Why Your First Layer Keeps Failing (And How to Fix It)

Your print has been running for two hours. You check on it and find a tangled mess of spaghetti filament, or worse — nothing, because the first layer peeled off the bed in the first five minutes and the nozzle has been drawing in the air ever since.

If this sounds familiar, you are not alone. A bad first layer is the single most common cause of failed prints. Nearly every catastrophic failure — spaghetti, warped parts, prints stuck so hard they crack the glass — traces back to the same root cause: the nozzle was either too far from the bed, too close, or the bed surface was simply not ready to bond with the filament.

This guide covers everything you need to know to fix it.

What bed leveling actually does

"Bed leveling" is a slightly misleading term. You are not necessarily making the bed physically flat — you are teaching the printer where the bed surface is relative to the nozzle at every point across it.

A 3D printer needs to deposit that first layer within a very precise gap — roughly 0.1–0.2 mm between nozzle tip and bed surface. Too far away and the filament won't bond; it just falls onto the bed like loose spaghetti. Too close and the nozzle scrapes the bed, the filament can't flow, and you risk damaging the print surface or the nozzle itself.

Because most printer beds are not perfectly flat — they bow slightly in the middle, or one corner sits higher than the others — the printer needs a map of these height variations so it can compensate on the fly as it moves. That is what bed leveling creates: a correction mesh the printer uses during the first layer.

Manual leveling: the paper test

Manual bed leveling is still the baseline, even on printers with automatic sensors. It gets the bed close enough that the auto-leveling sensor can do its job accurately.

The process:
1. Heat the bed to your printing temperature (the surface expands when hot — always level hot).
2. Home the printer (bring the nozzle to X0 Y0 Z0).
3. Disable the steppers so you can move the head by hand.
4. Slide a single sheet of standard printer paper between the nozzle and the bed.
5. Adjust the corner knob until the paper slides with slight resistance — you should feel it drag, but still be able to pull it free without tearing.
6. Repeat for all four corners, then recheck them all because adjusting one corner affects the others.
7. Finish with the center of the bed.

The target is the same slight drag at every point. Spend five minutes getting this right before relying on any sensor — a sensor can only compensate so much for a bed that starts wildly out of tramming.

Auto leveling: BLTouch, CR Touch, and eddy current sensors

Auto bed leveling sensors measure the bed surface automatically and build a correction mesh before each print (or on demand). The printer then raises and lowers the Z axis mid-print to follow the mesh.

BLTouch / CR Touch — a small probe with a retractable pin that physically touches the bed at a grid of points. Very accurate, works on almost any surface. The CR Touch is Creality's version with a metal pin instead of plastic. Both are retrofit upgrades for most budget printers (Ender 3, CR-10, etc.) and cost €30–50.

Eddy current sensors — used on newer Bambu machines (including the X2D) instead of LiDAR. An eddy current sensor sits on the toolhead just above the nozzle and measures its distance to the metal bed surface by emitting a magnetic field — the strength of the induced current changes with distance, giving a non-contact height reading at every point across the grid. Fast, no moving parts, and doubles as the Z-homing sensor during flow calibration.

Nozzle load cells — used on Prusa's CoreXY printers (Core One+). The toolhead detects the exact moment the nozzle itself physically taps the bed surface. Because the measuring point is the nozzle tip, there is zero probe-to-nozzle offset to configure. The measurement is inherently accurate regardless of nozzle size or filament residue.

Even with auto leveling, you still need to set the Z offset correctly. The sensor tells the printer the shape of the bed; the Z offset tells it the absolute height — i.e., how far above zero the bed actually sits.

Prusa Core One+ vs Bambu X2D: a real-world comparison

These are the two current enclosed CoreXY printers most commonly compared by serious beginners stepping up from entry-level machines. They cost very different amounts and take very different approaches to the first layer.

Prusa Core One+: open and repairable

The Core One+ uses a load cell built into the Nextruder — the nozzle physically taps the bed surface at each probe point. Because the nozzle tip is the sensor, the measurement is accurate regardless of nozzle diameter or what material is loaded. No offset to configure.

