Guide · Troubleshooting
Cut quality drops and the shop floor blames the file — so an afternoon goes to re-nesting and re-posting instead of the blade holder. This guide gives you a ten-minute sequence that decides whether the fault lives in the program or in the tooling.
A program fault reproduces itself exactly: the same toolpath produces the same wrong result from the first cut, on fresh material, at any position on the table. A tooling fault degrades: the job starts acceptably and drifts through the run, or fails only in certain table regions where hold-down or support is weak. That one distinction — reproducible versus degrading — settles most "software days" in two runs of the same file.
| Symptom | First suspect | Why |
|---|---|---|
| Every feature offset by the same amount | Program / compensation | A constant error points at a setting, not wear |
| Frayed corners, clean edges mid-run | Blade tip wear | Cornering loads a dull tip first |
| Good at table centre, poor near edges | Vacuum / support | Material lift, not code |
| Not through in one pass — but it was last week | Blade wear | File unchanged, result changed |
Run this before opening the CAM seat. Each item takes about a minute, and it goes in order — edge first, because that is where most of these runs end.
A tired oscillating blade shows a rounded tip and a shiny wear flat along the cutting flank; a staple or grit event can chip the carbide outright. Drag blades dull more slowly and show a polished, blunt edge, often with a burr forming on one side. Photograph the tip against a plain background with your phone at minimum focus distance — if the point photographs as a line, corner quality is already gone. Keeping last month's macro photo next to today's turns "the machine is acting up" into a dated wear record, which is also the fastest way to judge whether your reorder cycle matches real consumption.
Three signs point back at the code: the error appears identically on the first cut of a fresh blade; it repeats at the same coordinates regardless of material; and a second file shows the same behaviour after a full tooling reset. Then check the boring items first — origin shifts, the wrong tool selected in the post, compensation values inherited from the previous job — before the exotic ones. The point of the tooling-first pass is not that software is never wrong; it is that tooling is checked in minutes and software is debugged in hours, so the order matters.
If the pass lands on the tooling side, send this set and we will match a replacement blade to your holder before quoting:
Send the photos above — we verify the blade spec against your holder first, and we will say so plainly if the evidence points at your file instead.