Power Consumption of a Wemos D1 mini ESP8266

RE: Power Consumption of a Wemos D1 mini ESP8266

benaim — Wed, 29 Dec 2021 19:00:10 GMT

HI,

I received my wemos battery shield 1.3 [emoticon:578cee8aac0e41e9bcef50a2f8d281d5]

But... [emoticon:388f399768cb4510bc3d4c26fd6771f6]

I mount it on one Wemos D1 Mini, with one Lipo of 800 mA.

My project is one device who ckeck the moisture on one plant with capacitive sensor.

All, look good, battery charging, report level every 60 minutes, and I use Deep Sleep to have minumum consumption.

But after all is down after 24 hours... Why !!! [emoticon:c5352b3033f144c4ab70d527023c5190]

So I start to check consumption of all parts:

Wemos D1 mini, like I expected, +/- 60 µA in DeepSleep [emoticon:0e469978eb8947d18b9c41123db3725a]

Damned, my wemos battery shield 1.3 ( alone, without D1 ), drain the battery !!! +/- 50 mA... [emoticon:c3bc6fd04d0b4f1f9c1cc7d74eb1c174]

Whyyyyyy ?

On battery shield, all soldering look good. And my input batt ( without batt, of course ) is +/- 100 ohms ( if this can help to solve my problem )

All your good ideas are welcome [emoticon:5ced9b85a4a246058e4772ef2b9956f4]

RE: Power Consumption of a Wemos D1 mini ESP8266

harani — Sun, 05 Jul 2020 01:58:34 GMT

Your major limitation seems to be the minimum voltage requirements of the Wemos (3.56 volts) but a lithium battery is good down to about 3.2v before needing to be recharged. (you shouldn't discharge below approx 3.0v) so you are not utilising a significant portion of your available battery charge

Have you considered putting the lithium through a 3.7>5V boost converter and running the wemos off a constant regulated 5V. Even with conversion losses (you can get aroud 95% efficiency ) you should get way more rruntime out of your battery.

RE: Power Consumption of a Wemos D1 mini ESP8266

neilk — Fri, 15 Feb 2019 16:22:57 GMT

POSTSCRIPT

In the final section the blog I said the following:

The Wemos D1 mini is said to consume around 77 microAmps in Deep Sleep. The on-board USB-Serial interface is said to consume about 50 microAmps, making a total of 127 microAmps for the whole system.
Presumably an ESP8266 presentation like the Witty, which has its USB-Serial interface on a removable board, would consume less power when used without its USB-Serial interface. Depending on the balance between the power consumed when awake and the power consumed when asleep (depends on sleep duration), this could significantly increase battery life - room here for another investigation!

However, some fairly simple arithmetic suggests that these thoughts were somewhat ill-founded.

The ESP8266 element of the Wemos consumes around 77 microAmps in Deep Sleep
The CH340 FTDI to TTL chip forming the USB-Serial interface consumes about 50 microAmps all the time
The resistor chain (120k plus the built in voltage divider between A0 on the Wemos and the ESP8266 A/D pin) used to monitor the battery voltage results in a constant drain of around 100 microAmps or less.

This adds up to a total drain of around 227 microAmps, maybe less.

What does this represent in power terms?

227 microAmps for 1 hour represents 227 microAh; so for 100 hours it represents 22,7 mAh and this is being taken from a 2400 mAh battery (which may or may not not be full charged to begin with). Pretty negligible, I think.

My conclusion, therefore, is that almost anything which can be done to reduce the sketch execution time and/or the time spent using WiFi, could easily improve battery life. Failing any improvement there, any increase in the time interval between readings should definitely give a proportionate increase in battery life.

I hope my arithmetic is correct!!

Neil

RE: Power Consumption of a Wemos D1 mini ESP8266

genebren — Fri, 15 Feb 2019 01:51:40 GMT

Neil,

Not all Li-Ion batteries live up to there rated capacities. There are some super cheap batteries out there that a really junk. I ran into this problem when developing a Li-Ion battery pack for a robot project. The battery I was using during development would only achieve ~20% of the rated capacity. Replacing the battery with one from a reputable supplier (Panasonic) yielded a result where 100% of the rate capacity was actually observed during testing (see: Walky the Biped Robot - Power pack and Walky the Biped Robot - A new hope (actually a new body and battery) ).

On the issue of not reaching the expected voltage, I did find some reviews of the Wemos D1 Battery shield, where the authors saw the same issues (https://arduinodiy.wordpress.com/2017/01/02/reviewing-the-wemos-battery-shield/ ). Not all charger chips and/or charger designs are equal. I have found no end of problems with some of the cheap charger chips and even more so with bad charger designs.

Good luck finding the right combination of battery and charger!

Gene

RE: Power Consumption of a Wemos D1 mini ESP8266

fmilburn — Thu, 14 Feb 2019 23:20:37 GMT

Hi Neil (and Nico),

Great post with lots of interesting data and thoughts... If you haven't seen it there is a post by John Wiltrout on using cell phone type USB battery packs as power supplies. Also, on Nico's point about battery shields there is a recent YouTube video by Andreas Spiess on an inexpensive ($2.50 including shipping) power supply board with built in 18650 battery holder which among other things allows the use of solar panels. I ordered a couple of these and have so far been quite pleased with them.

Frank

RE: Power Consumption of a Wemos D1 mini ESP8266

ntewinkel — Thu, 14 Feb 2019 22:51:22 GMT

That's really interesting research, which is going to be very useful for me too, thanks Neil!

I still haven't even had time to implement the dual graph you put together. I want it for a similar reason - outside temperature and hummingbird feeder temperature.

And interesting note about the WiFi power draw - I'm surprised how much of a difference combining the server calls has made. I kinda thought you'd be able to do 2 or 3 in a row for similar power as just 1, given the setup effort, but that seems to be a bad assumption.

Are you still doing 1 minute intervals? I wonder if storing the information for 5 or 10 readings would again save a lot of power without losing much functionality. You would then only be sending the info to the server at those intervals (once per 5 or 10 minutes), rather than every time.

About the deep-sleep, does that fully reset the Wemos when it wakes up? ie, for the sketch it's like starting up from scratch?

Cheers,

-Nico

RE: Power Consumption of a Wemos D1 mini ESP8266

three-phase — Thu, 14 Feb 2019 22:24:48 GMT

Interesting blog to read, good luck with the further experiments.

Kind regards.