Bed leveling — more precisely, bed tramming and first-layer compensation — is the single most common source of failed first layers in FDM printing, and the All3DP bed leveling comprehensive guide identifies the core distinction that many new makers miss: leveling and tramming are different operations. Tramming makes the bed physically parallel to the gantry plane; leveling (or more accurately, first-layer compensation) measures and compensates for any remaining irregularities in the bed surface that tramming cannot correct. Automatic systems like BLTouch, CR Touch, and Bambu's force-sensing calibration handle both steps with minimal user input, but understanding what they are measuring and why the z-offset matters — independent of any probe system — prevents the confusion that leads to hours of failed prints from a misunderstood 0.1mm offset error.

Manual Tramming: The Foundation

Manual tramming adjusts the physical position of the bed so that each corner is the same distance from the nozzle when the printhead is positioned over it. The process is straightforward: heat the bed to print temperature (thermal expansion changes the geometry slightly from cold), disable steppers, and use a piece of paper as a feeler gauge at each corner while adjusting the leveling knobs or screws until the resistance to sliding the paper is consistent across all points. Four-point adjustment requires iteration — tightening one corner raises that corner and slightly affects the others — so multiple passes through all four corners are necessary until all reach the same resistance. A trammed bed is the prerequisite for any automatic compensation system to work correctly: a probe mesh cannot fully compensate for a bed that is 1mm higher on one side than the other, because the compensation range and the printable z-height are reduced by the uncorrected mechanical error. Manual tramming should be verified periodically even on printers with automatic compensation.

BLTouch and CR Touch: Probe-Based Mesh Compensation

BLTouch (from Antclabs) and CR Touch (Creality's implementation) are servo-actuated probe systems that extend a pin to contact the bed surface at a grid of points, measuring z-height at each location and building a surface topology mesh. This mesh is applied as a real-time z-offset correction during printing — as the printhead moves over regions that are higher or lower than average, the z-axis compensates to maintain consistent nozzle-to-surface distance. The probe offset from the nozzle must be calibrated in firmware before the mesh is accurate: an incorrectly set probe offset applies the right mesh shape at the wrong absolute height, producing systematic first-layer errors that mirror the mesh topology. Always verify the probe offset with a calibration print after installation. The probe must also be positioned to reach all corners of the mesh grid — a probe mounted too close to the bed edge produces inaccurate corner measurements that the mesh cannot correct.

Bambu Auto-Calibration and Force-Sensing Systems

Bambu Lab's approach uses a force sensor integrated into the toolhead to detect nozzle contact with the bed surface — the calibration sequence taps the nozzle lightly against the flex plate at each grid point and records the z-height at first contact. This approach has several practical advantages over magnetic or optical probes: it measures exactly where the nozzle will print (not at an offset position), is not affected by bed surface color or material type, and requires no separate probe hardware that can shift position over time. The resulting calibration is performed at print temperature with the actual nozzle that will be printing, which eliminates the thermal offset error that magnetic probes near a 100°C bed surface can introduce. Bambu's calibration runs automatically at the start of each print for a rapid check and offers a full calibration sequence via the interface. The limitation is that any nozzle contact with a foreign object on the bed surface — a stuck filament blob, a forgotten support remnant — can corrupt the calibration measurement, so bed surface cleanliness is critical before allowing auto-calibration to run.

Z-Offset: The Parameter That Controls First Layer Quality

Z-offset is the distance adjustment applied to every probe measurement to account for the gap between the probe trigger point and the actual required nozzle-to-bed distance for correct first-layer squish. Getting z-offset right is the final step after any leveling system is set up, and it must be dialed in for each combination of printer, bed surface, and filament type. The correct z-offset produces a first layer that is squished enough to bond firmly to the bed without so much squish that the nozzle drags through the material or the layer becomes indistinguishable as a distinct extrusion. Visual inspection of the first layer on a test square is the method: lines that are round and barely touching indicate the nozzle is too high; completely flat lines with no gap between them that look like a smooth sheet are correctly squished; lines dragging and smearing indicate the nozzle is too low. Live z-adjust during the first layer print — available on Marlin, Klipper, and most modern firmware — allows real-time correction without stopping and restarting the print.

When Each Approach Wins

Manual tramming without a probe is adequate for printers with flat, rigid glass beds and four-point adjustment — once a glass bed is trammed true, it stays stable for many sessions. Add a BLTouch or CR Touch when the bed surface has measurable warp, when the printer is frequently moved, or when large parts require consistent first-layer adhesion across the full build area. Bambu's force-sensing auto-calibration is the most hands-off system currently available — minimal setup, no external probe hardware, consistent results across sessions. For Klipper-based printers, adaptive mesh leveling — which measures only the sub-region the current print uses — reduces calibration time on small parts and is worth configuring once the basic probe setup works.

What It Means for Makers

Bed leveling is a recurring calibration task, not a one-time setup — verify it whenever first layers show inconsistency, whenever the printer is moved, and whenever the bed surface is replaced. Automatic systems reduce intervention frequency but do not eliminate the need to understand what the first layer is telling you. A maker who can read a first layer — identifying whether the problem is z-offset, tramming, surface contamination, or temperature — fixes problems in minutes. A maker who only knows to press Auto-Level repeats calibration runs on problems with a different root cause.

Sources