The printer arrived, I built it over a weekend, and then I spent the following week producing a small museum of failure. Nobody warns you about this part properly. The marketing shows you a flawless Benchy gliding off the bed. The reality, for the first dozen attempts, is a tangled bird's nest of filament, or a print that peeled off the plate at layer three and got dragged round the bed like a sad little tumbleweed.
I'd bought a kit, an i3-style machine, partly because building it myself felt like the honest way in and partly because it was cheaper. The build went fine. The lies started when I hit print.
My very first attempt didn't adhere at all. The filament came out, curled up, and the nozzle just smeared it about with magnificent indifference. The second stuck for a bit and then let go. The third produced something that was, technically, a shape, if the shape you were after was "abstract". I changed a variable each time, like a good engineer, and got nowhere, because I was changing the wrong variables.
The thing I was fiddling with was temperature, and flow, and print speed, all the numbers in the slicer that feel like they ought to matter. The thing that actually mattered was the half millimetre of air between the nozzle and the bed, which I'd been treating as roughly correct because I'd "levelled it" once on day one with my eyes.
First-layer adhesion is almost entirely about that gap, and that gap is almost entirely about bed levelling, and bed levelling done by eyeball is not levelling, it's wishing. The trick everyone eventually learns is the paper. You drop the nozzle to its zero height, slide a sheet of ordinary printer paper underneath, and adjust each corner until you feel a slight drag on the paper but it still moves. Do all four corners, then go round again, because adjusting one corner tugs the others out. It feels absurdly low-tech for a machine with stepper motors and a heated bed. It's also the single thing that turned my failure museum into actual objects.
The other half was the surface itself. Bare glass on the heated bed wasn't holding PLA reliably, so I went round with a glue stick, a thin even layer, which sounds like a bodge and is in fact the recommended approach. Between a properly set first-layer gap and a bit of glue, my next print stuck like it had always meant to.
The first thing that came off clean was, of course, a Benchy, the little tugboat everyone prints to prove the machine works. Mine had a touch of stringing and the funnel was a bit rough, but it was a boat, it had a hull and a chimney and it stood up, and after a week of bird's nests I was unreasonably pleased with it. It's still on my desk.
What I'd tell the version of me from a week ago: don't touch the temperature, don't touch the speed, don't touch anything in the slicer. Level the bed properly with a sheet of paper, set the first-layer gap by feel, and only once the first layer goes down clean and flat should you allow yourself to fiddle with anything else. Nearly every early failure is the first layer, and nearly every first-layer problem is that gap. The clever numbers can wait. The boring half a millimetre comes first.