Reddit:Electronics

Hello, a little update from my recent post. I tweaked few things and organized a bit better. I also added the remote control. If you could please check and review the boards, it would help me a lot.

Thank you in advance

Project Description – 24V DC Motor Drive System with BLE Remote Control 1. Overview

The project consists of a complete 24 V DC motor control system that integrates:

• A main control board based on the ESP32-WROOM-32E microcontroller,
• A high-power Pololu G2 motor driver (21 A version),
• A BLE remote control module based on the Raytac MDBT42V (nRF52832),
• And CAN bus communication for external system integration.

The system allows:

• Local control via onboard buttons and sensors,
• Remote control via Bluetooth Low Energy (BLE),
• CAN communication for multi-device coordination in industrial or vehicular applications.

2. Main Control Board 2.1 Power Supply Chain

• Input voltage: +24 V DC from a battery or industrial supply.
• Protection elements:
  • 5KP30A TVS diode for surge suppression.
  • Fuses (1 A for logic circuit, 15 A for motor branch).
• Voltage conversion:
  • Buck converter (XL4015) steps down 24 V → 5 V.
  • LDO regulator (AMS1117-3.3) converts 5 V → 3.3 V for ESP32 and CAN transceiver.
• Filtering: Electrolytic and ceramic capacitors reduce noise and stabilize voltage.

2.2 Motor Control Section

• Motor driver: Pololu G2 High Power Motor Driver (21 A).
• Control signals from ESP32:
  • PWM (GPIO27): Controls motor speed.
  • DIR (GPIO23): Controls rotation direction.
  • SLP (GPIO21): Enables/disables the driver.
  • FLT (GPIO22): Fault feedback from driver.
• The motor driver is powered directly from the 24 V line, while the logic operates at 3.3 V.

2.3 Local User Interface

• Buttons (GPIO25, GPIO26):
  • Forward / Reverse control for manual operation.
• Sensors (GPIO34–GPIO39):
  • Four digital inputs for limit switches.
• Buzzer (GPIO16 + n-MOSFET driver):
  • Audible feedback for warnings, alerts, or connection status.

2.4 Communication and Expansion

• CAN bus transceiver: SN65HVD230.
  • Connected to ESP32’s internal TWAI controller (GPIO32 TX, GPIO33 RX).
  • Differential signals on CANH/CANL for robust industrial communication.
  • Optional 120 Ω termination resistor.
• External connectors:
  • 12-pin screw terminal for sensor and Pololu connections.
  • 4-pin power connector (24 V IN, buzzer, GND).

2.5 Programming and Debugging

• Programming header connected to TXD0/RXD0 (CP2102 bridge).
• EN and BOOT pins are pulled up with 10 kΩ resistors but no onboard buttons are mounted — programming is done externally.

3. BLE Remote Control Board 3.1 Overview

The remote control unit uses the Raytac MDBT42V (nRF52832) module to wirelessly transmit control commands (button presses) to the ESP32 receiver using Bluetooth Low Energy (BLE).

3.2 Hardware Design

• Power supply: 3 V coin-cell battery (CR2032 or similar).
• Optional LDO: only used if other peripherals require regulated voltage.
• Crystal: 32 MHz main crystal + 12–15 pF load capacitors, depending on PCB trace length.
• Buttons: Two input buttons connected to GPIO6 and GPIO8.
• Programming interface: SWD (SWDIO, SWCLK, GND, VCC).
• Grounding: Central ground pad under the module connected to main GND plane.

3.3 BLE Functionality

• Configured as a BLE Peripheral that advertises only to the ESP32 receiver (not visible to smartphones).
• Sends short control packets on button press events.
• Uses low-power advertising mode to preserve battery life.
• ESP32 acts as the BLE Central, scanning for and decoding packets from the remote.

3 inch to 8 inch. Fab has been around since the 60s. Currently the 8 inch is our production size but the 6 inch is still used in the company and they float around as engineering wafers.

My simple lab in my dungeon. Recently picked up the Kepco Programmable Power Supply and Agilent 54622D oscilloscope from work. We’re moving buildings and they’re tossing a lot of stuff. I’m running an Intel NUC with Win11, HP Slice with Fedora, RPi 4b (in the 3D printed green and black Fractal case) with RPi OS, and a Dogbone running Debian. It’s a very simple setup compared to a lot of yours but I love it. A nice place to escape.

