metal fabrication handles the hopper walls, and clear polycarbonate or acrylic covers let users see jams. Save the machining budget for the shaft, the roller, and the side plates.

• 1
  Roller shaftTurned from 303 or 304 stainless; 316L if the robot sees humid club air.
• 2
  Side platesBore as a pair so the two holes stay parallel within 0.02 mm.
• 3
  Channel wallsSheet metal is enough; keep the machined tolerance for the roller path.

The sweep axis moves the launch head across the table. It is a low-speed, high-torque motion, so it does not need a fast motor. It does need a stiff pivot with almost no backlash. Any play in the pivot shows up as a delayed or overshooting move at the table edges, where the head has the most mass to swing.

The pivot housing is a machined block with a bored bearing seat, a mounting face, and often a hard stop. Bore roundness and face squareness are the two features to inspect. A bore that is round but not square to the mounting face will tilt the head, and the tilt changes the ball's landing zone as the head sweeps.

Belt or gear drive on the sweep axis needs a tensioner or an idler that can be adjusted in small steps. A slotted bracket is the usual answer: a machined slot with a sliding block lets a technician set tension without shims. Slot width should match the block within 0.05 mm to avoid the block rocking under load.

If the robot uses a linear rail for vertical or depth adjustment, the rail mounting face flatness matters. A face that is bowed by 0.1 mm will preload the rail unevenly and cause sticking near the ends of travel. Face milling in one pass usually holds flatness well enough.

• 1
  Pivot boreHold roundness and squareness to the mounting face within 0.02 mm.
• 2
  Tension slotKeep block-to-slot clearance near 0.05 mm to stop rocking.
• 3
  Rail faceFace mill in one setup; uneven preload causes end-of-travel sticking.

Aluminium 6061-T6 covers most brackets, housings, and covers. It machines fast, holds ±0.005 mm on critical features, and keeps the sweep axis light. Use 7075 when a bracket needs higher stiffness at the same thickness, for example a motor mount that flexes under belt tension.

Stainless 303 or 304 works for shafts and rollers that see constant ball contact. 303 machines more freely and is fine for a feed roller. 316L is the safer choice in humid training halls or where the robot is stored in a cold room, since it resists pitting better. 17-4PH is worth the cost only for a wear-prone shaft that also needs corrosion resistance.

Plastics appear in ball contact areas. POM and PA make good low-friction feed rollers and guides because they absorb ball impact and do not scratch the ball surface. PEEK is an option where heat or chemical cleaning is involved, but it costs far more than the job usually justifies.

Finishing depends on the exposure. Anodizing in clear or color gives aluminium housings a durable skin and hides tool marks. Hardcoat anodizing raises surface hardness on wear faces such as roller end caps. Bead blasting before anodizing produces a uniform matte look that hides small machining marks on visible covers.

• 1
  Brackets and housings6061-T6; step up to 7075 only where stiffness is the limiting factor.
• 2
  Shafts and rollers303 or 304 stainless; 316L for humid or cold-storage use.
• 3
  Ball contact partsPOM or PA for low friction and impact absorption.
• 4
  Wear surfacesHardcoat anodizing or electroless nickel on aluminium wear faces.

Not every part on a robot justifies a machined billet. Large flat covers, hopper walls, and simple L-brackets are usually cheaper as sheet metal. If a part has no tight bore, no critical face, and no wear surface, laser cutting and bending will do the job at a lower unit cost once quantity grows.

Injection molding wins when a plastic part runs in the thousands. A feed guide or a ball channel insert can be machined for prototypes and then molded for production. The machined version proves the geometry; the molded version proves the price. Doing both is normal and does not require a redesign if the wall thickness is kept uniform.

Casting suits housings with complex internal ribs where machining from solid would remove most of the material. A die cast or vacuum cast housing can carry the shape, then have only the bearing bores and mounting faces machined. That keeps the tight tolerances where they matter and leaves the cosmetic shape to the mold.

The honest rule: machine the features that touch a bearing, a ball, or a motor shaft. Form or mold everything else. That split usually removes most of the cost without touching performance.

• 1
  MachineBearing bores, motor mounts, roller shafts, pivot housings.
• 2
  Sheet metalHopper walls, covers, simple brackets, mounting plates.
• 3
  Mold or castHigh-volume plastic guides and ribbed housings with machined bore inserts.