Part rotation in circuit studio

[chende...@gmail.com]

Hi friends,

I need a bit of clarification regarding rotation processing, because it looks weird and I do not understand where the cause is. I attached some handwritten notes. I am working with circuit studio only, so I do not know how other CADs handle this issue.

In the attached picture there are parts rotation examples as they are handled by CS. I used for example a 2 pin component, with named pins 1 and 2. At each rotation angle, it is drawn how CS places it. The TOP column is top layer and the board is viewed normally, from the top. The BOT column is NOT how you see it in CS, because there is seen through the board, from TOP direction. BOT column is mirrored CS view, is how you would see the circuit board in your hand looking at the bottom side. On the notes are also written angles on the test feeder I use, is a pneumatic CL, 0603 part.

When I import data on openpnp, all values for rotation, x, y, layer are just copied into job tab/placements. When I place components on top, all is ok, but when I place on bottom, rotation is wrong. Just rotation, center x/y are ok. According to my analysis, a component picked with 0 degrees from feeder should be placed with 0 degrees when mounted on top and with 180 rotation when mounted on bottom, to preserve pin direction. In camera view, that circle which indicates 0 degree rotation stays the same, regardless if the component is 0 on top or 0 on bottom.

I do not get it. Is it something I am doing wrong? CS does not have any option regarding coordinate generation mirroring. It only asks if I need text/csv file and inch/meter coordinate type, nothing else. I have many CS designs sent for manufacturing at factories and nobody complained about rotation. Also, in the openpnp named CSV importer there are no options to mirror anything. Help? Thanks.

[Michael Anton]

In P-CAD, which is what I still use, the parts flipped to the bottom side are just mirrored in X when viewed through the board.  So, a part on the top looks just like you have drawn, and a part on the bottom looks like you have drawn when viewed through the board from the top side.  As a result, the straight export of the angles just works as is.  In your case, it looks like the angles on the bottom are rotated clockwise, rather than counter clockwise, so you would have to do a post processing step to fix these.

I haven't had the problem you describe when importing jobs from Altium, so I'm not sure why Circuit Studio would behave so differently.

[Richard Sim]

Altium also shows both top and bottom layers as you have drawn TOP and BOT respectively.

You are flipping the top component to the bottom of the board, not creating a new "bottom" component with pads on the bottom layer, correct? I cannot see CS behaving differently than Altium here, as it's based on the same code (I even double checked AD19, as CS was last updated in 2019).

--
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/b34df384-a67a-4fd7-bf45-c9af15abc313n%40googlegroups.com.

[chende...@gmail.com]

Yes, this is my usual workflow, with flipped components. After schematic is done, I create PCB file , import components from schematic and CS places them all on top. After arrangement, I select those required to be on bottom and I move them all changing 'layer' in their properties. This automatically changes the rotation component from 0 to 180 for ex. But funny thing is that in my case the rotation direction changes also. Components on top rotated from 0(in feeder) to 45(on board) must be rotated CCW, but when on bottom, they must be rotated CW. This is how I discovered it, on a test board. 0 and 90 components were placed ok-ish (those with 0 were actually placed with 180, but with passives does not matter and is not obvious on board) but those at 45 were placed rotated with 90.

Apparently only a post processing will solve it, as you indicated above. I think bottomangle=180-topangle is ok