Your bed probe reports the nozzle is 0.1mm from the bed but your first layer is 0.3mm thick because the probe’s standard deviation is ±0.05mm and your Z-offset was calibrated cold while you print hot. Bed probes are only as accurate as their calibration and their fundamental technology allows — and some probe types drift 0.05mm between room temperature and 100°C bed.
Bed Probe Technology Comparison: How Each Type Actually Works
BLTouch / 3D Touch: Mechanical Pin with Hall Effect Sensor
A solenoid-driven plastic pin extends, touches the bed, retracts, and triggers a Hall effect sensor at a precise retraction distance. The trigger point is the same every time (±0.005mm repeatability in ideal conditions). The pin is the physical probe — it actually contacts the bed, meaning it works on any surface: glass, PEI, BuildTak, bare aluminum.
The BLTouch uses an optical sensor for pin position. The cheaper 3D Touch clone uses a Hall effect sensor, which is temperature-sensitive — the trigger point drifts as the sensor warms up inside a heated enclosure. A genuine BLTouch is worth the extra $25 over a 3D Touch specifically for this reason.
Key specs: ±0.005mm repeatability (BLTouch), ±0.02-0.05mm (3D Touch). Works on any surface. Operating temperature: 0-45°C (the solenoid fails above 50°C ambient). Pin is sacrificial — if it crashes, the pin snaps and you replace it for $3 instead of replacing the entire probe.
CR Touch: Optical Sensor, Creality’s Proprietary Design
Functionally identical to BLTouch but uses an optical interrupter instead of a Hall effect sensor. Creality’s version has a metal pin (vs BLTouch’s plastic), which survives minor crashes better but can damage the bed surface. The optical sensor design is theoretically more temperature-stable than Hall effect, but real-world testing shows comparable performance.
Key specs: ±0.005mm repeatability. Metal pin (durable but can scratch beds). Creality’s 5-pin connector is different from BLTouch’s standard 3+2 pinout — requires an adapter or custom wiring for non-Creality boards.
Pinda / SuperPINDA: Inductive Proximity Sensor (Prusa Standard)
The PINDA (Prusa INductive Distance Axis) sensor detects metal distance without physical contact. It reads the steel sheet under the PEI build plate. No moving parts, nothing to wear out, nothing to crash. The SuperPINDA (current version) includes active temperature compensation — the sensor’s detection distance doesn’t change as the bed heats up.
The limitation: PINDA only works with metal build plates. PEI-on-glass requires a steel sheet underneath. Garolite (G10), bare glass, and thick BuildTak surfaces are invisible to inductive sensing.
Key specs: ±0.002mm repeatability (SuperPINDA with temp compensation). No-contact (no crash risk). Requires metal bed surface. Operating temperature: 0-120°C. No sacrificial parts — if it fails, replace the entire $35 sensor.
Inductive Probes (LJ12A3-4, PL-08N): Generic Metal Detection
The ubiquitous LJ12A3-4-Z/BX inductive sensor ($5-10) detects metal at 4-8mm distance. Detection range varies with temperature (typically 0.02mm/°C drift) and target metal type (steel vs aluminum vs brass). This is the “it works until it doesn’t” probe — acceptable on printers with manual bed leveling as a backup, but not reliable enough for automatic mesh compensation on a production machine.
Key specs: ±0.05-0.10mm repeatability. Temperature drift 0.02mm/°C. Only detects ferrous metals. $5-10 cost. 12-36V operating voltage (requires buck converter for 5V logic boards).
Klicky / Euclid: Microswitch Probes (Voron Standard)
A mechanical microswitch mounted on a detachable probe arm. The toolhead picks up the probe before homing and docks it afterward. The switch triggers at exactly the same physical displacement every time — ±0.001mm if the switch is quality (Omron D2F series). This is the Voron community’s standard because it’s nearly drift-free and costs $3 in parts.
Key specs: ±0.001mm repeatability. Zero temperature drift (mechanical switch). Manual docking/undocking adds complexity. Not available as a plug-and-play commercial product — requires printed parts and assembly.
