Schematic Review Request: ESP32-S3 Mini-1U + SDNand Flash + I2S Microphone for Drone Audio Logger

[emileletueur]

Hello all !

I recently created a small module designed to record onboard sound from drones equipped with Betaflight/Inav flight controllers. It's based around an RP2040 and an INMP441 MEMS microphone, with WAV audio files recorded to a MicroSD card via an SPI module.

It works well and I'm happy with it, but I'd like to miniaturize this module even further by eliminating the SD card and recording directly to SDNand flash memory. This would also remove the risk of vibrations and losing the MicroSD card during drone crashes.

For the new version using SDNand flash, the RP2040 or even the RP2350 are quite limited, so I'm turning to ESP32 chips, specifically the S3 which natively supports this type of memory. There's another reason too - I'm using MicroPython to manage the module.

For "maximum" miniaturization, I want to create a complete board with dimensions of approximately 16x32mm. Since this is my first board design, I'll start with the ESP32-S3 mini-1U version, which has the advantage of already including the basic peripheral components and "works right out of the box" - I'll just add the missing elements around it.

So I've created a first draft schematic in KiCad that I'd like to submit to you for review to identify any initial errors or more subtle issues.

I should mention that I'm a real beginner - my electronics engineering degree is long behind me and I only remember the basics .

One question I have at this stage is: can I do without the boot/reset buttons, and will I still be able to flash, or do I need to add RST/DTR management via transistors like in the ESP32 devkit schematics?

So it would be great if you could tell me if you see any issues and/or improvements to make before moving on to routing.

You will find below a PDF of the schematic.

schematic.pdf

(59.84 KiB) Downloaded 13 times

schematic.png (80.65 KiB) Viewed 843 times

[Sprite]

On buttons: If you can remove/reapply the power easily to reset the chip, I wouldn't bother with a reset button. While you don't need it that often, having the ability to put the chip into download mode (what the BOOT button does) might be invaluable if you ever flash firmware that breaks USB. You can probably get away by putting the signal on a header or testpad or something which you can then bridge to GND if you need to, rather than a button. RTS/DTR management is only useful if you flash the chip via UART, and you're doing it via USB here, so you can forget about that bit.

Few things:
- You must include a RC reset network on the EN pin, otherwise you'll get weird power-on issues
- You seem to feed the PCB from both USB and another power source (via VCC_5V). Is that second source OK with being back-fed 5V if USB is plugged in?
- The TPS73633 you use to feed the -S3 only can supply 400mA. I think we usually suggest a LDO that can do 500mA at least, otherwise you'll get brownouts (mostly when using WiFi/BT).
- While the -mini modules are great for size, they have the disadvantage that none of the pads are available anymore when soldered down. I'd suggest bringing out some of the important signals onto testpoints: GPIO0, U0TXD, U0RXD at least.
- Note that the SK6812MINI technically needs at least 3.5V to function. It may or may not work with 3.3V, but if you need guarantees, better to run it off a higher supply voltage. (Do see if you need to level-shift the IO signal in that case, though.)

[lbernstone]

In addition to Sprite's comments, you should have:
- a 100nF filtering capacitor on the esp32's Vdd line
- 5.1k pull-down resistors on USB CC1 & CC2 to ensure you get the 5V current you need

[emileletueur]

Thx for your replies !
I'll adjust according to your advises and repost a newest version soon !
Again thx !

[emileletueur]

Hello again everyone!
Today, I’m sharing my new version—it’s structured into multiple parts, and I believe I’ve fixed the issues you pointed out. I’ve decided to use just a single LDO, which is far more robust in terms of available current. I’ve also opted for a different RGB LED compatible with 3.3V. Finally, I’ve added a connector/header to access the GPIO0 pin and a small power indicator LED.
Note: No more ERC errors in KiCad.
Can you tell me if this version is better? Thx again and have nice day !

[Sprite]

Looks good, but a WS2812B also needs at least 3.5V of power to operate within spec. Probably easiest to run it off 5V and have a bit of level shifting for the data input, see e.g. the esp32-c6-devkit-c schematics for how we solve that on our devkits.

[emileletueur]

Hello everyone!
I’ve made some good progress here—I’ve been working on the PCB, which isn’t easy, but I have something to show you.
First, here’s the BOM (Bill of Materials) for the module—if you spot any inconsistencies or issues with the components, feel free to criticize or suggest improvements.
And then uou’ll find a view of the finished PCB design. As always, I’m open to criticism and suggestions, so don’t hesitate to share your thoughts.
Just to clarify, I’m still a beginner—I’ve been doing a lot of research, but I know I’m missing out on many best practices.
If needed, I can provide more screenshots.
Thanks to everyone for your past and future help, and most importantly—have a great day!

BOM.png (71.11 KiB) Viewed 669 times

Capture d’écran du 2026-01-14 10-25-33.png (174.64 KiB) Viewed 665 times

Capture d’écran du 2026-01-14 10-26-10.png (105.68 KiB) Viewed 665 times

[emileletueur]

I add an XRAY view of Ground plane and just for fun an 3D view of PCB:

[Sprite]

Looks good, I see no immediate issues with that layout.

[emileletueur]

Many thank Sprite for your reply !