mirror of
https://github.com/ajisaacs/OpenNest.git
synced 2026-10-29 21:41:08 -04:00
ajandClaude Opus 5.5
78ca08efe1
The nest report PDF gains an estimated cut time: the job total on the summary, an Est. cut time column in the plates table, and per-plate and all-copies times on each plate page. Times come from the cut simulation with the settings last used in Calculate Cut Time, plus clamp removal. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
49 lines
3.1 KiB
Markdown
49 lines
3.1 KiB
Markdown
# Cut-time estimates
|
||
|
||
Calculate Cut Time, Plate > Simulate Cutting and the [nest report](nest-reports.md) share one
|
||
timing model (`OpenNest.Core/CNC/CutSimulation.cs`, `ContourTiming.cs`, `MotionPlanner.cs`),
|
||
and the last-used settings are shared through `SavedCutParameters`.
|
||
|
||
## Inputs
|
||
|
||
- **Outside feed, hole feed, rapid and pierce** from the dialog or simulator. Inch rates are per
|
||
minute; millimeter rates are per second (the existing `Timing.CalculateTime` convention).
|
||
- **Skip etch-only contours** (on by default) leaves out contours that only scribe, as the
|
||
CI Fiber post's `SkipScribe` does, and joins the travel around them.
|
||
- **Acceleration** from a machine in Tools > Machine Configuration (`MachineConfig.Motion`,
|
||
in the machine's units and converted to the nest's). The choice is remembered in the
|
||
`CutTimeMachine` user setting. Until any machine is chosen, the first machine is used.
|
||
"None (full speed)" gives the old constant-speed estimate.
|
||
- A machine with no saved motion values uses `MotionConfig.Placeholder`: about 0.5 G cutting,
|
||
1 G travel and a 0.001 in (0.025 mm) corner deviation. These are stand-ins, not measured
|
||
values, and the picker labels such machines "(placeholder)".
|
||
|
||
## Model
|
||
|
||
- Moves are walked in plate cutting order. Cut, Display, lead-in/out and Break moves cut; cutting
|
||
strategies emit contour moves on Display, and the posts cut them. Rapids and Tab spans are
|
||
head-up travel, and suppressed moves are left out. A cut that follows travel gets a pierce
|
||
dwell. Scribes are not pierced.
|
||
- Contours follow the CI Fiber post's interior/exterior (`L2`/`L4`) rule: a cut-off, or a
|
||
closed contour wound clockwise (sampled without its lead moves), takes the outside feed.
|
||
Everything else, including open pieces, takes the hole feed. The CI Fiber post writes no feeds,
|
||
so these are stand-ins for the controller's process tables until calibrated.
|
||
- Each cutting run (between a pierce and the next travel) starts and ends at rest. Travel moves
|
||
are planned separately and also go from rest to rest.
|
||
- Each move follows a trapezoidal profile: it accelerates, cruises at the feed and decelerates. A
|
||
move too short to reach the feed peaks where the two ramps meet.
|
||
- Corner speed uses the junction-deviation formula `v² = a·δ·sin(θ/2) / (1 − sin(θ/2))`. A zero
|
||
deviation stops at every corner. Collinear joints do not slow.
|
||
- Arcs are capped at the centripetal limit `√(a·r)`, using the cut acceleration.
|
||
- A backward pass and then a forward pass keep every speed change within the acceleration
|
||
(Grbl-style lookahead).
|
||
|
||
Not modeled: jerk (S-curves), controller-specific lookahead limits, feeds from controller
|
||
process tables beyond the outside/hole split, an etch feed (etch kept in the timing runs at the
|
||
outside feed), and power ramping. With real controller values, expect roughly 5–15% error. Calibrate by comparing
|
||
a known program against the machine's own run time.
|
||
|
||
The plate time is the simulation total plus clamp-removal minutes, times the plate quantity.
|
||
Distances and counts in the timing dialog still come from `Timing.GetTimingInfo`, which counts
|
||
only Cut-layer moves as cut distance.
|