An auto bed leveling (ABL) probe drives around the bed, touches down at a grid of points, and records how far off each point is from where the nozzle expects it to be. The firmware turns that grid into a mesh: a map of small Z-axis corrections it applies during printing so the nozzle stays the same distance from the bed everywhere, even though the bed itself is a little higher in the middle, or tilted slightly to one corner, or has some other real, physical, non-flat variance.
That's the whole job: compensation. The bed is still not flat after running ABL. The printer is just now aware of exactly how not-flat it is, and adjusting for it in real time.
What ABL is good at
- Small, gradual variance across the bed surface, the kind that comes from a bed that isn't perfectly flat out of the box, or has warped slightly over time.
- Keeping first-layer height consistent across a large print, where a single manual level at the center wouldn't account for how the bed behaves at the edges.
- Catching drift over time, if it's run regularly rather than once and forgotten.
What it can't fix
| Symptom | Real cause |
|---|---|
| First layer still doesn't stick despite a good mesh | Bed temperature, surface cleanliness, or first-layer print speed, not leveling |
| One corner is consistently bad, others fine | Often a mechanical issue: a loose bed mount, a binding axis, or a gantry that isn't square, not something a Z correction can compensate for |
| Inconsistent extrusion across the first layer | Partially clogged or dirty nozzle, or moisture in the filament, unrelated to bed height |
| Mesh looks fine but prints still warp off the bed | Cooling, material shrinkage, or adhesion, not a leveling problem at all |
Before trusting a mesh, check the things ABL can't see
The easiest trap here is running the mesh again after a bad print, seeing it "succeed," and assuming the underlying problem is fixed. ABL will faithfully compensate for a warped bed every time, right up until the warp exceeds what the probe's correction range can handle, or the problem was never the bed to begin with. Before re-leveling, check four things the probe can't see:
- Is the bed surface genuinely clean, with no oil, dust, or old adhesive residue?
- Is the nozzle clean and free of a partial clog, especially after switching materials?
- Is the gantry mechanically square, not just electronically compensated?
- Is the first layer printing at the right temperature and speed for the material, not just the right height?
If all four of those check out and the first layer is still inconsistent, the mesh itself is worth re-examining. If they don't, no amount of re-leveling will fix what's really a mechanical or material problem wearing a leveling symptom.
Takeaway
Auto bed leveling compensates for a bed that isn't flat. It doesn't diagnose a dirty nozzle, a loose gantry, or wet filament, and re-running the mesh won't fix any of those. If a leveled bed still isn't printing a good first layer, look at what the probe can't measure before blaming the mesh.
