Closed feature request

Pour fragment count per net with the pad to rotate, as a review finding the pours gate can push on

Drew Owens · ·closed by John Lauer

What Quilter has

Power Plane Pour Planning (Aug 31 2026): inner-plane pours shaped to cover as many same-net vias as possible, and placement that orients same-net pins toward each other so a pour stays continuous. Their Dec 2025 pour generator already placed for pour continuity.

What Fields has today

The Ground view finds floating islands on ground; the pours gate in AI Flow checks area, connection style and vias. Neither says, per net, how many pieces a pour broke into or which pad to turn to make it one, so pour continuity is not an objective the loop can push.

Proposal

  1. Per loaded net, count the connected copper fragments on each layer after fill, and for a net with more than one, name the two nearest fragments and the pad or part whose rotation would join them (the solver already has the cell ownership and pad positions).
  2. Report it as a review finding with an action (rotate U3 90 degrees, or move C7 1.5 mm) in fixPlan, and expose fragments per net in fields.json.
  3. AI Flow's pours gate reads it, so a fragmented rail is a return to placement with a named move, the same way a hot tab is today.

Done when

A rail poured in two islands on the ESC is a finding naming the pad to rotate, and after the move the count is one.

Source: the Quilter assessment on the private page https://wiki.adom.inc/drew2/quilter-assessment (Oct 15 2026; Quilter facts are from their public changelog, https://www.quilter.ai/changelog). Filed by Drew's AI agent on Drew's behalf.

1 Reply

John Lauer ·

Built in adom-fields 0.7.20. fields.json fragments lists, for every poured net and layer, the pieces, their areas, how many the net's vias join through other layers, and the nearest gap. A split pour on a loaded or ground net is a pour-fragmented issue whose action names the pad in the gap (move it or rotate it 90°), or the net to reroute out of it. On the ESC, +VBAT's F.Cu pour is in 4 pieces, and the nearest gap (0.85 mm) has R14 pad 2 in it. Closing; reopen if it misses something.

John Lauer closed this ·

Log in to reply.