First layer calibration is largely automatic: the printer runs its mesh probe, performs a Z alignment calibration, and is ready to print without much input from you. A live Z adjustment is accessible by holding the control knob during the first layer if you need to fine-tune — but this is less prominent than on older Prusa models like the MK3. The Core One+ is more automated than its predecessors; don't expect the same interactive calibration experience you may have read about in older guides.

The real differentiator is what happens when something goes wrong. The Core One+ runs open-source firmware with a public GitHub, a thorough Knowledge Base, and a large community. Settings are documented and exposed. If your first layer is off, there is almost certainly a forum thread or a specific calibration step (like Z Alignment Calibration under Printer → Control → Calibration and Tests) that addresses it directly.

What this means in practice: Day-to-day, the Core One+ is not harder to use than a Bambu. The difference shows up over time: open firmware, a repairable design, and a community that documents everything.

Price: ~$1,199 assembled / ~$949 kit. The kit version takes 4–6 hours to build and is a legitimate option if you want to understand the machine before printing with it.

Chamber temperature: up to 55°C. Suitable for ABS, ASA, and PC with the enclosure closed.

Multi-material: Prusa now offers the INDX toolchanger system as an add-on (€899 for the 8-toolhead conversion kit). Rather than routing multiple filaments through one nozzle, INDX gives each material its own dedicated toolhead — so there's no cross-contamination, no wipe tower, and tool changes use only ~13 mg of purge material, roughly 5× less waste than filament-switcher systems. Tool swaps take about 12 seconds. It's a more capable approach than the X2D's dual nozzle at the cost of a significant additional investment.

Bambu X2D: automated and fast

The X2D uses an eddy current sensor on the toolhead — a departure from the LiDAR that made the X1C famous. It probes a grid across the bed before every print, runs flow calibration, and starts printing without asking you to do anything. The entire pre-print routine is automatic.

What Bambu calls "first layer calibration" is a background process — the printer prints a small calibration pattern, the sensor reads it, and adjustments happen in firmware. You watch the first layer go down and it is usually correct from the very first print after unboxing.

The X2D also has a dual nozzle system (mechanical switching, Bowden-fed) for two-material prints — a feature the Core One+ doesn't offer at base price. The AI toolhead camera monitors the print in progress and can pause if it detects a failure.

The tradeoff is opacity. When the first layer isn't right on a Bambu, it's less obvious what to adjust because you haven't been part of the calibration process. The Z offset exists in the settings, but Bambu's UX doesn't surface it prominently — most users never touch it, which is fine until the day they need to.

There is also a longer-term concern worth knowing about. In January 2025, Bambu removed direct cloud printing access from OrcaSlicer — a popular open-source slicer — and forced all connections through Bambu Connect, their proprietary middleware. When a developer created a fork that restored the removed functionality, Bambu sent a cease-and-desist and the project was shut down. The situation is legally messy: Bambu Studio was originally built on PrusaSlicer's AGPL-licensed codebase, meaning Bambu's own software is distributed under a copyleft licence — and sending legal threats to a developer using that same licenced code has drawn significant criticism from the open-source community. Right-to-repair advocate Louis Rossmann pledged $10,000 toward the developer's legal defence.

The practical implication for buyers: Bambu printers depend on Bambu's cloud infrastructure to function fully. If Bambu restricts that access further — or shuts down — third-party slicers stop working with their hardware. That is not a hypothetical risk; it has already started happening.

Price: $649. Nearly half the price of the assembled Core One+.

Chamber temperature: up to 65°C. A meaningful advantage for engineering materials like PA-CF or PC.

Which one should you choose?

If your priority is getting great prints with minimal fiddling — and especially if you want dual extrusion — the X2D is hard to argue with at $649. The automation genuinely works, and for most materials most of the time, you just press print.

If you want to actually learn how 3D printers work, prefer open-source software and community support, or simply don't want your printer's functionality tied to a company's cloud infrastructure, the Core One+ is the better long-term bet. The higher price reflects better materials, a repairable design, and an ecosystem that isn't going to lock you out.

