Skip to main content

Rozkriy — nesting and offcut accounting

An offcut is a cost line. As long as nobody counts it, it looks free

The challenge

Nesting by hand leaves more waste on the sheet than the owner will admit. The remnants pile up in the corner of the workshop and slowly turn into rubbish, even though half of them are perfectly good material for the next order. Nobody tracks them, because it is dull work and it does not show up in money until you count it.

The solution

The platform takes the order, nests the parts on the sheet accounting for rotation and kerf width, and records what is left as an asset. The nesting geometry was written for this job rather than wrapped around someone else's library, so it can be tuned to a specific machine and material.

  • ✓ An order reaches the floor already nested
  • ✓ What is left after the cut is recorded and offered to the next order
  • ✓ Kerf width and permitted part rotation are accounted for
  • ✓ Parts modelled in SolidWorks and Fusion 360, toolpaths in Fusion CAM

The unusual part

The same engineer draws the part in CAD, generates the toolpath in CAM and writes the web platform around both. Nothing has to be translated between disciplines — and translation is where most of the detail is lost.

Stack

Node.js · custom nesting geometry · SolidWorks · Fusion 360 · Fusion CAM · CNC

How it runs

An order arrives as a set of parts, not as a file somebody will open in CAD later. The platform nests those parts on the sheet itself, accounting for the rotations each part is allowed and for the kerf width, because the cut itself eats material that no drawing shows. What comes out is a nesting layout, and the order reaches the floor ready to be cut.

What is left after the cut does not disappear. The remnant is recorded as an asset with its own dimensions, and the next order asks about it first and about a fresh sheet only after that. It is this step that turns the corner of the workshop from a rubbish pile into a stock of material.

The nesting geometry was written for this job, so it can be tuned to a particular machine and material — a wrapper around somebody else's library gives no such freedom. Parts are modelled in SolidWorks and Fusion 360, toolpaths are generated in Fusion CAM, and the web platform is written in Node.js by the same engineer.

What to check before starting a similar project

  • • Whether the kerf width has been measured for every machine-and-material pair. A layout calculated with somebody else's figure looks tighter on screen than it turns out on the sheet.
  • • Which rotations a part is allowed. The grain direction or a pattern on the material can forbid turning a part, and the nesting has to know that before the cut, not after it.
  • • Where the offcuts physically live and how they are found. Tracking a remnant only works if the piece on the record can actually be pulled off the shelf.
  • • Who enters the order. If parts are redrawn by hand on the way in, the error appears earlier than any nesting algorithm can reach.
  • • What counts as a usable remnant. The threshold below which a piece goes to scrap is the owner's decision, and it has to be written down rather than left to the eye.

FAQ

Why not use an off-the-shelf nesting module?

Because the geometry is exactly the part that has to change. A ready-made module keeps its own logic fixed, while different machines and materials call for different rules on rotation and kerf. Geometry written in-house can be tuned to the workshop it serves.

What does tracking offcuts add if they are lying in the workshop anyway?

Visibility. As long as nobody counts a remnant it looks free, and it slowly turns into rubbish. Once it has dimensions and a place in the system, the next order can be laid onto it instead of onto a fresh sheet.

Who did the CAD, the CAM and the platform?

The same engineer. The part is drawn in SolidWorks or Fusion 360, the toolpath is generated in Fusion CAM, and the platform is written around both. It is an unusual combination, and it removes the most expensive part of the work: translating between disciplines.