After quickly brushing up some basic concepts in electronics, our instructors proceeded to demonstrate the Roland Modela machine to us. It was a surreal experience to mill my own Printed Circuit Board (PCB) and solder the components on it. We referred to Brian's webpage 'Building the FabTinyISP' to make our programmer.
Group assignment:
1. Characterizing the specifications of our PCB production process
Individual assignment:
2. PCB fabrication
3. Soldering and assembling the PCB
4. Checking and debugging the PCB
5. Programming the PCB
Watch the video below for a quick walk-through..
The same process as described above was followed for our individual assignment to mill the PCB for the FabTinyISP, following Brian's tutorial.
Soldering seemed very difficult at the first go but got easier and was even fun, once I got the hang of it. The trick is to heat the board with the soldering rod and place the lead on it. As the lead melts, place the component in the required orientation with the tweezers. Use the soldering rod again to smoothen the soldering.
I used the multimeter to check that the circuit is connected and all the solder joints are in place. The black and red ends of the multimeter were placed at two different points in the circuit. If the circuit was complete, the values would change and a long beep would be heard. The video clip of this can be seen towards the end of the video at the top of the page.
My circuit seemed to be intact when checked with the multimeter. However, when the PCB was plugged into the PC, both the red and green LEDs lit up. This is a bug as only the red LED should light up now. The green LED should light up only when the programme is loaded into the PCB.
Vinod, our instructor, helped me by checking the circuit with the multimeter again. He noticed that the forward bias voltage value across the two diodes D1 and D2 was over 700V on my PCB whereas it was ~650V in the other functional PCBs.We tried changing the diodes - didn't work. Then Vinod removed the IC and noticed that one half of the IC (3 pins) had not been soldered completely. He soldered it back and then the problem was solved! I wondered then why the multimeter debugging had not revealed this issue. Vinod said that the pressure applied by the multimeter pin would have pressed the IC down and completed the circuit. When the multimeter was taken off, the circuit would have been incomplete. This was quite an eye-opener.
I connected my PCB to our instructor, Yadu's programmer, such that the same connecting wire connected the first pin of one board with the other (check the video at the top of this page). The programmer was inserted into my PC. I had downloaded 'Crosspack' on my Macbook. I also downloaded the firmware source code, as per the instructions on Brian's tutorial. I didn't have to edit 'Makefile' as it already had the line
PROGRAMMER ?= usbtiny