PCB Design Review Request

Hi all,

Last spring I began chipping away at the wild idea of a "Standalone MIDI Sequencing Drumming Glove", the current demo of which can be found at the bottom of that linked thread. A more apt title moving forward might be "Wearable Drum Machine" or just "WDM".

I'm getting ready to pull the trigger on a JLCPCB order for this board -

https://github.com/ColbyDAllen/Wearable-Drum-Machine/tree/main/Version 1/Hardware/E-CAD

- but seeing as this is my first 4-layer board, I wanted to post here in case anyone has time to take a look. I understand we're all busy, so any feedback would/will be taken quite seriously.

Along with LED indicators, push buttons, and a rotary encoder, the 3 MAIN circuits of this board are as follows:​

1) Lipo battery charger / Power Delivery System
- For battery management, I've copied the Adafruit Powerboost 1000C closely in using the MCP73871 charge controller (U2), and the TPS61090 boost converter (U1)​
- For the system power delivery, I've chosen to route +5V in from the Teensy 4.1's microUSB port, leave the D-/D+ pins intact, and then route the T4.1's VUSB pin straight to U2. From there the +5V is either handed directly back to the T4.1's VIN pin (trace between VUSB and VIN to be cut), or is used to charge the Lipo battery. When on battery power the nominal 3.7V is stepped up via U1 before entering the VIN pin in place of the USB's +5V.​

2) Audio Processing / Audio Output
- I've broken out the SGTL5000 stereo codec + amplifier (U6) in order to save space, taking leaves from a few inspiring projects listed on this forum, including Matt Venn's "Teensy-FX", and Prajwal Mahesh's "Teensy Synth".​
- The onboard speaker circuit is closely copied from the Adafruit STEMMA Speaker board, from which I've employed the PAM8302 class-D mono amplifier (U7).​
3) Force Sensing Resistors
- The core function of this board will be to translate analog pressure readings from fingertips into A) SMF format MIDI files saved to the SD card, and B) audio output for user feedback as mentioned above in 2). I'm re-purposing four 3.5mm aux jacks (J3, J6, J7, J11) as part of the voltage-divider circuit used to make this happen with detachable sensors.​
In efforts to "floorplan" and route this circuit effectively, I've tried to apply a few guidelines mentioned in Rick Hartley's "How to Achieve Proper Grounding". One point I'm still trying to wrap my head around, starting at 1:58:20, explains advantages/disadvantages of the 4-layer GND/SIG_PWR/SIG_PWR/GND vs. SIG_PWR/GND/GND/SIG_PWR stackup. I've chosen to go with the latter in my design, but am unsure why Rick says that "...when you have GND on layers 2 and 3, you have to squeeze all of the signals and components and power pours on layer 1 of the board".​
I think in order to minimize EMI and/or common mode problems, you don't be switching back and forth between layers 1 and 4 by way of vias too often (to keep fields from "blooming"), but is this what he means?​
If you've taken the time to read all this, thank you! I don't expect a free consult, but if anyone's feeling generous enough to offer their advice on how I've laid this board out, I'd listen gratefully to your input😁

Last edited:
One big question I forgot to include above:

Might it be a good idea to split the ground planes for analog and digital, and then to star the grounds at a single point? Sources I've come across are about 75/25 yay or nay on this point. I guess it's really case by case?