Unable to Boolean join the original out of 5 cloned parts

c_piasec

Document is https://cad.onshape.com/documents/1607259a62c3cea2efc84d86/w/4b407f521e0613e8b247e888/e/fe335b89afc0b2e0ddaf6129

Referring to parts 7 and 8, I'm having issues boolean joining the last part 7 to the rest of the frame 8. Part 7 has been copied 4 times through transform, all rotated around the same axis, and all 4 copies have boolean joined successfully to part 8, so struggling to understand why the part they've been copied from won't? Would have expected either all of them to join, or none to join.

nick_papageorge073

Part 7 is touching the main part on a single point somewhere, resulting in "non-manifold geometry". If you move-face all of its outer side surfaces a small amount, then the union works.

glen_dewsbury

@c_piasec
I'm a little curious about some of the methods used in your part studio. Is this intended as an assembly or a casting? Why transform and copy when a
circular array will do? So many transforms instead of drawing in place. When adding and extrude there is an option to add to connected parts instead of a separate Boolean. Why not start your basic pattern at the origin since it will make locating other features simpler to locate?
There are a number of things that look puzzling to me that has increased the length of the feature list that seem unnecessary and a bit time intensive.
If this is intended to be an assembly, only one of the arms needs to be drawn then arrayed in the assembly.
Kinda looks like you may be applying techniques learned from an other program that aren't very effective in OS.

Have you looked at the learning center? Going through to introduction courses might be a big time saver in the long run.
https://learn.onshape.com/catalog?labels=%5B%22Self-Paced%20Courses%22%5D&values=%5B%22Onshape%20Fundamentals%22%5D
https://cad.onshape.com/documents/9e8fd4216c662978ed342e4a/w/3b7efba1c8a1bac1b60ea9bf/e/fb5c476e33e76eb7ec581a12

c_piasec

The document is a complete Frankenstein cut and shut, a second version that I had indeed done over the top of the first version, hence the transforms in moving parts around, and legacy design choices. Seeing as it's 99% there for my purposes (I could just use it in parts without any substantial drawback) I was hoping there was something small noticeable I was missing to just join the last piece. Appreciate if it's not really salvageable.
I'm a new user of OS, and due to what will be very infrequent use the time I've put into get this far is about as much as I can afford. Will look into those links. 

nick_papageorge073

Part 7 is touching the main part on a single point somewhere, resulting in "non-manifold geometry". If you move-face all of its outer side surfaces a small amount, then the union works.

c_piasec

Perfect, thank you very much!