Something is wrong with a transfer. There are droplets clinging to the tips, or the volumes come back low, or every well reads a little different from the last. The temptation at this point is to open the liquid class and start turning knobs until the problem goes away. It usually does go away eventually, and you are left with a class you cannot reproduce and no idea which change actually helped. Troubleshooting goes faster, and teaches you more, when you run it the other way around: start from the symptom, reason back to the liquid property behind it, and reach for the one parameter that property controls.
What follows is a symptom-first map. It assumes you have already met the parameters a class exposes and the liquid properties that drive them; here we connect the two in the direction you actually need them when a run misbehaves. The order matters as much as the map: look before you measure, rule out the deck before you blame the class, and change one thing at a time so the fix is something you can explain.
Look before you measure
The cheapest diagnostic you have is your own eyes, and most problems announce themselves before any balance or plate reader is involved. Run a single transfer that mimics one step of your real method and watch it happen. A short visual inspection catches the majority of faults while they are still free to fix.
- Droplets on the tips after aspiration: the liquid is being carried where it should not be, usually a dripping or wetting problem rather than a volume one.
- Droplets on the tips after dispense: the tip is not clearing fully, which points at blowout, adhesion, or an ending that breaks off too gently.
- Bubbles as the liquid dispenses: air is coming out with the liquid, which almost always means it went in too fast or is being dispensed above the surface when it should not be.
- An aspirate or dispense height that sits too high or too low: the tip is fighting the well geometry, not the liquid, and that is a labware or offset question before it is a class one.
- Channels that track the surface wrongly, or follow when they should not: level detection or the labware definition, again more about the deck than the parameters.
Only once the transfer looks right is it worth measuring it. Visual inspection tells you whether the mechanics are sane; measurement tells you whether the volume is true and precise. Doing them in that order stops you from chasing a number when the tip is visibly dripping.
The symptom-to-parameter map
When a symptom persists, trace it back to the property that causes it, because the property tells you which knob to reach for. The map below covers the faults you will meet most often.
Dripping and stray droplets
A liquid that drips is telling you about its vapor pressure or its surface tension. Volatile liquids, the alcohols and ethers with high vapor pressure, off-gas and shed drops; low surface tension liquids leave the tip too readily. Reach for anti-droplet control first, then a larger blowout volume to give the vapor room and clear the tip, and pre-wet the tip so the first transfer is not the worst one. For a low surface tension liquid a larger air transport volume gives the column a buffer behind it, and a shorter settling time limits the time a drop has to form. A stop back volume pulls a hanging drop back right after the dispense.
Bubbles on dispense
Bubbles mean air is leaving the tip mixed into the liquid, and the usual cause is a flow rate set too high, so suction or expulsion outruns the fluid and entrains air. Lower the flow rate, and for a viscous liquid lower it further and lengthen the settling time so the column can catch up. Dispensing into the surface rather than shooting a jet above it also stops a stream from beating air into the well.
Short fills and quiet under-delivery
When the delivered volume comes back low, separate two cases. If it is low and consistent, that is a trueness problem, a steady bias, and the fix is the correction curve: measure the real delivered volume across your range and tell the instrument to aim higher by the offset it needs at each volume. If it is low and erratic, look for a physical cause. A flow rate too high for a viscous liquid pulls in air instead of liquid and short-fills; evaporation robs small volumes in dry air before you ever measure them; a liquid that wets the tip walls leaves some behind, which an over-aspirate volume can compensate for on a single accurate transfer. Match the flow rate to the viscosity, and for small volumes control humidity and keep the tip near the surface.
Over-aspiration at low volumes
At 10 or 50 microliter scale a liquid can climb the tip on its own through capillary action, and you end up with more in the tip than you asked for. This is a low-volume, small-tip regime with its own rules: move to a smaller tip, keep the tip close to the surface, turn liquid following off, and tune the blowout to counter the capillary pull rather than feed it. A liquid with a low contact angle spreads thin and is hard to pick up at all, which is a related low-volume trap.
Carryover between wells
Carryover, where a trace of one well shows up in the next, is the enemy of a serial dilution, and one parameter causes most of it: the over-aspirate volume. The excess it draws and dispenses back, helpful for a single accurate transfer, becomes a route for contamination once you reuse the path. Remove the over-aspirate volume for serial work, use a fresh tip every step, and mix below about 80 percent of the well volume so you are not splashing liquid up toward the tip.
Wrong heights and bad following
If the tip aspirates air, crashes toward the bottom, or follows the surface into nonsense, suspect the deck before the class. Heights and level following are expressed relative to the labware, so a definition that says a well is deeper or shallower than it really is, or a plate seated a millimeter off, corrupts every transfer built on top of it. Check that the labware definition matches the physical labware and recalibrate the offset before you touch a single class parameter.
When it is the deck, not the class
Not every symptom is a liquid class problem, and the fastest way to waste an afternoon is to tune parameters against a fault that lives elsewhere. Before you commit to editing a class, rule out the things around it.
- The labware and offsets: wrong or uncalibrated geometry produces height and following errors that no parameter change will fix.
- The environment: temperature, humidity, and even atmospheric pressure shift how a liquid behaves, so a class that worked last week can drift when the room does. Confirm the conditions match the ones the class was built in.
- The instrument itself: a poorly maintained or miscalibrated handler will not deliver well however good the class is, and a worn syringe, a leaking seal, or a fouled pressure sensor shows up as noise you cannot tune out.
- The fluidic path and tips: the pump and syringe cap what is achievable, and a tip that does not match the one the class was built on invalidates the calibration.
A useful habit is to keep one validated class and one known liquid, water is ideal, as a reference. If water on your deck also misbehaves, the problem is almost certainly the deck or the instrument, not the class for your difficult reagent.
Change one parameter at a time
Whatever the symptom, the discipline that separates repeatable troubleshooting from lucky guessing is the same: change one parameter, run it, and see what moved. It feels slower because each iteration costs a run, and it is much faster overall, because you learn which knob does what for this liquid instead of arriving at a combination you cannot reproduce or explain. Let the property behind the symptom choose the knob, adjust only that, and keep the changes that helped.
Do not turn knobs until the problem disappears. Read the symptom back to the property that causes it, change the one parameter that property controls, and you will know why the fix worked.