Teaching Your Deck: Why a Correct Definition Is Not the Same as a Correct Position

The gap between software and hardware

The previous article established that the worktable is the software model of your deck. A perfect worktable, with the exact right carrier definitions, correct labware geometry, and properly ordered sequences, is necessary but not sufficient.

Even a perfect software model doesn't account for physical reality. Your STAR deck has tolerances. Carriers sit in tracks with small positional variation. The same carrier type placed in track 1 on one instrument may sit 0.3 mm differently than the same carrier on another instrument. Labware placed in carriers has its own manufacturing tolerances. Over time, carriers can shift.

Teaching is the process of verifying and correcting the robot's understanding of where things actually are on your specific deck. It's the calibration step that closes the gap between the worktable model and the physical instrument.

Skipping teaching, or treating it as optional, is one of the most common causes of tip crashes, missed aspirations, and systematic pipetting errors in Hamilton methods.

What teaching actually corrects

Teaching addresses three layers of positional uncertainty.

1. Carrier position on the deck

Carriers sit in numbered deck tracks, but the exact lateral position and the Z-height at which the robot approaches that carrier can vary. The Autoload system uses a barcode reader to confirm carrier identity, but it doesn't verify the Z-position of every labware slot on that carrier. A carrier sitting 0.5 mm higher than the definition expects means every labware position on it is 0.5 mm higher than the robot thinks, and every aspirate/dispense Z-position will be 0.5 mm off.

For a trough, 0.5 mm of Z-offset means the robot stops aspirating 0.5 mm further from the true bottom than intended, leaving extra dead volume. For a 384-well plate, 0.5 mm represents a substantial fraction of the total well depth and can cause tip-to-plate contact.

2. Labware Z-bottom

Even with a perfectly seated carrier, the labware inside it has its own geometry. The Z-bottom, the Z-coordinate of the lowest point the robot will send its tip, is defined in the labware definition. Teaching verifies that this value matches reality for your specific combination of carrier and labware on your specific instrument.

In Venus, the Calibrate Carrier step can correct carrier-level position offsets at runtime. For labware-level Z-bottom corrections, the teach process involves adjusting the labware definition directly in the Deck Layout Editor after physically verifying the tip position.

3. Liquid class positions within labware

Teaching also validates the aspirate and dispense positions defined by your liquid class, specifically the fixed height from bottom that is used when LLD is disabled. A liquid class specifying "aspirate 1 mm from well bottom" will only reach the correct depth if the labware definition's Z-bottom is accurate and the carrier position is correct. Both layers must be right.

How teaching works in practice

The teach process for a carrier

Teaching a carrier is a hands-on process done at the instrument. The general workflow:

1. Load the carrier physically onto the correct deck track.
2. Open Venus and navigate to the System Configuration Editor or the relevant teach function.
3. Use the manual jog controls to move the pipetting arm to a reference position on that carrier; typically the first position of the first labware slot.
4. Lower the arm carefully until the tip just reaches the reference point. For a plate carrier this is typically the centre of the well at the first position; for a trough it's the trough bottom centreline.
5. Record the XYZ coordinates at this position.
6. Compare these coordinates to what the software expects from the labware definition.
7. If there's a discrepancy, correct the carrier or labware definition to match reality.

The precision required depends on the labware format. For a standard 96-well plate, a teaching tolerance of ±0.3 mm is typically acceptable. For a 384-well plate (4.5 mm well pitch), the tolerance is tighter, ±0.2 mm or less, because the narrower well pitch means a position error can cause tip-to-wall contact or mis-positioning into an adjacent well.

Using Calibrate Carrier in a method

Venus includes a Calibrate Carrier Single Step that can be included at the start of a method to verify carrier positions at runtime. This step moves the arm to a reference position on the specified carrier and confirms the position is within tolerance. If the carrier has shifted or is missing, the step generates an error before the method proceeds to pipetting.

Including Calibrate Carrier for critical carriers, particularly for 384-well plates, deep-well plates being accessed at low volumes, or troughs at the end of a large reagent run, is good practice. It adds seconds to the start of a run but prevents far more costly errors downstream.

The Initialize step

Every Venus method should start with the Initialize Single Step. This step homes all axes and confirms the robot is in a known mechanical state. It's not teaching; it doesn't verify positions against labware. But it's the prerequisite for reliable position control throughout the method.

