My process is a bit rudimentary. I just have a folder with notebook paper in it. When I need to create a new system or subsytem I will draw it on paper just as it would look in any of these diagraming tools. Once ready I'll use lucidchart to replicate the drawing (schema or flowchart). Once it's implemented in code and committed in Git I will go back to the electronic copy, mark in red or other colors anything that will change in the next refactor or implementation. As I refactor I change the red makrings to black indicating it is now part of the system. The paper becomes an artifact of the first revision.

Sometimes I'll break out a new paper and do the initial drawings to test my knowledge of the system and see if I have complicated the design. This system works well for me.