Summary Report

Aim of the project

Aim of the project was to learn about the facilities at FabLab. Secondary goal was to produce a working automatic house plant watering device.

Final product and its parts

Final product consists of a laser-cut acrylic box on legs as water reservoir, a 3d printed (with SINDOH) mounting plate for the valve system and servo and a "sock" for the tube-closing clip's lip for the servo to act against, an Arduino Uno and an Arduino shield (PCB designed in EAGLE and milled with Roland SRM-20) containing the user interface and connections to moisture sensor and servo motor. Moisture sensor consists of two wires with loops on the ends for nails or screws to stick into the soil. One wire is connected to a digital output pin, the other to an analog input. During measurement output pin is set to high and the voltage reaching the input pin is used as a measure of soil moisture. To reduce wasted electricity and electrolytic corrosion the output pin is set to low when not measuring.

Design sketches were made on paper and in Krita. Laser cut parts were designed in Inkscape. 3D printed parts were designed in Autodesk Fusion 360. Initial electronics design and programming was done in TinkerCAD and finished in EAGLE and Arduino IDE, respectively.

Initial sketches featured a water bottle as the water reservoir, but an acrylic box on legs was made instead. UI faceplate was omitted, but usage of the device was implemented as planned.

Device is operated by inserting moisture sensors into the soil so that they do not touch (and short-circuit), but are still close enough to each other that a meaningful measurement can be made. Water reservoir is filled and watering tube is pointed at the soil. Pressing button 1 cycles through water dose sizes (indicated by LEDs). Pressing button 2 calibrates the "dry" point to that sensed by the sensor. All three LEDs blink after successful dry point calibration.

• Link to repository
  • Direct link to final code
• Link to Final Presentation

Lessons learned

Filament 3D printed PLA is not watertight and needs some kind of surface treatment. UV cured 3D printing might have been a better option.

Hot glue is not ideal for making acrylic boxes, chemical welding with acetone or similar solvent would have been better.

Small changes to a single layer PCB can be made with hand tools, but their mechanical resilience is not as good as "getting it right" on the board to begin with - we had trouble with copper strips delaminating off the PCB after too many fixes.

There are many ways to screw up milling PCBs - watching others make as many mistakes as possible is almost as good as learning from your own mistakes!

EAGLE software's image export works differently between Windows and Mac computer. On Mac fabmodules.org creates file for milling machine that mills too large a board (about twice as big than using Windows). Also scaling the image down didn't yield wanted results.

Recommendations?
Plan, plan and schedule!

Feedback

Brainstorming ideas and making physical things was fun.
Course offered a great way to introduce Fablab environment.