Powering Wemos D1 Mini ESP via 5V pin

Hey guys,

So ive been using a powerbank to power my prototype through the mini USB port on the wemos board during programming and prototyping. Plan is/was to power the thing through the board's 5V pin using the same powerbank to save space.
I'm using a strip of 5 Neopixel LED's and a push button, and control the board over WiFi.

Here's the problem:

When I connect the powerbank to the 5V pin, instead of the USB port, my LED becomes flashy, and won't behave properly (first LED only, random colors, very jittery).

Any ideas why it wont work? I put a capacitor across the power supply, as well as across the LED strip but nothing seems to resolve the issue.

Could it be the D1 isnt meant to be powered directly from the 5V pin? The data sheet doesnt seem to indicate this..

Thanks! (I'm a bit of a novice)

1 Like

I have exactly the same issue and it makes me going crazy :open_mouth:
When I power my wemos via USB and the LEDs via 5V pin from wemos everything is working.
But I think it is no good idea to power my 144 LEDs directly from wemos, therefore I added a 5V supply powering wemos over 5v-Pin and the LEDs with capacitor -> flickering all over the strand :frowning:

Do you have a solution now?

When looking at the schematic https://wiki.wemos.cc/_media/products:d1:mini_new_v2_2_0.pdf
you see the small difference, a diode B5819W and a polyfuse.


So by powering the 5V rail directly you raise the voltage for the LEDs by around .5V,
that could make it harder for them to identify the 3.3V input.

If it works via USB you could insert a similar diode before going to the 5V pin,
that should at least create the same voltage environment.


I was stumped at the same problem. To power the WemosD1 Mini from a battery give it at least 6v.

To power the WemosD1 Mini from a battery give it at least 6v.

No, don't do that. Max is 5.5V. Even 6V could damage.

This is the opposite advice to when using an Uno. With Uno you need around 6.5V~7V minimum if powering through its regulator (Vin pin or barrel socket). But then Uno is a 5V device (Wemos is a 3.3V device), and Uno has a different kind of regulator to the Wemos. The Wemos has what is known as a "low drop-out (LDO) regulator". These have their advantages. They are much more efficient and better for battery powered circuits, but one disadvantage is their much more limited input voltage range.

I know this post is quite old but this might be helpful for someone else.
As Whandall already said voltage at 5V pin will be a bit higher than if powered through USB connector (4,3V) and this makes it harder to recognise the 3,3V input at the datain of the first LED. To solve this problem I can think about 3 things that can work (do one of the three):

  1. add a diode between your power supply and the 5V pin to simulate the voltage through USB connector
  2. add a level shifter before the data_in of the first led to raise the voltage : Shifting Levels | Level Shifting 3.3V microcontrollers and NeoPixels | Adafruit Learning System
  3. instead of the level shifter you can do this: Cheating At 5V WS2812 Control To Use 3.3V Data | Hackaday

All very well and good, but the 74AHC125 is absurd overkill.

Two cascaded gates of a 74HC04 which is likely to be more readily available and cheaper will perform the task just fine, and a 74HCT14 would be perfect.