I bought a printer expecting a manufacturing tool and got a temperamental hobby instead, which I'm told is the correct outcome. The famous calibration cube came out beautifully, exactly once, and lulled me into thinking I'd cracked it. The next four prints were a museum of ways a thing can go wrong: corners lifting off the bed, a first layer that looked like someone had dragged a fork through it, and one memorable print that detached entirely and continued extruding into mid-air. That last one is a sight every owner sees and nobody warns you about.
The recurring villain was bed adhesion. PLA on a cold or greasy bed does not stick, and a print that doesn't stick in the first layer is already finished, it just doesn't know it yet. I'd been printing on a bed I'd handled with bare hands, depositing exactly the oils you don't want there. A wipe with isopropyl before each print fixed more problems than any setting did. Boring, unglamorous, and the single biggest improvement of the week.
The warping was a temperature problem. The corners of larger flat parts cooled and contracted faster than the middle, peeling up at the edges. Two things helped: nudging the first-layer bed temperature up a touch, and adding a brim so there was more surface gripping the plate at the perimeter. I also slowed the first layer right down, on the theory that a slow, well-squished first layer has every chance to bond and a fast one is just hoping.
What surprised me is how much of this is feel rather than numbers. The slicer gives you a hundred settings and the temptation is to treat it as an optimisation problem, tweak everything, measure, repeat. But the print itself tells you more than the settings do. A first layer that's too high looks ropey and you can see gaps between the lines; too low and it's translucent and smeared. Once I learned to read the first layer and stop the print early when it looked wrong, I stopped wasting an hour watching a doomed object slowly assemble itself.
A week in, I've got a small pile of usable parts and a slightly larger pile of failures I'm keeping as a reminder. The thing nobody tells you is that the failures are the point, at least at first. Each one taught me a setting I now understand rather than copied. I'll take that over a machine that just works and leaves me none the wiser.