linux-card

It's been a year and a half since the last sharing
of the Linux card machine project. The reason for the long delay in updating is mainly because the project has not made any substantial progress. Although we have done six PCB fabrications and modified the PCB repeatedly, the entire process has not been working. So, I will continue to update. Of course, this is also because my consistent principle is that any original content I share must be verified by myself as much as possible.
Fortunately, all the problems were perfectly resolved within the past week. The board was successfully driven, u-boot compiled and ran successfully, and Linux ran successfully, so I started updating again. First, let me show you the picture of the successful boot:

u-boot successfully driven


, Linux successfully running,


front and back


of the board.

Today, I'll share my blood, sweat, and tears history of board making and designing. I made it 6 times, soldered it countless times, and stayed up countless nights... but it was worth it. The picture below shows the record of the six board making attempts, from top to bottom, from the first version to the sixth version. Although the sixth version still has some problems, it can run normally after jumpers and shorting. In
the first version
, I didn't even know what a package was. When designing the board, I involved all kinds of packages, including 0805, 0603, 0402, and even 01005 and 0201 packages. So the first board was basically impossible to solder. Of course, this is why the packages on the boards I have now are still a mess. After all, I can't waste the components I bought.
Second version... The third version
involved redrawing and fabricating the circuit board. After the second soldering, it was discovered that the chip pins were misaligned, causing a persistent short circuit. This led
to the third version. Another change between the second and third versions was the removal of the LED from the serial port. I actually referenced another expert's schematic, who connected an LED instead of a light-emitting diode, suggesting a diode should be used. I drew a 1N4148 switching diode. Regardless of its rationality, drawing something haphazardly is definitely inappropriate. This serves as a reminder to anyone copying circuits: always understand the underlying principles before copying, or you'll only end up hurting yourself.
The third
version had an incompatible SD card slot, so a new board was made, since JLCPCB boards were readily available for free.
The fourth and fifth versions
had the chip driving correctly, but the serial port would disconnect immediately upon pressing the reset switch. After comparison, it was discovered that the serial port connection was incorrect. I was using a CH340E chip, and RXD and TXD should correspond to the DTX and DRX of the F1C200S chip, respectively—reading corresponds to transmitting, and transmitting corresponds to reading. I hadn't known this before and drew it incorrectly, leading to this blunder.
The fifth version was then created. The fifth and
sixth versions were basically unchanged, except that because the serial port wasn't working, I referred to other people's schematics and added a 1uf filter capacitor to the reset pin.
I also added some voltage test points.
However, just when I thought everything was finally over and I was about to power on this little board, fate played another trick on me.