Vision Compositing - SOT428C

[Litterio Andrea Guainella]

Hi all,

I try to draw SOT428C (DPAK-3) package but the error: "Error: Invalid | No solution found. Cannot isolate corners with compositable X and Y symmetries. Check footprint." it prevents me from using the bottom vision.

Seems all steps for pin definition are ok but I am struggling for this.

Any suggest? I attach some details for more clarity (I am using CP40 NT4)

LAG

[Litterio Andrea Guainella]

I just solved issue rewrite with right dimensions the pads

LAG

[mark maker]

Hi LAG,

must have overlooked your email, sorry.

It was evident from the cross-hairs ruler that it was not symmetrical:

_Mark

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/435cdb55-d5c6-4476-83a2-df89ee8ded2cn%40googlegroups.com.

[Litterio Andrea Guainella]

Hi Mark,

thank you for the clarification. I completely overlooked that a component should be symmetrical in shape. I will keep in mind. :-)

LAG

[mark maker]

>  I completely overlooked that a component should be symmetrical in shape.

Not the whole part needs to be symmetrical, but some pairs of corners, either across X or Y or both, i.e., diagonal across the center.

It needs to determine X position, Y position and rotation, and can only do so, if there are some symmetrical pairs of corners available.

Examples:

• Square parts need only two diagonal corners to provide all three.
• Rectangular parts can again use one diagonal to nail X, Y, but need one extra corner to get the rotation.
• But it will also find symmetry in parts that are not entirely symmetrical.
  

Symmetry in the pairs is required, because the theoretical footprint specs never match the vision detected features 100%. Pads will always appear larger or smaller, or closer to/farther apart from the center. Pads and pins are slightly beveled, the vision stages may erode away some of the contour, etc. These errors are unavoidable.

Only by assuming that the errors are the same on both sides, can we cancel them out. Hence they need to be symmetrical.

https://youtu.be/P-ZudS7QQeE

_Mark

To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/5e98bb54-bf5c-489c-b869-bbb1d540ed59n%40googlegroups.com.

[Litterio Andrea Guainella]

> But it will also find symmetry in parts that are not entirely symmetrical.

Ok... In fact I was wondering how things could work here in this point of the video with offset pads.

Now it's all clear. Thanks again for clarifying that

> Symmetry in the pairs is required, because the theoretical footprint specs never match the vision detected features 100%.

It is certainly a sad truth...

LAG