The twenty-four-minute job that took two hours
Take the numbers this page opens with. A nested sheet with 1,450 inches of toolpath, cut three quarters of an inch deep in quarter-inch passes at 180 inches a minute, is 24 minutes and 10 seconds in the cut. That is the number CAM reports, and it is correct. Then the head spends 40 seconds on rapids, two tool changes take six minutes, and getting the sheet on and off takes four. The sheet is 34 minutes and 50 seconds. Three of them plus twenty minutes of setup is two hours and four minutes.
Nothing in that gap is waste. It is the ordinary structure of a routing job, and the reason quoting from the CAM figure alone runs short by a factor that grows as the parts get smaller and the sheets get fewer.
Only one term multiplies by the pass count
The depth passes multiply the cutting time and nothing else. Rapids do not repeat per pass in most posts, tool changes certainly do not, and the sheet gets loaded once whether you take it in one pass or six. So going from two passes to three on a job that is already half setup adds far less than fifty percent to the day, while doing the same thing on a deep, single-sheet, one-tool job adds almost exactly fifty percent.
That is worth knowing before you decide where to spend effort. On short runs of small parts, the lever is the handling and the tool changes. On long production runs of deep cuts, the lever is the depth per pass and the feed rate, and the setup has already amortised itself into nothing.
The pass count always rounds up
Three quarters of an inch at a quarter inch a pass is exactly three. Three quarters at 0.26 is 2.88, which is three passes, and the third one takes off 0.23 of an inch. Push it to 0.38 a pass and you get two passes with a second one that removes 0.37. But at 0.4 a pass you get two passes where the second removes only 0.35 — the arithmetic is 1.875, and the fractional part is what the last pass gets. A depth per pass just under a clean divisor buys you a whole extra trip round the path for a skim that removes almost nothing.
Rapids are not free and are not fast
The rapid rate on the spec sheet is a top speed reached on a long move. Between two small parts nested two inches apart, the machine accelerates, decelerates and never gets near it. Entering the machine top rapid here will understate the repositioning time on a busy nest, sometimes badly. If you have never measured it, run a nest with the spindle off and time it.
Where this sits
The spindle speed and chipload side of the same cut is on the machining speeds and feeds calculator, and the safe-rpm question for a large-diameter bit is on the router bit speed calculator. What the depth passes do to the tooling bill is on the bit life and cost per part calculator, which uses the same pass count. If you are laying the parts out rather than cutting them, the sheet nesting yield calculator and the cut list optimizer come first. A laser job has a different time structure and is on the laser cutting time and cost calculator, and the extraction the router needs is on the dust collection calculator.
Questions people ask
Where do I get the toolpath length?
Most CAM packages print it in the operation summary or the setup sheet, sometimes as "cut distance" or "feed distance". If yours does not, the crude version works surprisingly well: add up the cut perimeter of every part in the nest, add the interior holes, and add the lead-ins. What you must not do is use the outline of the sheet, or the sum of the part outlines when the parts share cut lines. If your machine logs run time, the honest route is to run one sheet, time it, and back the effective feed out of the result — that folds in the corner slowdowns this page cannot see.
Should the rapid distance include the lifts?
The Z lifts and plunges are usually a small fraction of the XY repositioning distance, but on a nest of many small parts they stop being small. Fifty parts means fifty lifts and fifty plunges, and at a plunge rate well below the cutting feed that adds up. If your post reports a rapid distance, use it and it will already include them. If you are estimating, add the number of separate cut regions times roughly twice your clearance height to the XY figure, and remember that plunges happen at plunge rate rather than rapid rate, so the result will run a little optimistic.
Why is the setup only counted once?
Because that is what setup means here: the work you do to get the first sheet cutting and never do again for this job. Posting the file, squaring a fence, finding the right collet, cutting a test slot. The work you repeat for every sheet belongs in the load and unload figure instead. If a job genuinely re-fixtures between sheets — different hold-down for a second operation, say — then that is per-sheet work and should move into the load line, where it gets multiplied.
My real jobs take longer than this. What am I missing?
Usually one of three things. The feed rate entered is the programmed feed rather than the achieved feed, and a nest full of small arcs never reaches the programmed number. Or the rapid rate is the machine top speed rather than what short moves actually manage. Or the load and unload figure only counts the sheet handling and not the vacuuming, the labelling, the sorting into stacks and the walk to the rack. The fix is the same in all three cases: time one real sheet end to end and adjust whichever figure is furthest from what you observed.
Does this work for a mill rather than a router?
The arithmetic does, because path length divided by feed rate does not know what kind of machine it is on. What changes is which terms dominate. Milling metal runs at feeds an order of magnitude below sheet routing, so the cutting term swamps everything and the setup share goes to almost nothing on any job of length. Milling also tends to involve far more tool changes per part, so the change count and the time per change matter more than they do on a router where one or two tools cut the whole sheet.