Bed Probe Accuracy and Feature Comparison
| Feature | BLTouch | CR Touch | SuperPINDA | LJ12A3-4 Inductive | Klicky (Microswitch) |
|---|---|---|---|---|---|
| Technology | Hall effect + plastic pin | Optical + metal pin | Inductive + temp comp | Inductive (raw) | Mechanical microswitch |
| Repeatability | ±0.005mm | ±0.005mm | ±0.002mm | ±0.05-0.10mm | ±0.001mm |
| Temperature Drift | <0.01mm (0-45°C) | <0.01mm (0-50°C) | <0.005mm (0-120°C) | 0.02mm/°C (severe) | <0.001mm (any temp) |
| Bed Surface Compatibility | All surfaces | All surfaces | Metal only | Ferrous metal only | All surfaces |
| Contact Method | Pin touches bed | Pin touches bed | Non-contact | Non-contact | Switch touches bed |
| Crash Protection | Plastic pin (sacrificial) | Metal pin (durable) | None needed | None needed | Switch arm flexes |
| Probe Speed | Slow (pin extend/retract) | Slow (pin extend/retract) | Fast (continuous sensing) | Fast (continuous) | Moderate (pickup/dock) |
| Installation Complexity | Medium (wiring + mount) | Medium (Creality-specific) | Low (Prusa native) | High (voltage, mount) | High (3D printed, assembly) |
| Cost (May 2026) | $35-40 | $35-40 | $35 | $5-10 | $3-10 (DIY) |
| Best For | All-around reliability | Creality upgrades | Prusa, metal beds | Budget builds, metal beds | Voron, ultimate accuracy |
Common Bed Probe Mistakes
Mistake 1: Calibrating Z-offset with a cold nozzle and then printing hot. Thermal expansion moves the nozzle tip 0.04-0.08mm downward when it heats from 20°C to 230°C. Your “perfect” cold calibration is now 0.05mm too close to the bed. The first layer is over-squished and your nozzle drags. Always calibrate Z-offset with the nozzle at printing temperature and the bed at printing temperature.
Mistake 2: Setting probe grid too coarse. A 3×3 grid on a 220×220mm bed has 110mm between probe points. If the bed has a localized high spot between two probe points, the mesh won’t detect it. Use at least 5×5 (25 points) for beds above 200mm. For 300mm beds, use 7×7.
Mistake 3: Using an inductive probe (LJ12A3) with an aluminum bed. The LJ12A3 detects ferrous metals — steel, iron. Aluminum is invisible to it. You’ll get “PROBING FAILED” because the sensor never triggers. Some aluminum beds have steel spring steel sheets on top (PEI-coated), which the sensor can detect through the PEI, but bare aluminum is a no-go.
Mistake 4: Mounting BLTouch too high relative to the nozzle. The BLTouch pin must extend 2-3mm below the nozzle tip to trigger before a crash. If the probe body is mounted flush with the nozzle height, the pin extends but the nozzle hits first. The probe never triggers and the print head drives into the bed. The BLTouch mounting bracket must position the probe body 2-3mm above the nozzle tip.
Mistake 5: Enabling multiple bed leveling methods simultaneously. Marlin’s AUTO_BED_LEVELING_BILINEAR and MESH_BED_LEVELING conflict. Enabling both produces a mesh that’s generated with one method and used with another — your probe data is ignored. Pick one method and disable the others.
Mistake 6: Not running M48 probe repeatability test after installation. The M48 test probes a single point 10-20 times and reports the standard deviation. If your BLTouch shows σ > 0.005mm, something is loose — probe mount, gantry, or wiring. Fix the mechanical issue before trusting the mesh.
Mistake 7: Running bed mesh with the nozzle oozing filament during probing. The ooze creates a blob on the nozzle tip that the probe can’t account for. The mesh thinks the bed is higher where the blob is. Heat the nozzle to 160°C (below oozing temp for most filaments) during probing, or retract 5mm and wipe the nozzle before starting the mesh.
⚠️ Safety Notice: Bed probe installation involves wiring into your printer’s mainboard. Incorrect wiring can damage the probe, the mainboard, or both. The BLTouch and CR Touch use 5V logic; wiring them to 12V or 24V destroys the sensor instantly. For inductive probes requiring 12-36V, use a voltage divider or optocoupler to protect 5V logic pins. Always verify pinout with a multimeter before connecting. When a probe fails, the printer may drive the nozzle into the bed — enable
Z_MIN_PROBE_ENDSTOPas a hardware failsafe where possible.
As we covered in our first layer calibration guide, the probe gets you close but the final Z-offset tune is visual. See our motherboard upgrade guide for dedicated probe ports on aftermarket boards.
For builders upgrading their bed leveling system, the uavmodel ANTCLABS BLTouch Smart V3.1 includes the standard 3+2 pin connector, an extended 1.5m cable, and a mounting bracket kit compatible with Ender 3, CR-10, and most open-frame printers.