Comparing manual and automated handling

A liquid handler is a research tool for repeating a volume after a person has decided the volume is the right one. Comparing manual and automated handling means listing the variances each one removes and the new ones the deck adds, then deciding which check proves the run you are about to trust. Manual forward technique remains the reference habit in accurate micropipetting technique. The robot does not make that habit obsolete. It freezes one version of it in a file.

What the application can claim

Use automation when the same map repeats, when fatigue has been the scatter, and when the liquid is within what the head and the tip can move. The claim you can support is that programmed wells received the programmed action more uniformly than a tired hand. The claim you cannot support, without a check, is that the programmed volume matches the specification, that the plate was in the right orientation, or that a viscous reagent behaved like the water used to teach the head.

A person sees a bubble, a drip, and an empty reservoir corner. A robot sees thresholds if it has sensors, and otherwise it sees nothing. Clot detection and liquid-level detection are optional classes of hardware, not a moral property of automation. If your head lacks them, your method needs a visual check of the first cycle and a fill height with margin. If your head has them, test whether a viscous liquid trips the clot alarm. The branch is to slow the aspirate in the programme or to keep that liquid manual.

Tip-touch, dead volume, and deck layout

Tip-touch is the automated cousin of touching the well wall. It can finish a forward dispense. It can also pellet a bead of liquid on the wall above the reagents if the height is wrong. Verify touch with dyed water and the real plate type. A different manufacturer of 96-well plates can change the well depth enough to turn a good touch into a crash or a miss.

Dead volume sits in reservoirs, troughs, and tubes the tip is forbidden to reach. Specify it for each labware definition. Manual users often tilt a tube. Robots usually cannot. The last programmed aliquots are where air aspiration shows up as scatter. Leave a remainder on purpose, the way reverse pipetting leaves a remainder, and do not plan experiments that consume the theoretical bottom of the bottle.

Deck layout errors are systematic and confident. A plate rotated so A1 is where the programme thinks H12 is will be pipetted perfectly into the wrong wells. A tip rack of the wrong height teaches the head to pick air or to crush cones. A dropped tip in a nest shifts every later coordinate. The application limit is that the map in the software and the plastic on the deck are the same object. Photograph or checklist the deck at the start of a run you will publish.

Tip fit still rules. The head was checked with a tip family. A bargain rack that almost mounts will leak on every channel and the leak will look like biology because it is so even. Filter tips change the air path slightly and may be required by the risk assessment. They are not a biosafety cabinet.

A workflow for adopting a method, and the branch when the robot and the hand disagree

Write the manual method first: volume, forward or reverse, pre-wet, tip family, temperature. Translate it into speeds and heights. Run dyed water or a gravimetric series at the programmed volume. Convert mass with water density at the measured temperature, about 0.998 grams per millilitre near room temperature, from a table. Compare the mean and the scatter with the specification you claim. Error numbers live in the manufacturer specification and, for classical pipettes, in ISO 8655-2:2022. ISO 8655-1:2022 remains a terminology and user-recommendation standard, not a robot certificate.

Branch if the robot is tighter than the hand but the mean is shifted. Look at tip-touch height, blow-out settings, and pre-wet before you edit the volume number to force agreement. Branch if the robot is wider than a rested hand. Look at tip seating, teaching, and a reservoir low enough to catch air. Branch if a power cut stops a run mid-plate. Do not restart blindly. Record the last completed well from the log, check for half-filled wells, and decide whether the plate is still interpretable. The head does not lose its calibration because the power failed, and the plate may still be ruined.

A manual check of the same map, by a person using checking a pipette between calibrations logic on the hand pipette, tells you whether the disagreement is the head or the chemistry. Keep both instruments inside their own specifications.

Safety and the limit of automation

A robot can spill a larger reservoir than a hand. The risk assessment names containment, tips, and waste. A filter tip on a head is still not a cabinet. This comparison is research method design. It is not a diagnostic claim and not a statement that unattended operation is appropriate for every liquid. Stay inside what your institution allows a deck to do alone.

What an enquiry should include

State volumes, channel count, tip family, whether you need level or clot sensing, the labware heights, and the calibration document required for the head. Say which liquids are viscous or volatile so dead volume and speed are not copied from water. Instruments are in the scientific instruments catalogue, tips and reservoirs in the laboratory consumables catalogue. Send the deck constraints with the quotation request.