I bought a 3D printer expecting a manufacturing machine and got, instead, a small and patient teacher of the phrase "it's probably the first layer". A month in, I have a drawer of failed prints, a working benchy, and a much better understanding of why people who own these things talk about them the way they do. This is the diary I wish I'd read before I started.
The first layer is the whole game
Nearly every failure I had in the first fortnight traced back to the first layer. If the first layer goes down clean, evenly squished, well stuck to the bed, the rest of the print mostly looks after itself. If it goes down badly, you are watching a slow-motion accident for the next four hours and pretending it might recover. It will not recover.
Two things fixed most of mine. The first was bed levelling, which on my machine is a manual paper-drag affair: slide a sheet of paper under the nozzle at each corner and adjust until you feel a slight drag, not a jam and not free movement. The second was the first-layer height and flow, tuned by printing a single-layer test square and looking at it under a desk lamp. You want the extruded lines to just merge into each other with no gaps and no ridges. Too high and you get a loose raft of spaghetti that peels; too low and the nozzle drags through what it already laid down.
Adhesion, and the great bed-surface argument
My early prints kept letting go of the bed partway through, curling up at a corner like a sandwich left out too long. The internet has strong opinions about bed surfaces, and most of them are right for somebody. For me, on PLA, the answer was boringly simple: clean the bed. Skin oils are the enemy of adhesion, and I'd been handling the plate constantly while levelling. A wipe with isopropyl alcohol between prints did more than any glue stick or hairspray trick I tried.
I also learned to use a brim. A brim is a single-layer skirt printed flat around the base of the model, attached to it, giving the part more contact area to hold on with. You snap it off afterwards. For tall thin things it's the difference between a print and a small plastic tumbleweed rolling around the bed.
Temperature is not one number
I assumed "PLA, 200 degrees" was a setting, like an oven. It is a starting point. Different filaments, even different colours of the same brand, print best at slightly different temperatures, and the way you find out is a temperature tower: a model that prints a stack of blocks, each at a temperature five degrees lower than the last, with a gcode trick to change the hotend setpoint between sections. You print one, you look at which block has the cleanest bridges and the least stringing, and you write that number on the spool with a marker. It feels like overkill. It is the single most useful calibration print there is.
Stringing, by the way, is those fine cobweb threads strung between the towers of a print, and it's mostly a retraction problem: how far the printer pulls the filament back when it moves across a gap. Too little and it oozes; too much and you grind the filament. Another tower, another afternoon, another number on the spool.
The failures I'm glad I had
- A print that shifted layers halfway up, every layer above offset by a few millimetres, because a belt was loose enough to skip. I'd never have checked the belts otherwise.
- A part that warped off the bed and got dragged around, welding itself to the nozzle into a blob the size of a walnut. Cleaning that taught me how the hotend actually comes apart.
- A beautiful print that was the wrong size, because I'd scaled it in the slicer and forgotten. No machine fault at all. Pure operator.
Where I've landed
The thing nobody tells you is how much of 3D printing is not printing. It's levelling, cleaning, calibrating, tightening, and writing numbers on spools. The actual printing is the easy bit, the bit the machine does while you make tea. Once I stopped expecting it to be a press-button appliance and started treating it as a small workshop tool that needs maintaining, the failure rate dropped off a cliff.
The benchy on my desk is not a masterpiece. The hull has a faint seam and the chimney is slightly rough. But it printed clean on the first attempt after a month of attempts that didn't, and I am unreasonably pleased with it. Onwards to something that isn't a boat.