A single-nozzle Bambu printer levels itself against one reference point: its only nozzle. The H2D cannot work that way, because it carries two independent hotends on one toolhead, and Bambu’s documentation confirms it reads bed contact through two eddy current sensors, not one. It also adds a calibration step no single-nozzle Bambu machine needs: aligning the second nozzle’s position against the first. Treat the H2D like a P1P or an X1 Carbon with an extra hotend bolted on, and a print that would look fine on those machines can come out shifted or misaligned here for reasons that have nothing to do with the model file.
How the H2D’s bed leveling works
Z homing lowers the active nozzle until it touches the heatbed; an eddy current sensor detects the contact and the printer records that height. There are two of these sensors, one per hotend: the left sensor sits behind the heating block, the right sensor above it. Homing will not lower a nozzle until it reads at or below 140°C. If the hotend is hotter than that, the printer homes the X and Y axes first and waits for the nozzle to cool before it descends to the bed.
Leveling uses the same sensors for a different job: compensating for small height differences across the bed caused by an uneven heatbed or normal assembly tolerance, so the nozzle-to-bed gap stays consistent everywhere the toolhead travels.
The H2D runs this check in three ways. A per-print toggle in Advanced Options, labeled Auto Bed Leveling, runs a quick pass automatically before each job; leave it on. The same routine also exists as a standalone entry in the printer’s Calibration menu, meant to run after reseating or replacing the build plate, hotend, or heatbed assembly, or after the printer has been moved. A third pass, High-Temperature Bed Leveling, runs before printing ABS, ASA, PC, or PA-family filaments, since the bed does not hold the same shape once it is heat-soaked. Bambu treats this as its own dataset rather than folding it into the standard pass, and the printer stores and applies both automatically.
Clean the nozzle and the plate before running any leveling pass. Residual filament stuck to a nozzle tip can throw off the height reading, the same failure mode documented across the rest of Bambu’s current lineup.
If auto-leveling itself fails, usually after shipping damage or heatbed service, the fallback is manual bed tramming: loosening the screws under the heatbed and adjusting the four corners by hand with a hex wrench, feeler paper, and a leveling test file run from a USB drive. The H2D ships pre-leveled from the factory, so this is a maintenance step, not something to repeat routinely — see Unboxing and Safe First Setup for what the printer does automatically the first time you power it on.
Once leveling is confirmed good, a first layer that still will not stick is a bed adhesion problem, not a calibration one.
Dual-nozzle Z-offset: calibrating the second nozzle
This is the step with no equivalent on a single-nozzle Bambu printer. Only one of the H2D’s two hotends prints at a time. A linked mechanism lifts the idle nozzle and swings a flow blocker across it so it cannot ooze onto the model while the other side is working. Every time the toolhead switches which nozzle is active, that nozzle has to arrive at exactly the height and position the print expects, or the sections printed by each nozzle will not line up.
The fix has its own entry in the Calibration menu, right next to Auto Bed Leveling: Nozzle Offset Calibration. The heatbed contains a dedicated calibration coil, separate from the two eddy current sensors in the toolhead, built specifically to support this measurement. Together, the toolhead sensor and the bed-mounted coil let the printer compare exactly where the left and right nozzles sit relative to each other and store the difference as an offset, so both nozzles print into the same coordinate space instead of two slightly different ones.
The symptom telling you it is time to run this calibration is specific: a visible layer shift or misalignment appearing right at the point where a print switches from one nozzle to the other. That is a different failure signature than a leveling problem, which shows up as first-layer inconsistency spread across the whole bed rather than a step at one height.
On a single-nozzle machine like the A1 mini, P1P, or X1 Carbon, there is only one hotend, so its position is the only reference frame that exists. There is nothing to reconcile it against, which is exactly why this step does not appear anywhere in those printers’ menus or manuals. It is unique to the H2D’s dual-nozzle hardware.
Nozzle Offset Calibration can fail for mechanical reasons that mimic leveling failures: filament residue on a nozzle tip or the heatbed throws off the reading, and specific HMS codes exist for a weak, damaged, or misread calibration sensor signal. If the calibration keeps failing, check for stray filament before assuming a sensor has actually failed.