Fantastic!!! STM32-based project with an LCD display and a PIR + temp module.
Simple, cheap, and surprisingly effective — perfect for those hot nights when you’re too lazy to turn the knob manually 😎

im k

I just parts-salvaged a Metz camera flash with a burnt-out charging transformer, and found this vintage beauty on the inside.

Tell me what you think!

This bloody LM2576-33 (new) gave out 10V instead of 3,3V. Killed two STM32 before I figured out WTH was going on.

I am, of course, going to widlarize it, as it it written in the ancient scrolls.

After addressing the issue with the shorted Kelvin Leads this instrument a FNIRSI LC2010E, it has so far proven to be a handy tool to have going above and beyond my Fluke DMM.
Symptoms were erroneous readings and it would fail the lead calibration check on the short setting. FNIRSI support has also has responded well sending a replacement.

Today I milled another working pcb (for practice) and its turned out great! In the basic TP4056 chip, only replaced the leds with bigger ones.

Qualcomm acquired Arduino. This is a result of that acquisition. That was quick!

Link to official page

- Qualcomm QRB2210 (- 0.4mm pitch BGA package)

- STMicroelectronics STM32U585

- 8 layer board

As you can see it's a green little toroid with a "secondary" made using green metal wire used for flowers etc 3 windings. Gave the coil some 19MHz 24VPP Sinus wich gave me a result of 5.89VPP 19MHz on osciloscope but with multimeter i got 149.9V in both AC and DC.

Not to sure if this belongs here but i aquired these with an order of vintage NOS. I am trying to find a way to scan these and digitize them correctly. Ill post or link the scans once i can get them done.

DigiKey is known for comically over-packing their orders. A regular $50 order usually leaves you with a lifetime supply in ESD bags and packing material. But today they really went "hold my beer!".

I wanted to start a small side project to 'calm down' from my last big one. And because we need something to replace the old FM radio in our dining room, I started building this, an active smart speaker.

Currently, I have what you can see on the pictures. I pulled all the parts out of an old multimedia system my mother gave to me. The donor-board can be seen on one of the pictures. It was to big to fit in the speakers I am planing to use, so I decided to split it up by transferring the components onto a self soldered PCB and throwing out unnecessary parts out in the process. The amplifier board was quite easy and done in a few hours, but the power supply took quite long as I paid careful attention while building it because you know, things plugged into the mains.

The PSU originally put out multiple voltages for not only the Amp itself(24V) but also the Vacuum fluorescent display and other shenanigans I don't need. I threw out everything except the 24V for the amplifier and the 5V rail to power the RPI and micro-controller that I will put into the device for the 'smart' part. I still have to isolate the bottom of the Power supply and build a small shielding for it to eliminate noise as it will be sitting directly behind the amplifier part.

Like I said, apart from these two PCBs I will also be putting a RPI1 and a STM32 with a LCD screen and rotary encoder into this thing to give it streaming capabilities.

I will keep you up to date on the progress!

I wanted to share a common issue with the LM358 that might help others troubleshooting similar problems.

The Problem (Left Circuit)

Built a simple non-inverting amplifier (gain ≈ 4.9) using an LM358 with ±9V rails. The output showed significant crossover distortion around zero-crossing - you can see the characteristic "flattening" in the waveform.

Root Cause

The LM358 uses an NPN output stage that pulls high well but relies on an internal current source to pull low. When driving high-impedance loads (like a scope probe directly), there's insufficient sink current to rapidly transition through zero, creating a dead zone.

A Solution (Right Circuit)

Adding a 1kΩ pull-down resistor (RL) from output to the negative rail (-9V) completely fixed it:

• Provides a continuous current path to the negative supply
• Enables smooth zero-crossing transitions
• Result: much cleaner waveform with minimal distortion

Key Takeaways

1. LM358/LM324 require careful output loading considerations in bipolar configurations
2. Pull-down resistor to negative rail (typically 1kΩ-10kΩ) enables proper operation
3. This is in the datasheets but easily overlooked in practice
4. For true rail-to-rail with minimal distortion, consider CMOS op-amps (TLV274, MCP6004, etc.)

Hope this helps someone debugging similar issues! The LM358 is a low cost and accessible op-amp great for general or educational/hobby use, but understanding its output stage limitations is key for clean signals. This came up while documenting some lab exercises, and I thought it was worth sharing since it's such a common gotcha.