Any tips for a LiPo battery, charger and solar panel for an always-on node?

ah there’s the sun, time to test this one on an empty 750 mA LiPo

indoor behind glass clouded day test :wink:

@BoRRoZ : Huge solar panels, what kind of power requirements do you have ?

you think 6V - 2W - 330 mA = huge ?

I put this setup yesterday afternoon in the sun - it filled the LiPo during that time with 360 mA

but I have some smaller panels coming soon

0.6W 6V 80 x 55mm

1W 6V 110 x 60mm

2W 6V 330mA 136 x 110mm

A small node with a very small solar panel looks nice, could work very well in certain countries with many sun hours, but during winters and a clouded weeks you’ll need a bigger battery and a bigger solar panel to charge it faster.
All depending on how much ‘juice’ your node&sensors wants to consume of course.

It depends of your power requirement. An RFM95 transmitting data + ATmega328 is about 35 mA. About 1.5 second is required to transmit 10 bytes in SF12, 0.5 sec for the application. Data are send every 5 minutes, which gives 352/360012=0,233 mA/hour, I add 4 uA for standby power consumption.
With 1000 mA, it can last 175 days. If the device is connected to TTN with fair policy restriction it will be a lot more. (times 12 ?)

Of course this simple figure does not take into account self-discharge nor temperature factor but lot of modern uC are less hungry than the ATmega328, so I was just curious about your application :stuck_out_tongue:
Also, Charles spoke about ESP8266 which I don’t know very well the power consumption…

The figures I used are from my sensor connected to solar panel, even if I still need to measure all of them. Also, I tinker around the solar panel yesterday and it seems that my boost converter is far from efficient ! I hope to have some time to characterize it after my vacations but connecting the solar panel directly to the TP4056 works where using the boost did not. By the way, I can’t wait to receive my CN3065 to improve the efficiency :smile:

they seemed to work very well :yum:

@BoRRoZ, I think I missed one important information how consumption can be can be 35mA ? RFM95 datasheet says (several so not sure what the real value)

One page => 120mA

Another One => 90mA

If I understand what you’re writing, does this mean Lora never use PA_BOOST?

I use LTC3105 for solar controller (because need small device and panel ~ 40x15 mm)
Real consumption for RF part ~ 80mA and small battery + supercap provide 2-3 seconds to transmitte. Device size ~ 35x20 mm with sensors, gps.

@charles : True I made a mistake, but we should transmit at 14 dB max in EU, so around 30mA I assume. I will record it after my holydays, I only checked standby current to be sure that the module enters in sleep mode.
@x893 : nice ! Did you post some data somewhere about them ? Which duty cycle ? I wanted to try a supercap only…

@Oliv
oh yes 14dB max, that’s the trick, and so why it’s limited to ~30ma :wink: which in fact is perfect for me, got a node that can’t go over 90mA :wink:

thanks for sharing

received the board and a little disapointement is, that the LiPo connector is 0.1" which is not standard for all my batteries I have.
That should be corrected or at leased two solder holes added.
I would also like 2 solder points/holes close to the micro-usb as input from the solar panel
otherwise the board is of high quality!

still searching for the optimal combination for a small sensornode , and its a sunny day yeah !!

wow … an 18650 with a real capacity of more then 3100 mAh
(discharged for > 3 hours @ 1 Ah to 3V)

one of the best i’ve seen : panasonic NCR 18650B (used in el. cigarettes)

  • warning - these batteries are unprotected

Guys,

I mounted R-Divider breakout board, time to get some code working, but ready to measure from any 3.3V CPU with I2C and no descent Analog input, ohh, did I say ESP8266 ? :wink:

  • Solar Voltage Panel
  • Battery Voltage
  • On Board Voltage

Here are picture, as you can see, seems PCBs.io silk is much better now, really close to OSHPark :wink:
PCB

Assembled
I’ve just put 1% 1M+1M Resistors and 100nF filtering, so we divide input voltage by 2 (so up to 6.6V input)
Mounted

Assembled and stacked with ADS1115
Satcked

1 Like

I was looking at the schematic but couldn’t find the chip you use :sunglasses:

@BoRRoZ
It’s just a R-Divider (if you already have Analog input on device) and if not (like ESP8266) you can stack it with ADS1115 board (I2C 4 ADC inputs) like below. You can see github readme for more information

You can find this ADS1115 board (new version smaller) on ebay for less than $3

ah ok tnx :sunglasses:

and you know… not all LiPo’s are equal - never believe what’s printed on them

1 Like

Hey, yeah I know that, how do you find the real value, you count time to fully charged to discharge at specific AMP ?

yes , I have 2 possibillities, the Maynuo M9812 electronic load and the Turnigy Accucell 6
The Maynuo have a special battery testfunction and can, when connected to PC spit out information/graphs

You set the end voltage (normal between 2.75 and 3.2 V depends on the buildin protection circuit/datasheet)
and you set the discharge current (between 0.2 and 1 C)
then push start and go do the dishes/do some shopping and when reached the end voltage it stops discharging and give you the mAh.

what I really need is a logging power tool… like Ladyada
have been busy finding a diy solution

just curious, did you ever measure the cheap ‘action’ 2000mah AA powerbase batteries?

no, have been testing chinese Ultrafires … that’s really crap