Flow and extrusion calibration
Flow calibration on the H2D works the same way it does across Bambu’s current lineup. Flow Dynamics Calibration estimates a K value, the printer’s version of pressure advance, that compensates for the delay between commanding an extrusion rate and the filament actually responding to it. In automatic mode, the printer runs this itself and saves the result to that filament’s profile; you do not need to inspect a printed line by eye unless you choose manual mode instead.
There are specific, documented reasons to rerun it: a filament brand or model you have not calibrated before, a nozzle worn from use, a freshly installed nozzle (manufacturing tolerances vary unit to unit), filament that has picked up moisture, or a change to the max volumetric speed or print temperature saved in that filament’s profile. Because the H2D’s two hotends are separate extrusion paths, a new filament or a nozzle swap on one side is its own trigger and does not calibrate the other side automatically.
A second, separate tool, Flow Rate Calibration, checks the ratio between expected and actual extrusion volume. Most users will not need it, since official filaments and printers ship pre-tuned already. It is meant for defects that persist after everything else checks out, or for unusual materials like foaming filaments that expand heavily when heated. Run flow-related calibration after bed leveling and nozzle offset are both confirmed good, not before. Flow cannot fix a print that is already failing because the two nozzles are out of alignment or the first layer never seated correctly.
Chamber temperature and calibration accuracy
The H2D’s chamber uses active heating rather than relying on residual warmth from the bed and hotend. A PTC heating plate and a circulation fan bring it to a documented maximum of 65°C, and once it reaches target, the heater’s power output drops to hold that temperature instead of running at full output the whole time. That is a separate system from the heatbed, which is documented on its own and reaches up to 120°C.
None of the leveling or offset data described above is captured at a fixed temperature. A printer measured cold, before the chamber and frame have warmed up, is not in exactly the same physical state as the same printer an hour into an enclosure-heated ABS or PC print, because metal expands slightly as it warms. The separate High-Temperature Bed Leveling pass built into the H2D is a practical acknowledgment of that: stored and applied alongside the standard leveling data, it means the printer does not treat cold-frame numbers as good enough once the chamber and bed are heat-soaked for an engineering filament.
The practical fix mirrors Bambu’s own first-layer advice: if a print is giving inconsistent first layers despite a good leveling result, Bambu recommends preheating the heatbed and holding it at target temperature for ten to fifteen minutes before judging the first layer, rather than trusting the very first pass. The same thermal-stabilization logic carries over to the chamber on an enclosure-heated print: a leveling or nozzle offset result read on a cold machine right before switching straight into a hot, enclosed print is describing two different physical states of the same printer.
When to re-run calibration
Different triggers call for different calibrations:
- Nozzle swap on either side: run Nozzle Offset Calibration and Flow Dynamics Calibration for that hotend.
- Hotend replacement: run Nozzle Offset Calibration and a full Auto Bed Leveling pass, since geometry near the bed can shift slightly with a new heating assembly.
- Heatbed or build plate replacement, or any heatbed disassembly for service: run Auto Bed Leveling, and fall back to manual bed tramming only if that pass then fails.
- After transport: run Auto Bed Leveling first. If it fails, treat that as the documented signal to check for shipping damage before moving on to manual tramming.
- A new filament brand, or the first use of a filament on a given hotend: run Flow Dynamics Calibration for that side.
- A visible step or misalignment at a nozzle switch point: run Nozzle Offset Calibration regardless of anything else that changed recently.
- Any HMS code referencing homing, leveling, or the nozzle offset sensor: treat it as a prompt to run the matching calibration, not just a message to dismiss.
None of these fix a mechanical fault by themselves. A leveling or nozzle offset pass that keeps failing after a clean nozzle and a clean bed points at hardware, such as a loose sensor connector or a damaged coil, and it is worth checking against the printer’s own HMS code before repeating the same calibration again.
If a leveling or nozzle offset failure keeps coming back after all of that, read it against the H2D’s actual HMS code rather than guessing further. Once bed leveling and nozzle offset are both confirmed good, the site’s general calibration order guide covers where these steps fit against temperature and flow tuning.