...and I did it the hard way.
In the past couple of days I've succesfully switched from EasyEDA to KiCad for my newest effort at trying to get ipdm56v2 finalized.
The key things were:
I now have practically the same workflow as with EasyEDA, but in a much better behaving, faster, more capable and more ethical program and I'm pretty sure I won't be ever giving this up.
https://github.com/freerouting/freerouting
This is the open source autorouter I've been using. It has a long history of ups and downs and is written in Java... but as the current end result, it's capable enough to take a LOT of manual, tedious work off my hands
Basically it does what an autorouter should do:
With those three aspects, it will eventually succeed. And that's all I care about as I open youtube to watch funny cat videos after spending hours (days?) just placing parts on the board
Designing complex circuit boards with a reasonably capable autorouter is a day and night difference compared to not having one available. Not having one is like coding without version control. Having to stress about every little thing and not having the proper ability to go back and restart from a known good state really kills productivity and enjoyment
I've locked down all traces and vias that have critical dimensions or positioning. Due to this, I can at any point in time just select all unlocked traces and vias and delete them all, and then move components freely around, optimizing them or making room for more. And when I want the traces back, the autorouter will do it for me, every time, no problem, customized for whatever part placement I did this time. The only thing required from me afterwards is minor DRC related adjustments
Talking about version control, I'm now managing schematic and layout versions in Git too. This also makes a huge difference to how reliably one can progress, because one is always standing on "solid ground". Permanent, labeled, branchable, diffable history is such a game changer