The problem is finding the atom lite schematic. It is not clear what voltage stabilizer was used. I suspect it has a voltage drop between input and output of ~1.0-1.2V. I connected a li-po cell (4.2V 900mAh) directly into the 5V input. I measured the cell voltage and sent the data to the local blynk server every 20 seconds. Additionally, I measured the internal temperature of the esp32 chip and it oscillated at a constant 50-51°C. With this connection the chip worked for about 6.5 hours and then turned off or stopped working properly at 3.41±0.1V cell voltage. So the esp32 chip probably stopped working at ~2.2-2.4V on a 3.3V supply line. I don't know how usb-uart converter will work with supply voltage lower than 5V. A colleague from "Esp Poland" forum tested power supply for ws2812 diode and at 3.9V the diode went crazy. Screenshot from blynk contain: cell coltage esp32 internalge temperature atom lite uptime/working time [image: 1624622423004-screenshot_20210523-193127-01.jpeg]

@kine90 said in ATOM HUB AC/DC, RS-485 VOLTAGE CONVERTER, VOLTAGE PROBLEM: I am working on a project where the switch is controlled by the M5 PIR SENSOR. When I power the ATOM from USB everything works fine, when I power it with a 12V 3A PSU thought the RS485 voltage converter the PIR begins malfunctioning (fake triggers). I checked the voltage of the two groove outputs (from the hub and from the atom lite) under USB and 12V power. As long as USB is connected, the interfaces shows a quite stable 4.98V. When the USB is unplugged, leaving only the 12V power source, the voltage drops to 3.3-3.28V, if the atom is removed it goes back to 5V I guess this is causing the PIR to misbehave, since it requires 5V. What´s wrong? According to labels on the components and to the HUB schematics, I expect 5V 2A from the DC/DC converter. [image: 1617022245614-1-resized.jpg] [image: 1617022237243-2-resized.jpg] [image: 1617022251558-3-resized.jpg] [image: 1617022264667-4-resized.jpg] [image: 1617022252907-img_20210329_144131-resized.jpg] [image: 1617022277105-switch_schem-resized.jpg] Based on your photo's you haven't connected the 12V to the correct pins of the RS485 connection. 12V needs to be connected to the RS485's red and black pins.

I have also no chance with Arduino and UIFlow :(

check the soldering of the connectors on the tailbat as i had the datapins break

Have you tried the "Cold boot" method i posted in the forum? windows has a preinstalled driver that breaks the current driver requiring a very specific method to remove and install.

Lobo Micropython is out of date. You need the 1.7.7 firmware that is available in the M5Burner firmware installer found on m5stacks download page.

@felmue Tried both. 3.3V has a bit higher current. Strange, if there's just an LDO (guessing - no schematics!). Also removed IR and RGB LEDs. Tried with Arduino and Espressif IDF deep-sleep example. No success. NodeMCU with Wroom32 is kind of OK (1.6mA), by the way. But mechanically too large ...

@ajb2k3 I wasn't aware of ultrasonic I2C! Wonderful!!! Here I've only been able to find GPIO ones Thank you veyr much for clarification

Thanks @robalstona , this works perfect, after some hours of struggeling I can now program the device.

Hi @greenleaf , Yes, it's Baidu AI cloud service.

Hello @alien on the M5Atom there are no predefined pins assigned to RX2 and TX2. GPIO21 and GPIO25 are assigned to I2C so I would not use those. I have successfully tried GPIO19 for RX2 and GPIO22 for TX2. Good luck! Felix

@felmue said in Atom Lite with Earth sensor powered by Tailbat: If, however, you are open to different hardware you can find boards which are designed specifically for low power consumption. I chose Atom Lite with Tail Bat because of: Out of box solution e.g. board with battery including case. It is small and it looks nice. Cheap / good price. Bunch of sensors which can be easily connected. I don't know about any other HW which fits these categories. I'm using also other products for home automation which can be on battery for 2 years at least but it is x3 more expensive so I'm kind of surprised with the result :(

I've posted a guide a few times about using the "cold boot method" to install the drivers and windows and mac cause issue with corrupting the driver if the cold boot process is not followed.