Side-by-side

The Z offset: the most important single setting

The Z offset is the vertical distance between where the sensor triggers and where the nozzle tip actually touches the bed. Get this wrong and no amount of bed leveling will save you.

Too high: the nozzle is too far from the bed. Filament comes out and sits on top of the surface rather than pressing into it. The lines look rounded and don't bond to each other. The print will detach with almost no force.

Too low: the nozzle is dragging into the bed. Filament is squished so flat it blocks the nozzle path, creates ridges, or damages the surface. You will hear a scraping sound.

Just right: the first layer lines are slightly squished — flattened, not round — and they bond side by side to form a smooth, slightly shiny surface. This is called "elephant foot" if it's overdone, but a tiny amount of squish is exactly what you want.

How to dial it in:
- Print a first layer test (a 100mm × 100mm single-layer square at 0.2 mm height is perfect).
- Watch it print. Adjust the Z offset in 0.05 mm increments until the lines look squished flat and slightly glossy.
- The lines should not have visible gaps between them, and the surface should look smooth, not ribbed.

Most slicer front-ends (Prusa Slicer, Bambu Studio, Cura) let you make live Z offset adjustments during the first layer. Use this — it's the fastest way to find the right value.

Bed surfaces: glass, PEI, and textured PEI

The surface your filament adheres to matters as much as the leveling itself. Almost every current printer ships with a PEI spring steel sheet — if yours came with bare borosilicate glass (common on older or very budget machines), a PEI upgrade costs €15–20 and is the single best first improvement you can make.

Smooth PEI (spring steel sheet): the current standard. PLA, PETG, and ABS all bond well when hot. The magnetic sheet pops off the bed — flex it and prints release almost by themselves as it cools. No glue needed for PLA. PETG can stick aggressively; a thin layer of glue stick acts as a release agent.

Textured PEI: the micro-texture grips filament better, especially useful for PETG and flexible materials. It also hides layer lines on the bottom surface of prints. Downsides: the texture transfers onto the bottom of your part, and ABS/ASA can bond so hard it damages the surface.

The short version: smooth PEI for clean bottom surfaces and easy PLA/PETG release; textured PEI if you print flex or want that matte bottom finish.

Clean the bed before every print

This one sounds obvious and is the most frequently skipped step.

Skin oils, dust, and residue from previous prints create a barrier between the filament and the bed surface. A first layer that bonded fine yesterday may peel today because you touched the surface to check if the print came free.

Wipe the bed with 90%+ isopropyl alcohol (IPA) and a lint-free cloth or paper towel before every print. Let it evaporate fully before you start. This takes thirty seconds and prevents a significant percentage of adhesion failures.

Do not use acetone on PEI — it will damage the surface over time. IPA only.

First layer settings in your slicer

Even with a perfect physical setup, bad slicer settings can ruin the first layer. The three most important:

First layer speed: slow down. 20–30 mm/s for the first layer, regardless of your normal print speed. The filament needs time to bond before the nozzle moves on. Most slicers already default to a lower first layer speed — check that it's enabled and not accidentally set to 100%.

First layer line width: wider lines have more surface area and bond better. A first layer width of 120–150% of your nozzle diameter (so 0.48–0.6 mm for a 0.4 mm nozzle) is a standard recommendation. This is sometimes called "extrusion width" or "line width" depending on your slicer.

First layer height: 0.2–0.3 mm for a 0.4 mm nozzle is the sweet spot. A thicker first layer is more forgiving of slight bed imperfections. Some people use 0.3 mm specifically for this reason, even when printing at 0.2 mm layer height for the rest of the model.

A reasonable starting point for PLA on PEI:

Quick troubleshooting reference

Not getting consistent results?

Dialing in a 3D printer takes time, and some printers are simply harder to tame than others. If you'd rather skip the troubleshooting and just get a well-printed part, that's exactly what our 3D print service is for — you send the file, we handle the settings, and you get a finished print.