Tell us about yourself!
I think the instructions should be updated to correct the error. As is, its just a matter of time before something burns. Adding a diode to block supply from the battery would be a sensible step until the author decides to correct the topology.I ended up building my own design from scratch. It's fully scallable too. I made a few PCBs and have given them to a few friends. I got carried away but finally have a robust design. It consumes ~40ma with wifi logging. Minimal heatsink ing needed.It can handle 100V on the input and so far have one of them running with 5 * 200w panels. I found electrical protection quite challenging to account for all scenarios including MOSFET and logic controller failures (important as I am away from home for periods at time). But it's solid now.
A picture of a few revisions. They handle about 15A continuous output and have temperature throttling of the MOSFET long durations above 15A. Starting to get it to fit on one board now.
Just remove it from the code. Originally I was experimenting with periodically increasing the step size to help find the MPPT point. My idea was that if it got stuck on a local max a periodic step would force it to move and re-evaluate.