unfortunately, it is not possible to supply sensor kit #1 with a battery. The average consumption is about 1 Watt. And the SDS011 is also dependent on a minimum voltage of 5V (exactly 4.8 Volts) and at least 100 milliamper during measurement.
I have tested kit1 only with the BME280 sensor with a battery (5200 mAh/19 Wh, Output 5V-2A) to charge a mobile and it worked more than 24 hours until discharged. When adding the SDS011 sensor, can’t this battery be used for occasional measurements?
Well, technically one can use battery powered solution. It is not documented yet and officially supported.
But there is already a pull request with battery support. If (and when) it will be accepted the official manual might follow shortly…
Part of the of the problem is the power supply for the MCU is a linear regulator.
USB supply is 5.2V Required voltage is 3.3V so it’s only about 60% efficient if you use a switching regulator you can get this into the high 80s-90s
You should also think about removing the USB to serial chip from the equation as that’s just a waste.
Then remove any leds.
If you want to run from batteries I would be looking at a 2S battery pack and two switching regulators one always on 3.3V for the MCU and then a switched 5V regulator for the SDS011 controlled by the MCU.
Just to give you an idea the average current of a NodeMCU is about 40mA and with a switching regulator and no usb chip it goes down to about 20mA.
I’ve been running my sensor for a week using a powerbank. It seemed to work OK, though I did wrap the powerbank in wool last week as we had sub-zero temperatures. 6000mAh lasted 1.75 days.
I used a 20,000mA powerbank last week and it lasted about 6 days.
I am trying to power a minimal sensor (nodemcu + SHT35) by means of a powerbank and had success with an older Anker device. I then got a larger one with solar (PowerCore 20000), hoping that this would increase the time on battery but even on trickle charging mode the Anker shuts down after 2h.
I understand that the purpose of a powerbank is to charge a battery instead of acting as a battery.
What is your experience? Are there powerbanks with high capacity which don’t switch off on their own? Could the firmware periodically switch a load to tell the powerbank that it’s supposed to stay on?
I will share some experience from my experiments. For the past six or so months I’ve been testing solar powered setup with NodeMCU + SDS011 + BME280 build inside a Stevenson shield and positioned on pure South view. For the experiment I chose Rivacase 10000mAh Li-polymer power bank. This setup was working uninterrupted for days. The only problem was that when I tried to use constant connection with a charger the electronics started to go crazy at the end of the charging process. I have to disconnect the charger manually every time the bank was fully charged. Still once disconnected from the charger the Boost inverter inside seamed to work pretty nice. So I decided to slightly modify the power bank. I needed decent protection circuit that will keep battery in working voltage range while constantly charging it and draining power from it at the same time. For that purpose I used a TP4056 boards. And that did the trick. That last setup, separating the battery from the boost inverter with that protection board was running with no issues in outside environment during the whole winter till last week. Then I run a battery test from fully charged until fully discharged. Here is the result chart:
So I could assume any Li-polymer power bank with a decent quality will serve well for this purpose using its Boost inverter to get 5V. But you will need to install separate protection board.
I’ve also disconnected the circuit to the status LEDs of the original board in order to save a few milliamperes.
Can you post your setup somewhere. We talked about such a system yesterday evening in the weekly call. I could then give you the contact of someone in California who want to build a sensor with a solar panel.
Sure thing @pjg. May be I could create a post in my blog and then I’ll share the link here. I’ll need a day or two for that.
Here we go: Solar powered NodeMCU & SDS011 dust sensor – Georgi's blog
I hope I haven’t missed some important information. If so I can always edit it. I hope my experience during the experiment may be helpful to other people.