Hopping one week ahead

Due to the arrangements in our local class, I will do this week as if it was the 8th week in FabAcademy, and return to CNC milling in a week or two.

PCB production process

I downloaded this linetest for traces (image below), opened it with Inkscape, scaled the traces to match millimeters (with a 0.5mm wide yellow rectangle on top of the 0.02inch trace), exported it as an SVG, and opened it in KiCad with Import -> Graphics. For some reason the import didn't include the numbers, so I added them to User.1 layer. That probably won't work as also the front copper layer is going to be removed under them. So I returned to Inkscape to do some adjustments and spent too much time trying to figure out why the numbers disappear. So I let it be.

For creating tool path from Kicad to CopperCAM I rely on this tutorial at Aalto wiki. First I ran DRC to make sure there are no violations. Then created the gerber file.

Roland SRM-20

For the use of this machine, I relied also on another the documentation at wiki.aalto.fi. We have two types of tools for milling on this machine:

• Tool 1 is 60 degree ~0.4mm V-shaped bit and it's used for engraving
• Tool 2 is 0.8mm router for cutting and drilling

Tool 1 is used first for engraving the routes because it's easier to keep the board in place when not yet cut smaller.

I started with milling my own PCB design for the reaction game. So I opened CopperCam on the FabLab computer and began preparing the files for milling. I opened the boards F.Cu gerber-file (File -> Open -> New circuit) and clicked yes to If the highlighted line is the contour. I did the same again for the Edge.Cuts-layer with File -> Open.. -> Additional layer and clicked yes for contour.

The layers weren't aligned so I followedthe instructions to align them. It didn't work because for some reason I couldn't select the edge.cuts layer. I also realized I hadn't placed the via in bottom right corner in Edge.cuts-layer so I returned to do it on KiCad (on another computer) and exported the gerbers again (via OneDrive). It didn't go as planned either, because the via wasn't visible even if I did it on edge.cuts-layer. I tried to make a simple circle instead, but for some reason KiCad only drew rectangles even though the draw circles was selected. I quit KiCad to reopen it and it wasn't to be found in the applications. I searched for it and eventually found it and it still didn't draw circles.

At this point I had already used 1.5h trying to figure some basic things out and I hadn't even got started with the milling.

Well instead of circles, I made two half circle arcs and connected them and this way I got the hole visible in gerber view. The hole is for possibly attaching the servo motor to the board as for the possiblity of using the board as a console (I'm not sure if it's going to work). Then I uploaded the new files again to OneDrive to move them to the milling computer, and now OneDrive gives and "Unknown Render Failure" out of the blue. This seems to be my workflow all the time: struggling with random obstacles that are not about the actual work that I'm trying to do. So I emailed the files to myself and opened them in CopperCam and now the edge.cuts-layer comes up only as a dim rectangle in bottom corner.

I tried sending it again and suddenly it appear as it's supposed to. But I still can't align it with the front layer because I can't select the edge.cuts area at all, it won't respond to any clicks. Maybe I have done something wrong / not done something right in exporting from KiCad.

I turned to Gemini to ask how to align the layers. It can't seem to help with the issue I'm having on CopperCam but recommends aligning them already in KiCad so I try that. I try sending all the my KiCad files to the lab's drilling computer and open them on KiCad. It appears that KiCad is not configured on the drill computer, so I end up doing it on my computer anyways.

Around a few more corners, I select Place -> Drill/Place File Origin and set the origin in the bottom left corner of the board. Exporting to gerbers, I selected Use drill/place file origin. Also Gemini instructed me to generate Drill files (which I hadn't done previously) and select Origin: Drill/Place file origin.

Again I emailed these to the lab's drilling computer, opened them on CopperCam, and now I had them already aligned. Finally the first step in the process and it's already almost lunch time. But I managed to manage my frustration.

Now trying to open the drill files, CopperCam places holes in somewhat random place. Apparently I need the holes for the XIAO board to be attached to the board. After a few tries, I realized that by toggling the Units to Imperial then the holes where in their right place. The NPTH.drl-file is not opening at all, but it seems I don't need it.

I calibered the board to be 1.56mm so I set the dimension to 1.6mm and margin to 1mm. I also checked that the tools in CopperCam are the same that are in use (match the image at aalto wiki). At Parameters -> Selected Tools, I check that the Active tools match the instructions.

Next I clicked Calculate contours and now it shows that 60DEG V bit is used for isolation milling, and I got confirmed by lab assistant (and Gemini) that it's OK. I also set number of successive contours to 2 and extra contours around pads to 0 (to make the contours shallower than the previous ones printed here.) Also the other two options below are as supposed to.

I also heard that the engravong depth should be shallower than in it is at the moment, so I go to Tool Library and for V-bit I set Maximum depth per pass to 0.07mm. To make it update, I quess I have to start from beginning, which is quick because I have done it at least 5 times already. Well it didn't, so I set the depth in Selected tools instead.

Then I'm ready to export toolpaths. I click Mill icon and check the sections, leave Mirror X unchecked and have the XY-zero point to South-West corner. Click OK and check that this resulted in two files for both engraving and cutting/drilling. Getting here took me almost 4 hours and I need to leave the lab now, so I continue the actual milling tomorrow.

To be continued...

Bungard

We also have this PCB milling machine for more advanced use with also more advanced configurations. Unfortunately I had to leave the class and couldn't participate to the end, so the process documentation for using this machine is coming somewhere in the future.

Making my own PCB