Skip to content

feat(ark-flow): add Rev 2 case in STEP with board-slot clearance fix - #7

Open
dakejahl wants to merge 1 commit into
mainfrom
feat/ark-flow-case-rev-2-step
Open

feat(ark-flow): add Rev 2 case in STEP with board-slot clearance fix#7
dakejahl wants to merge 1 commit into
mainfrom
feat/ark-flow-case-rev-2-step

Conversation

@dakejahl

Copy link
Copy Markdown
Contributor

Summary

Adds the ARK Flow case as STEP (with matching STL) at Rev 2, based on revised files a customer sent us, with one interference defect corrected before publishing. Rev 1 STLs are kept until a test print validates Rev 2.

Problem

The case/ directories in this repo only ship STL, which is a triangle mesh — customers can't mate to it, dimension it, or modify it without a mesh-to-solid conversion. That is what prompted the customer to send STEP files back to us.

Comparing his files against Rev 1 showed they are not just a format conversion — the geometry changed. The significant change is the board slot: Rev 1 seats the board on a shelf 0.8000 mm below the parting face against a top-half ledge 0.7218 mm above it, giving a 1.5218 mm slot. That is exactly the modelled PCB thickness of ARK_Flow_Rev_2_3D_Model.step, to four decimals — zero clearance, which fits the CAD model but not a real board with normal thickness tolerance. His revision opens it to 1.0000 + 0.7000 = 1.7000 mm.

The revision also introduced a defect. The connector cut-out back wall in the bottom half moved inward from y=4.8531 to y=4.6190. Rev 1 cleared the bottom-side component at x[10.94, 18.94] by exactly 0.2000 mm; the revision spent that entire margin and went 0.034 mm past it, for 0.092 mm³ of interference. Every other change in the revision moves outward to add clearance, so this reads as an oversight rather than intent.

Solution

Restored that one wall to its Rev 1 position, leaving the rest of the revision intact. The cut removes 1.5735 mm³ and the part remains a single valid solid with an unchanged outer envelope.

Verified with boolean interference using ark-hardware-jigs:

Check Result
Bottom + top vs ARK_Flow_Rev_2_3D_Model.step, board seated 0.000000 mm³
Bottom vs ARK_Flow_MR_Rev_1.0.step 0.000000 mm³
Top half against bottom half (halves mate) 0.000000 mm³
Board slot vs 1.5218 mm PCB +0.1782 mm clearance

Both halves are written in the same coordinate frame as ARK_Flow/model/, matching the existing Rev 1 STLs, so the case and PCBA open already assembled.

Two things to confirm before this is merged:

  • Revision number. Rev 2 is our label. If the customer's files correspond to an official revision, or if this should instead land upstream in ARK_Flow and be mirrored here per manifest.tsv, the naming and the manifest rows should follow that.
  • ARK Flow MR. ARK_Flow_MR/case/ currently holds byte-identical STLs to ARK_Flow/case/. Rev 2 clears the MR board too, but it is not added there in this PR — leaving MR on Rev 1 is an inconsistency worth a decision.

Adds the ARK Flow case as STEP (plus matching STL), based on revised files
supplied by a customer, with one interference defect corrected.

The Rev 1 case slot is 1.5218 mm, exactly the modelled PCB thickness, so it
has zero clearance for real board tolerance. Rev 2 opens it to 1.7000 mm.

As supplied, the revised bottom case moved the connector cut-out back wall
from y=4.8531 to y=4.6190, which clipped a bottom-side component by
0.092 mm^3. Restored to its Rev 1 position, giving 0.20 mm clearance.

Verified with boolean interference against ARK_Flow_Rev_2_3D_Model.step and
ARK_Flow_MR_Rev_1.0.step: 0.000000 mm^3 for both halves with the board
seated, and the two halves mate at 0.000000 mm^3.
@dakejahl
dakejahl requested a review from NickJ98 August 15, 2026 04:28
@dakejahl

Copy link
Copy Markdown
Contributor Author

@NickJ98 let's print this off and verify it. It would also be good to find the .step files for the other cases as .step is strictly better than .stl

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant