Compact motion for sample transport
A sample shuttle runs at low duty and low torque for years. The jam case, not the running case, sets the drive, and brush life is set by hours the mechanism barely accumulates.
- Wedge
- Medical and laboratory equipment
- Catalog reviewed
- 2026-08-18
- Configurations screened
- 32 published
A sample transport moves a rack, a carrier, a slide or a tube between stations: a load port to a centrifuge, a decapper to an aliquotter, an input drawer to a track. The drive is usually a belt or a rack and pinion carrying a horizontal mass over a fixed stroke, or a carousel indexing a fixed angle. There is no process force. The load is friction, acceleration, and whatever a jammed or mis-seated sample does when the carrier meets it.
Same reflected-inertia model as any horizontal indexing axis. J = m·r² for the translating mass at a belt pitch radius r, plus pulley and belt inertia; T = J·α + µ·m·g·r. Two things distinguish transport from pipetting indexing. The moves are long and the times are seconds rather than tenths, so the acceleration term is small and the friction term is comparable to it or larger. And the friction coefficient is not a constant: guides in a sample-handling instrument collect spilled specimen, dried reagent and label adhesive, so the value that matters is the end-of-service-interval value, not the clean one.
What normally controls the selection
The jam. Everything else has enormous margin. A transport axis runs at a few percent of the continuous rating, at a small fraction of the rated speed, for a duty cycle measured in single-digit percent. What actually removes a candidate is that when a rack fouls, the drive can push hard enough to crush a tube, damage a carrier, or break the output stage, because the stall torque delivered through the ratio exceeds both what the mechanism should apply and what the gearhead is rated to carry. After that, the second consideration is total running life over a ten-year instrument program, where the relevant quantity for a brushed motor is running hours and for a brushless motor it is bearing and grease life.
Four conditions, and they are usually decided by four different numbers.
Sizing on one of them is how a mechanism gets a part that works on the bench and fails in the field.
Failure modes and misleading specifications
What follows is what actually fails, and which published number sends people the wrong way.
| Failure mode | Physical cause | What it looks like in the field |
|---|---|---|
| Crushed tubes or damaged carriers | The drive can develop far more force at stall than the transfer needs. Nothing in the mechanical path limits it. | Damage concentrated at one station, always in the same direction of travel, correlated with a specific rack or carrier tolerance. |
| Output stage failure with no wear | Repeated stall events at full command. Each one loads the output stage above its rating. | A tooth root crack or a spun output shaft on a unit whose gears are otherwise unworn and whose brushes have barely started. |
| Failure to start after a service interval | Static friction risen with contamination, on a drive sized against the clean running friction. | Intermittent no-motion faults that clear when the guide is cleaned. Current trace shows the drive drawing stall current without moving. |
| Position loss on a belt drive | Belt tooth jump during the stall or during recovery from it, on an open-loop axis referenced only to a home sensor. | Cumulative position error that resets at every homing operation, growing with the number of jam events. |
| Specification | Why it misleads in this application |
|---|---|
| Continuous output torque | The axis uses a few percent of it. Selecting on this number tells you nothing, and it hides the fact that the same configuration can develop many times that torque at stall. |
| Rated life in hours | A transport axis running 40 transfers an hour at 1.5 s each accumulates one minute of running per hour. Ten years of that is under 1 500 hours. Hours are rarely the constraint. Starts, reversals, and grease age usually are. |
| Efficiency | At a few percent of rating, efficiency changes the current by microamps. It matters for the thermal calculation on a hard-working drive and not at all here. |
| Duty cycle rating expressed as a percentage | A 25 percent duty rating implies the limit is thermal. On this axis nothing is thermal. The limit is a single-event torque limit that no duty percentage describes. |
The tradeoffs that matter here
Spur or planetary reduction
Brushed coreless or brushless
How the jam is limited
Size a belt-driven rack shuttle, then check what it does to a tube when the rack fouls.
- Transfer stroke
- 300 mm
- Transfer time
- 1.5 s
- Moving mass
- 0.8 kg, rack plus carrier
- Pulley pitch radius
- 15 mm
- Transfer rate
- one every 10 s
| Step | Expression | Result |
|---|---|---|
| Peak carriage velocity | v = 0.300 m ÷ (1.5 s × 2/3) | 0.300 m/s |
| Acceleration | a = 0.300 m/s ÷ 0.5 s | 0.600 m/s² |
| Peak drive speed | n = 0.300 m/s ÷ 0.015 m = 20 rad/s | 191 rpm |
| Total drive-shaft inertia | 0.8 kg × (0.015 m)² + 0.4 × 10⁻⁴ | 2.20 × 10⁻⁴ kg·m² |
| Acceleration torque | 2.20 × 10⁻⁴ × (0.600 ÷ 0.015) | 8.80 mN·m |
| Friction torque | 0.08 × 0.8 kg × 9.81 m/s² × 0.015 m | 9.42 mN·m |
| Peak torque | 8.80 + 9.42 | 18.2 mN·m |
| RMS torque over the 10 s period | √[(18.2² + 9.42² + 0.62²) × 0.5 ÷ 10] | 4.59 mN·m, 7.9 percent of the 58.1 mN·m continuous rating |
| Output torque at a hard stall | 11.8 mN·m × 32.5 × 0.73 | 280 mN·m, 1.43 times the 196 mN·m gearhead rating |
| Belt force at a hard stall | 0.280 N·m ÷ 0.015 m | 18.7 N against whatever the carrier hit |
Result
MM-B2419-S00325-240A resolves the running case with very large margin: brushless, 32.5:1 spur, 24 V, rated 212.3 rpm against the 191 rpm demand and 58.1 mN·m continuous against a 4.59 mN·m RMS demand. Brushless is the right class here not because of torque but because a transport axis in a ten-year instrument accumulates real hours on a high-throughput track and generates no brush debris in a sample-handling enclosure. The jam case does not resolve on its own. At 24 V with no limit the drive develops 280 mN·m, which is 1.43 times the gearhead rating and pushes 18.7 N into the obstruction. A current limit set to bound the output at, say, 40 mN·m holds the belt force to 2.7 N and keeps the gearhead at a fifth of its rating. That limit is part of the mechanism design.
What would change it
The pulley radius is the strongest lever on the jam force: force at the belt is inversely proportional to radius, so a smaller pulley makes a jam worse for the same torque limit while also raising the required speed. Contamination is the biggest uncertainty in the running case, and doubling the friction coefficient doubles the friction torque without touching the acceleration term, which the 58.1 mN·m continuous rating absorbs without complaint. If the deck cannot accept a 37 mm gearhead envelope, the brushless option disappears and the choice becomes a coreless brushed planetary with a brush-life argument attached.
A short qualification plan
Each step names the measurement, not the intention.
-
Measure friction at end of service interval
Breakaway and running force on a guide that has been through the contamination the instrument will actually see, not on a clean sample.
-
Set and verify the torque limit
Measured belt force at the commanded limit, at both temperature extremes, on at least five drives. The torque constant varies with magnet temperature and so does the limit.
-
Run the jam case deliberately
Repeated stalls against an instrumented obstruction, then teardown. Confirm no cumulative damage to the output stage.
-
Accumulate transfers, not hours
Life test counted in transfers at the real duty and dwell, with position accuracy checked against the home reference throughout.
-
Prove homing recovery
Position integrity after a jam and recovery sequence, including the case where the belt has jumped a tooth.
Tools, products and articles for this problem
Send the transport axis for an application review.
Send the stroke, transfer time, moving mass, drive geometry and the damage threshold at the sample. You get back the running case, the jam case, and the torque limit that separates them.