The assignment of this week has two parts: a group assignment to characterize the design rules for our PCB production process and a second one, individual, to make an in-circuit programmer by milling and stuffing the PCB and test it.
To do the group assignment we milled a trace width test in two different machines, and with different parameters. We used the png file we were provided in the fabacademy page of this week.
We used the site http://modsproject.org to generate the files for the machines.
The machines we used are the Roland SRM-20 and a cheap generic one which is new.
In the following screenshots I'm showing the process to generate the files in modsproject.
First, we choose the machine we are going to use.
Then we have to select the png file. In the first place, we generate the file for to mill the traces of the circuits
We choose the 1/64 endmill for the traces. The depth is the default one as well as the offset number (4 for the traces). The stepover is 0.5. We tested both climb and conventional directions, with the results are showed later.
Then we have to set the speed, the origin, jog and home (where the machine goes after finishing the job). We chose 3mm/s for the traces, with the origin in (0,0,0), 12mm for the jog and the final home in (0,0,20)
After connecting the otput to the "save file" node, we can calculate
The result is a file with the path that the machine should follow
To generate the file for cutting the border, we have to upload the png of the border of the board.
The process is the same but we ned to adjust some parameters. In this case we choose the 1/32 endmill, and we set the depth in 1.75mm because is the thickness of the board. The offset number is 1 in this case. We tested also both climb and conventional directions.
Regarding speed, we setted 1.25 mm/s
Once we have both files, we can send them to the machine in order to mill the boards. First of all we have to choose a big enough board to fit our design and stick it to the plate of the machine. We used double sided tape to do that.
Then we have to put the endmill in the head of the machine. In the first place we use the 1/64 endmill.
After setting the "X,Y,Z" position in the software, we can start milling.
Once the traces are done, we can cut the border. We have to change the endmill with the 1/32 one and adjust the "Z" again.
After the work is completed we can take the board out of the machine
We tested both machines in conventional and climb directions. The results were quite similar in both machines (they were better in the Roland, but regarding the price difference the results were not that far away). Regarding the direction of cutting, in conventional the results were finer, but the traces were thiner. It seems that in climb, the machine takes less material away when milling, although the borders of the traces are not that smooth.
For the individual assignment I chose to make one FTDI and one UPDI boards
I decided to use the Roland SRM-20 because the results were better.
First of all I generated the files in modsproject, with the following parameters:
Then I milled them in the Roland. The result was good so I just needed to do it once.
After cutting both boards, I soldered all the components in place, following the board schemes.
The tools I used to solder were:
And the proccedure I followed:
The FTDI:
The UPDI:
The final step was to plug the FTDI to the computer to see if it detects it. After some checkings with the multimiter I connected to the computer and it was detected as FT230X Basic UART.