PCB Vias

It's a pain to solder vias with wire from one side of a PCB to the other, and if the vias are under an SMD IC then it makes soldering the IC tricky.

Different methods I have found or dreamt up are maybe possible:

1. Put wire through and solder it. This is easy and thevobvious way but has problems outlined above.

2. Use proprietary conductive epoxy paste in the hole. There's a commercial system that does this and it seems to work. It looks a bit messy and I'm not sure of the cost at the moment.

3. Put wire in the hole (wire that is thick enough to not fall out) and then cut it to a small distance from the surfaces. Then use a press to compress the wires and form a small rivet with very flat heads. This should then fit under the ICs. I'm not sure how well the rivet would handle age and oxidantion of the copper. Would it have conduction problems?

4. Use PCB pins. these are tapered pins that you put in the hole and snap off then solder on both sides. Problems are that they protrude above the surface like wire. they also don't seem to be made any more.

5. Use test pins. Similar to 4., but they are still made and sold.

6. Use small 0.8mm copper rivets.

7. Use a 3D printer to fill the via holes by printing conductive copper into the hole. I like this idea as it means that the hole drilling gcode or gerber file can be used to position the 3D printer and all the holes can be filled automatically. Problems include how well the plastic will fill the holes and how well the conduction will work. Conductive plastics for 3D printing are also not that conductive and it looks like a via would end up as around 50-150R, which is maybe usable, but care would be needed when placing them in a circuit.

I'm still investigating...

More PCB Milling

I'm experimenting with milling more detailed and double sided PCBs. So far I have tried a smaller cutter (D bit) that was 0.1mm, but I suspect turned out to be a bit larger than that. The fine tracks have disappeared in a few places so it's not usable, but overall it's encouraging.

The four large holes are my attempt at registration pin holes, which failed as the PCB software auto centres the artwork so when I milled the traces they weren't registered with the pin holes properly. I need to experiment with methods of creating these holes.

The second attempt at the double sided OLED watch PCB was more successful. I mistakenly drilled the via holes from the wrong side, I hadn't remembered that the holes are drilled from the bottom of the board, so I milled the top side tracks then drilled holes. Of course they were wrong and this PCB has an extra set of via holes. It is probably just about usable, but I am going to make a new board more suited to milling and also add some of the circuit changes as well.

This second PCB looks a lot better. I cleaned the copper up using some fine wet and dry paper as there's a lot of rough burrs as it comes off the milling machine.
The blue PCB is the one I had made in China, for comparison.

The registration using pins worked well, the program I wrote to generate the pin holes now uses the board outline, which has to be the correct size.

Collet Holder

This is the sort of thing that the CNC router is great for. I wanted a simple way to hold the collets that came with the CNC spindle, there's 13 of them and it's easier to use them if they are stored in order. So I made this holder:

Each collet is in a hole that is made up from four holes of different radii and depths so the collet is held snugly.

Making this without a CNC machine would be a trial indeed and simply not worth it. With a CNC it's trivial, well almost.

OLED Watch V2

The second version of the OLED watch PCBs turned up yesterday, so this morning I assembled a board and programmed the firmware. The boards are in a nice blue colour this time:

The assembled board is quite neat:

The firmware ran first time, as I had altered the breadboard version to match the layout and adjusted the firmware. Let's see how well it works...

3040 CNC Demo

The Chilipeppr software that you can use to drive the TinyG controller comes with a demo file. Once I'd calibrated the CNC so it moved 1mm for every 1mm it was told to (instead of 4mm for every 1mm it was told to), I had a go at carving the demo. Before the recalibration I had terrible problems getting the machine to do my bidding. I don't know why I didn't realise that I needed to calibrate it, it seems silly now that I didn't. Why I thought the two bits of kit (TinyG and the router) would have worked together out of the box(es) I don't know. Anyway, here's a video of the machine carving part of the demo logo:

I've just used some masking tape to hold the workpiece down, it's fine for a quick test.

Bandsaw

The chair project is moving on a bit: I'm going to make a full size model of the chair design using scaffolding planks. These will be routed to shape using templates, the shapes roughly cut before routing. I tried using the Japanese saw and the electric jigsaw, but they both had trouble cutting vertical to the line. So, finally, I have bought a bandsaw. I know, it should be the first tool you buy for the workshop, not the last, but I buy tools as I need them and I never really needed it up until now.  I decided to buy the Axminster BS11-INV, as it has an inverter ( and hence speed control) on the motor, which allows the blade speed to be slowed to the extent that you can cut metal with an appropriate blade. 10mm thick steel apparently.

I've built a base for the saw to sit on, which is itself on a wheeled base so I can move the saw around the workshop. It took three hours one morning to set it up, quite a bit of tweaking here and there was necessary.

Variable speed is nice as you can use the best speed for the job at hand. It looks like it will be a useful tool.

It looks like I won't be able to use scaffolding planks in the final design, the supply is too tricky. I'll have to hunt around for another supply of wood, and that will hopefully let me get something closer in size to the parts so I won't have to remove too much material on the thicknesser.