Setting Initialize mode to 1 (Always initialize) ensures the robot homes at the start of every run, not just the first run after power-on. This prevents accumulated mechanical drift from affecting position accuracy over multiple runs.

What happens when teaching is wrong or absent

Tip crashes

The most visible consequence of incorrect teaching is a tip crash: the robot drives a tip into the labware, carrier, or deck surface. This can bend or break tips, damage the pipetting arm, crack labware, and in severe cases damage the deck hardware. Tip crashes most commonly occur when:

• The Z-bottom is set lower in the software than the physical bottom of the labware
• A carrier is placed on the wrong track and the robot navigates to a position that's physically occupied by a different carrier
• Labware is taller than the definition expects, so the robot navigates at a height that intersects with the labware wall during traverse

Silent pipetting errors

More common and more dangerous than crashes are silent pipetting errors, situations where the robot completes the method without hardware errors but the pipetting is systematically wrong.

If the Z-bottom is set 1 mm above the physical bottom of a well, the robot aspirates from 1 mm higher than intended. In a deep-well plate this may have no practical effect. In a 96-well standard plate where the working depth is 10 mm, 1 mm represents 10% of the available depth, and at low volumes the robot may be aspirating air.

If carrier X/Y positions are off, multi-channel aspirations may be addressing wells that aren't centred under the tips. For a 1000 µL tip in a wide 96-well plate, this may be acceptable. For a 50 µL tip in a narrow 384-well plate, off-centre positioning causes tip-to-wall contact, meniscus disruption, and volume errors.

Accumulated errors across a plate

Teaching errors small enough to be invisible in a single run can accumulate into visible failures across a plate. An aspirate position that's 0.2 mm off-centre has minimal effect in well A1 of a 96-well plate. But if the carrier position error applies uniformly across all 12 columns, the error at column 12 may be 0.2 mm plus whatever manufacturing tolerance exists at the far end of the carrier, potentially 0.5 to 0.8 mm from column 1 to column 12.

This is why teaching should be done for each carrier, not just the first one on the deck.

Teaching best practices

Teach every new carrier, not just new carrier types. Two physically identical carriers may sit differently on the same deck due to track wear or carrier-to-track clearance variation. Teach each carrier individually when it's placed on the instrument for the first time.

Re-teach after any deck modification. Moving a carrier to a different track, replacing a worn carrier, or adding a new active device requires re-teaching the affected positions. Don't assume the previous teach values carry over.

Document your teach positions. Record the XYZ coordinates confirmed during teaching in a method development log. If the labware definition is updated or the carrier is replaced, you have a reference to compare against.

Verify teaching empirically for critical aspirate positions. After teaching a carrier, run a test dispense from the relevant labware into a waste plate or tube; a dye dispense works well. Visual inspection of the dispense pattern confirms the robot is addressing the correct positions. For troughs, a test aspiration at the taught Z-bottom followed by inspection of the tip (liquid present or dry?) confirms the Z-bottom teaching is correct.

For 384-well and smaller formats, test the extremes. Teach verification at position A1 of a plate confirms the near-end position. Also verify position P24, the far end of a 384-well plate. Carrier-level X and Y offsets accumulate across the full plate length, and an error that's invisible at A1 may be significant at P24.

Teaching as part of method development, not an afterthought

The most reliable approach is to treat teaching as a formal step in method development, with its own checklist and sign-off, before the method goes into routine use.

A typical teaching workflow for a new method:

1. Finalise the worktable with all carriers and labware defined.
2. Load the physical carriers and labware onto the instrument.
3. Run the Initialize step.
4. Teach each carrier position, comparing software-expected coordinates with physically measured coordinates.
5. Correct any discrepancies in the labware definition or carrier placement.
6. Run a dry test (no liquid, no tips) to visually confirm that all arm movements navigate to the expected positions without contact.
7. Run a wet test with dye to confirm aspirate and dispense positions produce correct results.
8. Document the final teach positions and the labware definition versions used.

This process takes time up front. It saves run failures, reagent waste, and hardware repair costs downstream.

With a verified worktable and confirmed teaching, the next layer is the logic that drives the robot through your labware: sequences.