Introducing DG577YI - A Geolocation Recorder - Chapter PCB

2022-09-20

A geolocation recorder aiming at enhanced lifespan.

Basic Info

“DG” is the code for the circuits that I make in my leisure time.

“577” is a number to differentiate different modules. It is a random number I chose.

DG577YI is the 2nd generation of DG577 series. It is a GPS recorder. Very simple, right?

The geolocation data will be read out from the GPS module, and then written into a micro SD card. Potentially, the data can also be transferred to a cellphone or PC using Bluetooth.

It also features:

Quectel L86 as the GPS module, which supports multiple modes on demand.
ESP32-PICO-MINI-02 as the MCU, a ubiquitous module for small DIY gadgets.
DG002YI integration. DG577YI is the first module in which DG002YI plays its role. You may check more details via the link.
Solar cell to charge the battery and power the device. Also supports USB charging.
GPS backup battery.
3-color LEDs to indicate the status
2x side buttons to control the device.
CM1624 as the SD card EMI filter.

Overall the board is simple and does not have many challenges in design. It is roughly equivalent to an electrical engineering bachelor’s degree final project.

Why Do I Need This?

This project was inspired by an App - Fog of World, where you can log your location and “defog” the map of the world, like what you do in a PC game. I really like this concept, whose core laying behind is to explore the world.

However, geolocation is a kind of privacy. You don’t want to leak the date and time that you visit somewhere. I don’t fully trust the apps and any web-based services essentially. If you store the data on their servers then you lose control. The ONLY solution is to build the whole infrastructure from data acquisition to data management, then to data visualization.

So why not develop an app and use the phone as the recorder? Three reasons:

It doesn’t allow for an “attention-free” experience. You have to turn on “recording” when you want to log a trace, and need to stop it when the journey ends.
It drains the phone battery fast. GPS can be power-hungry. I have to save energy for other components.
I’m not familiar with App development. I can learn it and build the things I want, for sure. But re-sharping my existing skills has a higher priority at this moment, in my mind.

Another benefit of extra hardware is that it scales. I can build a few as needed and then track my move seamlessly.

My ultimate goal is to automate the whole pipeline and forget about it. Then whenever I’m interested in my “Fog of World”, I can open the website that I host internally, and check my routes.

The other side of the coin (disadvantages): (1) need extra charging process. (2) Risk of battery expansion or short circuits - I didn’t add many protections right now. Nonetheless, I will take these risks. No pain no gain.

The Design

The schematic design and PCB layout took me a couple of nights. The most time-consuming part, as always, is to choose which components to use. You may wonder, if this is the 2nd generation, where is the 1st generation? What I can tell you is that the MCU I chose for the 1st-gen is not optimal, and its flash and ram are too small to fit the FATFS software stack. So I have to re-design the whole thing (and add more features in the meantime).

The board dimension is 40 x 40 mm, with 2 layers. Unfortunately, the major components are not resizeable, like ESP32-PICO-MINI and Quectel L86. I’m using 0603 SMD components in this generation as this is still proof-of-concept status. In the next generation, I will use 0402 components and find more ways to shrink the board’s size. The 5050 LED can be smaller in the next-gen as well.

After I develop all the code (FW) for this module. I will work on the 3rd generation.

Build It!

As usual, I chose PCBWay to manufacture the boards.

My option is 1.2 mm thickness, 2-layer board, white color soldering mask, and black silk layer texts.

One caveat here is that the minimum spacing in the soldering mask layer is >=0.22 mm. For the CM1624 EMI filter, the pad distance is so small. PCBWay customer service folk reached out to me and provided the guideline to resolve this. Finally, I shrank the size of the pads. Thanks for their heads up and great services, otherwise I may risk failing this version of PCB.

The boards and steel stencil came as exactly as I expected, in less than a week. They are blasting fast!

Check these photos for its great appearance:

Soldering is no news. Applying solder paste and baking for a couple of minutes - Done!

I have to manually solder the through-hole objects, which are the solar panel, Li-ion battery, and DG002YI module. At this moment I didn’t solder the solar cell and the battery yet. I will wait until the software development part is completed.

Next Steps

Tasks for future chapters:

Software development. So far the device is up and running. I’m writing some demo code to spot-check the functionality of each module. The next step is to let them run organically.
Whole device assembly and test (Battery and solar cell). I will bring it with me to benchmark the geo-tracking accuracy, battery life, data model, etc.
3D printing enclosure (box) and final assembly. Hopefully, it will be something wearable.

Gallery

Check out these photos:

The depth of field of the 100 mm Macro lens is so shallow that some components are out of focus. Here is another one taken by the phone:

Lesson Learned

0603 components are so large (in a sense)! The PCB layout looks dense, but they seem to be sparse on the PCB.
The reset buttons are still large. I will look for a smaller form factor in the next-gen.
Quectel L86 needs extra soldering as its pads are larger than other components.