19. Project Development

Methodology ?

Because I work before in a big IT company I usually use Agile or Scrum. So I usually split project in many tasks, define priority, sprints and so on.

I always says “a big and complicated project never exists, it just many small projects that include many small simple things”.

Working alone on a project is very difficult because we are always head on.

At AgriLab, we use redmine. On my previous job, I was using Jira. Theses tools are very useful when you are working in team. But I don’t know to use special tools, I work alone on my project, I don’t need to synchronize with myself (my other 22 personalities agree with me), I made it old school with papers (It’s seems stupid but it’s so relaxing to take time to just use a pen. It helps me to get my head out of work.) I also use OneNote, I love use it, no real rules, almost free to put anything everywhere like a paper. I can also put it in the cloud, synchronize everywhere. For private notes, I work only local and made my own backup. For quick local text note, I use notepad also.

For this final project, I work in spiral. Because of my work and COVID it was very useful, my only big problem is to finish at least spiral 1 and have something working to show at the end. I have the global idea of my project since the first weeks of fabacademy but I really started to fully think on how to make it only in the middle.

I work with just planning functionalities in a lightweight agile methodology. I have functionalities and nesting of them, but I have also many questions I can’t answer before working on each parts each week. It was only 2 weeks before final presentation I really know what parts can I reuse, which microcontroller I will take, can I keep TFT screen like wanted or I have to switch on B plan? The main thing on my project is something I always do in all my project : the well known B plan! I always have at least B plan (and often C and D), fall-back escapes.

what tasks have been completed, and what tasks remain?

All the first spiral is done. The litter robot works & I have basic functionalities on it. Cats are using it. Now I’m in long testing in real condition. I also checking in litter still works with a long uptime, no memory leak, no electrical failure.

After that test, I will see if all my components’ choice was good. If it succeed, I can directly go to step 2. If not, I will rework step 1 before.

what’s working? what’s not?

I have all what I want in first spiral. I reuse all I need on litter robot (excluding motherboard & front panel as planned) and I have all factory functions.

what questions need to be resolved?

All questions for spiral 1 are resolved.

For spiral 2, I have to find a new screen with same ILI9341 & XPT2046 in a correct form factor and correct price before working on new board & new front panel.

When going to online add-on, I have to define what functionalities I implement & in which order. Will I make a first version and keeping it upgradeable or go directly to full version? If I want to make it upgradeable, will it be in OTA or flash with computer? Will I keep SAMD21 and add connectivity with an ESP8266 (or ESP32) in master/slave communication or will I remove SAMD to reduce the cost (but more development for basic stuff)?

I have made some tests with FDX-B reader but range isn’t enough for me. If I decide to keep it, how can I have better range? Should I wire a new antenna or something else?

what will happen when?

With all I have learned, I want to enhance project and work on further spiral. Doing all spirals and ideas will be the best but at least my next goal is to work on v2 and design an optimized board, find another tft screen and make a thinner version of the front panel. After this, I want to purpose a kit for other owners of this litter robot.

My next goal is to work on my other litter robot (from same brand), it’s the newer version but finally electronic also fails after 2 years… And parts are very expensive, for example basic offline motherboard cost 90€. They add a new system to detect if the bin is full, it’s just IR leds and sensors… it costs also 90€ and I see in forums that part fails often. You can’t just remove and deactivate that part, you just have to pay…

what have you learned?

Doing fabacademy was for me an opportunity to push my limits and learn something new every week. I have of lot of work at AgriLab and I’m often frustrated to haven’t enough time to test new things, use more machine we have. In my mind it was “It’s now or never for that”.

Even if I was a former developer and know how to code in many languages, I wanted to learn new languages. It helped me to reconcile with development and having fun again.

Electronic (behind resistor, LED and arduino) was totally new for me and I love learn that, it’s a great entry point into that world. I hope have time to work more on this and improve my skills. I learned circuit design, reading electronic datasheets, circuit milling.

3D design for CNC and be autonomous for milling my objects was new for me because I haven’t got the time before. It’s a big entry point for my projects in my car company !

In bulk: mkdocs & typora, animation & render in fusion360, how to use (correctly) blender, photogrammetry, nodeJS, working on a method to enhance 3D scan, take time to benchmark prusa mk3 and ultimaker, use the Form 2, use the Markforged mark 2, ink/stitch, kicad, use python/nodeJS/mathematica to talk with boards.

I have many ideas for new workshops at AgriLab, I just hope having enough time for that!