T
tscircuit•2mo ago
Ray Kholodovsky

Ray Kholodovsky - warning messages from updatin...

warning messages from updating to latest. not sure if related to, but see how footprint pads on bottom look a different way, while 3d model is rotated some other way.
No description
No description
No description
32 Replies
Ray Kholodovsky
Ray KholodovskyOP•2mo ago
looks like plated holes are completely not rendering in all my tsx 3d models
Ray Kholodovsky
Ray KholodovskyOP•2mo ago
plated holes broken all around
No description
Seve
Seve•2mo ago
CC @Techmannih yea we were literally discussing this the other day plated holes not rendering in 3d viewer is very odd, CC @Abse maybe @Ray Kholodovsky can you right click and select "Jscad Renderer", I'm curious if it's manifold-specific thanks for the report- this is high priority
Ray Kholodovsky
Ray KholodovskyOP•2mo ago
Yes, they render with jscad selected
No description
Seve
Seve•2mo ago
kk it's a manifold issue (CC @Abse or @shibo )
Ray Kholodovsky
Ray KholodovskyOP•2mo ago
yes please, I made the mistake of updating to try to get any latest features before showing updates to team. any way to revert back to older version for time being?
Seve
Seve•2mo ago
yes absolutely, just do bun add tscircuit@1.2.3 or whatever your previous version is hopefully you have a version control record of it let me know if you still see issues, it automatically loads some features from latest (but not related to the viewer)
Abse
Abse•2mo ago
just a normal plated hole will do to repro ?
Ray Kholodovsky
Ray KholodovskyOP•2mo ago
yes, on .798 now, was possibly on .766 before
Seve
Seve•2mo ago
@Ray Kholodovsky yes if possible upload circuit json
Ray Kholodovsky
Ray KholodovskyOP•2mo ago
Abse
Abse•2mo ago
thank you
Ray Kholodovsky
Ray KholodovskyOP•2mo ago
wait am I gonna miss any major feature rendering things if I stay on jscad renderer for now?
Abse
Abse•2mo ago
no both should jscad and manifold should render the same things
Ray Kholodovsky
Ray KholodovskyOP•2mo ago
ok sweet so staying on jscad for now here is what the "component on bottom of pcb rotation is wrong relative to pads rotation" looks like now. 
    <resistor
        name="R1"
        //footprint="kicad:Resistor_SMD/R_0402_1005Metric"
        footprint="0402"
        resistance={resistance ?? "-1"}
        supplierPartNumbers={{
            lcsc: lcsc ? [lcsc] : ["-1"]
        }}
        pcbX="0"
        pcbY="0"
        layer="bottom"
        pcbRotation="110deg"
    />
    <resistor
        name="R1"
        //footprint="kicad:Resistor_SMD/R_0402_1005Metric"
        footprint="0402"
        resistance={resistance ?? "-1"}
        supplierPartNumbers={{
            lcsc: lcsc ? [lcsc] : ["-1"]
        }}
        pcbX="0"
        pcbY="0"
        layer="bottom"
        pcbRotation="110deg"
    />
No description
No description
Abse
Abse•2mo ago
CC @Techmannih
Seve
Seve•2mo ago
there's no rigor around CCW/CW rotation for bottom side of board right now
Ray Kholodovsky
Ray KholodovskyOP•2mo ago
I was just wondering whether "110 deg" in kicad means the same direction as it does in tsci
Seve
Seve•2mo ago
yea it's ambiguous for the bottom layer because it depends on the position of the viewer i.e. if the person is LOOKING at the bottom of the board, the CCW direction is different than if they're seeing through the board i think we're assuming that the viewer is always viewing from the top side, but this distinction gets lost especially as people start trying wacky math @Techmannih we cannot fuck this up
Seve
Seve•2mo ago
trivial repro i think
No description
No description
Seve
Seve•2mo ago
it also looks like the behavior with regard to the pads is correct, which is very good
Ray Kholodovsky
Ray KholodovskyOP•2mo ago
yeah... but the actual rotation needs to be absolute to something, and this should all be generated THEN the user can orbit the 3d model however and it doesn't change. Can you just replicate exactly however kicad does all this? cuz I'm already used to it and haven't noticed any major issues/ objections to it (I don't think). -rotate a component on the top of the board 30 degrees and it rotates about the center in this direction. -flip it to the bottom of the board and it is oriented in this direction. -etc This is all based on 2d views looking from the top, I think. The 3d view should only ever exactly replicate what the 2d pcb view already shows, IMO
Seve
Seve•2mo ago
yea this is exactly what we should be doing, always top-relative CCW
Seve
Seve•2mo ago
confirmed bug in GLTF viewer as well
No description
Seve
Seve•2mo ago
fwiw @Techmannih you must have implemented the pad rotation correctly, because AFAIK that is all correct 👍
Abse
Abse•2mo ago
patch on it's way
No description
Techmannih
Techmannih•2mo ago
There’s a CAD model rotation issue on both layers, as we discussed earlier. The rotation currently supports only 90°, 180°, 270°, and 360° I have added repro in gltf.
Seve
Seve•2mo ago
here's my repro on tscircuit.com fwiw
Seve
Seve•2mo ago
Abse
Abse•2mo ago
@Seve https://github.com/tscircuit/3d-viewer/pull/530
GitHub
fix: render circular plated holes correctly for manifold viewer by ...
Circular plated holes weren’t rendering because the board drill used the outer diameter, removing the copper ring during clipping. This change switches the board drill to use hole_diameter , ensuri...
Abse
Abse•2mo ago
Ray Kholodovsky
Ray KholodovskyOP•2mo ago
cad model rotation seems correct now
No description

Did you find this page helpful?