I sent this version to fab in November of 2023, but didn't finish bringing up major board components until recently.

Past Bringup

I've scattered board bring up progress in previous updates, but I'll repeat previous conclusions here:

• The USB micro connector I chose cannot be soldered to a board, so I have to change the footprint.
• The 24-pin FPC connector is too close to the board edge given the dimensions of the e-paper's ribbon cable, so I need to move it back ~2mm.
• The surface mount JST connector on the carrier board layout is so tight I need pliers to remove the battery. I need to replace it with a looser fighting connector.

Frontlight Dimming

I've been calling the light panel on the GDEY042T81-FT02 a "backlight", but physically, it is a front light. I complained about dimming it with the TPS61165, but it turns out all I had to do was increase the PWM frequency to within in the ICs documented specifications. I continued to see problems after increasing the frequency when connecting the Raspberry Pi GPIO to the piEreader HAT PCB with jumper wires, though. Once I directly put the piEreader HAT on a Raspberry Pi, I no longer saw dimming issues. Probably due to increased impedance from the jumper wires.

Battery Monitoring

I was able to read parameters from the BQ27441 IC when connected to the Sparkfun Battery Babysitter . I wasn't getting sensible readings from the piEreader carrier board because of a stupid schematic error.

You'll notice that the SDA and SCL pins on the BQ27441 aren't actually connected to the I2C pins of the CM4 module. They're just pulled up to 3v3. A stupid mistake. Again the code works on the Babysitter development board, but correcting battery status will require a board rev.

Luckily I placed status LEDs on the BQ24075 IC. Directly connecting the 5V and GND pads to a DC power supply resulted in expect behavior from the power good, battery present, and charging LEDs.

Next Steps

Previously I've talked about adding audio to the project. After doing some research, it looks like this will require fiddling with Linux configuration to make things work. I'm not against this, but I'm still not convinced a Linux board is the best platform for this project. The decision that forks the road to embedded Linux or a micro-controller is minimizing MuPDF to a size that can fit on a low resource platform. This won't be easy and might not even be possible, but that will be my focus for the time being.

For the latest hardware and software updates please visit the project's repository.