Incorrect implementation of differential error calculation

[svenwanzenried]

Hi @ussserrr
Thanks for sharing this neat library! I use it to control the temperature in an chamber I use for smoking meat.
Now while implementing, I looked at your code and I think the calculation of the D error is not done correctly.

You are currently calculating the difference between the input and the last_input. But I think it should be the difference between the Perr and the last Perr. Because we want the D portion to react fast to jumps in the error value.

Do you agree?

[ussserrr]

Hello! There are many variations of PID algos and implementations, each of them is tweaking some parameter to better suite some specific cases and show better performance on different machines. I believe you're talking about one of these formulas there. Anyway, it was some kind of student project for me long time ago already and I don't really recall such small quirks anymore :)