Differences between S1 and S1 Pro based on Pre-Release Firmware V1.0.1.0

updated by 4 min read #3d printing #reverse engineering #hardware

This post compares the firmware of the FLSUN S1 3D printer with the pre-release firmware of the S1 Pro, which had just been announced when I wrote it. It lists what the configuration files and scripts reveal about the new model’s AI detection, calibration, thermal settings and cloud features.

Update, October 2026: FLSUN launched the S1 Pro on 25 November 2024, a week after this post. Everything below is based on the pre-release firmware and may differ from what shipped. For example, the firmware allows a bed temperature of 140 °C, while the product page lists a maximum of 120 °C.

Background#

I was looking into the firmware upgrade methods of FLSUN’s latest 3D printers to develop a jailbreak that gives remote root access. While researching, I noticed the announcement of a new flagship model, the S1 Pro.

I compared two firmware versions: S1 V1.0.6.3 and the pre-release S1 Pro V1.0.1.0. If you want to look for yourself, my repository contains all the versions I could find. Both appear to be further developments of the FLSUN Speeder Pad firmware, which is based on Debian, or more precisely Raspbian.

Summary#

  • AI detection: upgraded from YOLOv5 to YOLOv8 and YOLOv10 models, plus segmentation models for print monitoring.
  • Thermal settings: a higher maximum bed temperature (140 °C instead of 130 °C), on top of the hotter nozzle, and refined PID settings.
  • Print recovery: more elaborate macros for power-loss recovery, filament handling and saving the printer state.
  • Calibration: what looks like a self-adapting pressure advance.
  • Klipper configuration: more macros for motion control and print customisation.
  • IoT: new cloud settings and SSL support for MQTT-based communication.
  • Image capture: changes to image capture and process management, possibly including a helper for video streaming.
  • Cooling: a lower maximum fan power, perhaps for quieter operation.

Overall, the S1 Pro firmware focuses on precision, automation and user experience. FLSUN may also be preparing a cloud platform for device management; at least they are consolidating the firmware upgrade paths for different models.

Key differences#

AI detection#

The biggest changes are in the AI detection. The flsun_func/AI_detect/ directory contains several new files and scripts that the S1 firmware does not have:

  • YOLO models: The S1 Pro ships rknn_yolov8_demotool and rknn_yolov10_demo; the S1 uses yolov5.
  • Segmentation: Files such as rknn_ppseg_demo and segzkpre.rknn point to AI-based image segmentation, which could let the printer analyse print layers in finer detail and detect problems more reliably.

Newer YOLO models and segmentation should improve real-time defect detection during a print. I hope FLSUN brings these models to the S1 as well.

Scripts and process management#

The aidetect.sh and before_printing_run.sh scripts changed substantially:

  • Conditional logic: The S1 Pro adds checks on the layer height (z_height) and more detailed logging, which helps with tracking a print and with troubleshooting.
  • Streaming: mjpg_streamer is replaced by pushStream, which suggests a new way of streaming camera images during a print.

Calibration#

  • Motor calibration: The S1 Pro has more elaborate macros for motor calibration, such as CLEAN_CALIBRATE_MOTOR.
  • Self-adapting pressure advance: The extruder configuration of the S1 Pro contains a self_adaption_pressure_advance parameter that the S1 lacks. If it does what the name says, it adjusts pressure advance during the print instead of using a fixed value.

Thermal and fan settings#

  • Bed temperature: The S1 Pro configuration allows up to 140 °C, the S1 up to 130 °C. This would allow a wider range of materials.
  • Fan power: The S1 Pro limits the fan to a maximum power of 0.55, the S1 to 0.90. This is probably a trade-off between cooling and noise.

G-code macros#

The S1 Pro configuration adds many G-code macros that were missing or simpler on the S1:

  • Power loss and state recovery: more elaborate macros for the filament motion sensor, power loss and state recovery, so interrupted prints can be resumed more reliably.
  • Bed levelling: macros to save and restore the levelling state, so it doesn’t have to be recalibrated as often.

Networking and system settings#

IotCfg.json and SysSetting.ini show changes to networking and security:

  • IoT configuration: a new UpdateHttpRequestUrl parameter, which points to cloud-based interactions and firmware updates.
  • SSL: an ssl file in the S1 Pro firmware, which suggests encrypted communication with the cloud.

Together, these point to a more connected device with remote management.

Hardware hints#

Some things stay the same. The firmware still expects a 720p webcam, which is usually good enough for ML detection. The improvements are mostly in software.

Conclusion#

In its pre-release state, the S1 Pro firmware is an incremental improvement on the S1. When I wrote this, I was looking forward to the final hardware and its firmware release.

As an S1 owner, I would like to see these software improvements, especially the AI ones, come to the S1 as firmware updates, without having to buy an S1 Pro.

FLSUN also uses a big chunk of GPL-licensed code in different components, so please remember to publish that source code. 🤓

Sources#

cd ~/projects