Open bug report

Z-origin audit: 6 adom STEP models render sunk or half-sunk in KiCad (+2 machine contacts need a call)

Colby Knox · 1mo ago

Z-origin audit of all hardware STEP models: the adom-owned defects (anchor issue)

Swept every hardware page's STEP (945 pages; Z measured from VERTEX_POINT geometry only, so construction points do not skew it). KiCad's convention, and what the wiki's "Open in Footprint Editor" flow assumes when it attaches a page's own .step to the installed footprint, is origin on the board surface with the body extending +Z. The adom-owned models that violate it:

Fully below the origin (renders sunk INTO the board)

page vertex Z range note
adom/ads1220irvar -1.0 .. 0 the original report: VQFN body entirely below Z0
adom/hfj11-1g41e-l12rl -21.59 .. 0 RJ45 jack exported with its TOP at the origin

Centered origin (renders half-sunk)

page vertex Z range
adom/bma580-561000 -0.6 .. +0.6
adom/act1210l-201-2p-tl00 -1.8 .. +1.8
adom/tpd2eusb30drtr -0.5 .. +0.5
adom/sot6pin -3.2 .. +3.2

Needs a human call, not obviously wrong

adom/machinecontactlarge (-3.75 .. 0) and adom/machinecontactmedium (-1.65 .. 0) are 100% below the origin. If contacts are meant to hang under the board that is by design; if they mount on top, the export is inverted like the RJ45.

Explicitly fine, no action

The molecule, scaffold, and machine-pin pages measure 55-73% below Z0 because their pins extend below the board plane to plug a carrier: the below-Z depths match the pin lengths exactly. That is correct geometry, not a defect.

Fix for the defect rows: re-export with the origin at the bottom face (mounting plane), body up. Same-defect issues filed with the other owners: aravk/cl21a106koqnnng#1 (8 Samsung caps, one run), barrett-land/skrpace010#1, john/nrf54l15#1.

0 Replies

Log in to reply.