Gantry robots: motion above the wet deck

A gantry’s job looks trivial: pick a microplate up, put it down somewhere else, hundreds of times a day. What makes the job possible at all is not the robot but the object. The microplate is the shipping container of the wet lab — a rigid rectangle with a standardised footprint and registered corners, agreed at roughly 128 by 86 millimetres by the Society for Biomolecular Sciences — and every instrument on the line accepts it in the same orientation. Integration is an interface problem before it is a motion problem.

A container standard, not a robot problem

Because every plate has the same footprint, one gripper serves the entire line, and because every instrument presents the same docking geometry, positioning demands are millimetre-class, not micron-class. Nests have chamfered openings that funnel a plate arriving a few millimetres off-target into place — passive compliance doing the work active servo control otherwise would. The plate is also mechanically forgiving in one direction and not in others: it tolerates small tilt, but a torsional mismatch between gripper and nest binds the plastic and skews wells. This is why commissioning an automated line is dominated by teaching station coordinates and verifying grip heights, not by tuning motion — the kinematics were never the hard part.

The motors are never the limit

A full 96-well plate with 200 µL per well plus its plastic shell weighs tens of grams. Any laboratory motor moves it at any speed the designer wants. What objects is the liquid. Accelerate an open vessel sideways and its free surface slopes, then oscillates: sloshing. In a single small well the first slosh mode is fast and hard to excite, but in a reservoir or reagent trough the free surface is wide, the natural slosh frequency low, and a brisk acceleration turns the contents into a standing wave that can slop over into neighbouring wells. So accelerations — not peak speeds — are engineered conservatively, and moves are smoothed with jerk-limited profiles. Then there is settle time: a plate set down while still ringing on a compliant gripper will not sit square in its nest. Cycle time in practice is dominated by acceleration ramps, grip and release moments, and the wait for oscillation to die — which is why swapping in faster motors changes almost nothing, and why transport of open, liquid-filled labware is deliberately slower than the hardware could go.

Why the robot lives overhead

The deck of a workcell is the scarce resource: every square centimetre occupied by a robot arm is a square centimetre not holding an incubator or a reader. An overhead rail on gantry axes spans the line without touching the deck, reaching every station from above; where a full rail is too much, an arm on a stand serves a cluster of instruments. The transport layer also inherits the liquid handler’s physics on timing: a plate waiting on a transfer station is evaporating from its edge wells, and a move that splashes cross-contaminates the assay downstream of any robot fault. Motion above the deck is therefore the visible half of a bargain whose other half is scheduling — deciding when each plate should move, which the orchestration layer owns.