Splitting & Joining
Hide seams where the eye already expects a line
Cut along a panel gap, a belt, a hairline, a roof edge — anywhere the model already has a hard boundary. A seam on a smooth cheek or a flat wall needs hours of filler; a seam tucked into an existing crease needs none. Plan the cut in Slicer Pro before you think about clearances.
Three keys beat six
Alignment keys exist to register two faces, not to hold them together — that is the glue's job. Three well-spaced pins locate a face perfectly. Six pins means six chances for one to be 0.1 mm proud, and one proud pin holds the whole joint open.
Print one test pair before all twelve segments
Slice off two 15 mm coupons that carry the same key geometry and print those first. Twenty minutes now, versus discovering after 30 hours that every socket on the model is 0.15 mm too tight. Tolerance Calibrator turns that measured coupon into the clearance number to use.
Number the segments in the model, not on a sticky note
Emboss or engrave an index on the inside face of every piece before exporting. Twelve grey segments in a box all look identical three days later, and a mirrored pair that goes on backwards is a full reprint.
Dry-fit the whole model before any glue opens
Assemble the entire thing with low-tack tape first. Tolerance stack-up is real: eight joints that are each 0.1 mm proud become a 0.8 mm gap at the far end. You can sand that out dry. You cannot sand it out once epoxy has cured.
Match the glue to the plastic, not to the shelf
CA is fast and brittle, epoxy fills gaps and takes load, and solvent welds ABS/ASA into a single piece. A thin, even bead beats a generous one every time — squeeze-out is wasted strength and extra cleanup. Adhesive Calculator names the right one and the quantity in grams.
Before You Slice
Repair the mesh first, always
Slicers do not refuse broken geometry — they guess at it, and the guess shows up as a mystery hole on layer 400. Thirty seconds in STL Doctor is the cheapest insurance in the hobby, and scanned or sculpted files are almost never watertight.
Let orientation cut your supports, not your support settings
Rotating a part 20° often removes more support material than any amount of tuning tree-support parameters. Score the options in Orientation Optimizer and take the stance with the least overhang area and the best bed contact.
Hollow it before you argue about infill
On a display piece, the debate between 10% and 15% infill is worth a few grams. Hollowing the model is worth most of the spool. Add drainage holes at the same time with Hollowing & Drainage so trapped resin, water, or powder can escape.
Add walls, not infill, for strength
Going from two perimeters to four does far more for a load-bearing part than doubling infill density, and costs less material. Infill mostly stops top surfaces from sagging; walls carry the load.
Nest the whole batch in one pass
Printing segments one at a time pays the heat-up, first-layer, and cool-down tax on every single one. Let Build Plate Packer rotate and nest them onto as few plates as possible — then check the total against your spool with Mega Estimator.
Read the G-code, not the preview
Scrub the first five and last five layers in the G-code Inspector. Gaps in the bottom layers mean an adhesion problem you have not met yet; travel moves across top surfaces mean zits you will have to sand off later.
During the Print
Never start a 30-hour job on a part-used spool
If the job needs more than about 80% of what is left on the spool, fit a fresh one. Runout sensors save some printers and betray others, and a splice mid-print on a visible surface is permanent. Weigh the spool; don't eyeball it.
Earn the first layer, then walk away
Watch the first layer complete — the whole thing, not the first corner. The overwhelming majority of long-print failures are visible in those few minutes, and a restart costs nothing at that point. After layer one, checking on it hourly changes nothing.
Fix the draught, not the temperature
Corners lifting on a big flat part is usually a room problem, not a slicer problem. Close the window, add an enclosure — even a cardboard box will do — before you start raising bed temperature and losing dimensional accuracy to it.
Change one setting per reprint
Change three things and a successful print teaches you nothing. Work down the ranked causes in the Failure Diagnoser in order, one change at a time, and verify on a small tower rather than the full part.
Finishing
Prime before you judge your seams
Filler primer reveals every flaw you cannot see on raw plastic, and hides the ones that were never worth fixing. Sand to 400, prime, then look — you will usually find the joint needs one pass, not the five you were bracing for.
Weight the base before you regret it
A metre-tall hollow print is a sail. Leave an access hole in the underside and pour in sand, plaster, or steel shot before you close it up. Retrofitting ballast into a finished, painted, glued model is not a job you want.
If You Only Take Three
Repair the mesh before slicing, print one test pair before the full set, and dry-fit everything before any glue opens. Those three alone prevent most of the expensive failures on this list.