I very much doubt the fun was found in the physical act of placing units. Instead I’d wager it was in the immediacy of the chaos, and correction of it, as you dealt with the smaller problems — and the micro-optimizations you could apply being involved with every minor detail of the system.
When you switch to blueprints, you move up the abstraction chain; there’s still some optimization in constructing blueprints but largely you’re now just plumbing… chaining together higher level components, simply ensuring inputs == output, and all the microoptimizations are hard to justify because they disrupt your blueprints. The patterns and structures have reduced to only what you’ve specified in your library (where previously you were discovering them out of the many operators the game provides).
But the literal act of placing objects wasn’t the point, yet that’s all the game has become.
I've found that for me, I lose interest in a factory game the moment I have to scale up. It's when progressing through the game means copying an existing solution to a problem and pasting it (a bunch of times) somewhere else rather than coming up with new solutions.
In general, I get bored when I've figured out the solution to a problem but it takes a long time to get it implemented. Which is also why building conveyor belts is typically my least favourite part of any factory game and why I very rarely refactor my base.