Node running on supercaps and a solar panel.
-
Hi,
I'd like to share my experience of running a small sensor, which currently only reports voltage every 5 minutes on supercaps with the tiny solar panel.
My setup:
-
A stamp size atmega328p sensor with ATSHA204A. Sleeping consumption is around 5 uA.
-
The harvesting board by @ceech with two 10F supercaps.
https://www.ebay.co.uk/itm/BQ25570-thermal-solar-energy-harvester/332071662285?hash=item4d510372cd:g:yG8AAOSwJ7RYXkVQ -
a small solar panel.
The harvester is done on BQ25570 with 3.3v converter for sensors. It is harvesting as low as 100mV which is excellent - no need for a the sun to charge supercaps. Can also charge supercaps indoors.
In the beginning I fully charged supercaps to 5V and then disconnected the solar panel. Voltage was slowly dropping though my sensor did not sense it as it is connected to 3.3V converter. However, as soon as voltage dropped below 3.3V, the the sensor could see it dropping. The drop is around 0.2V a day, which means that from 5V to 2.9V without the solar panel disconnected, the sensor was sending voltage to controller every 5 minutes for more than 10 days!!! 10 days without a single change! Then as soon as I connected the solar panel, voltage recovered literally in minutes. See the graph below
Very impressive in my view. TO DO - I'll try to assemble a more complex weather node reporting 6-8 parameters to the controller with 10-15 minutes sleeping time to test this setup.
-
-
Hi,
I'd like to share my experience of running a small sensor, which currently only reports voltage every 5 minutes on supercaps with the tiny solar panel.
My setup:
-
A stamp size atmega328p sensor with ATSHA204A. Sleeping consumption is around 5 uA.
-
The harvesting board by @ceech with two 10F supercaps.
https://www.ebay.co.uk/itm/BQ25570-thermal-solar-energy-harvester/332071662285?hash=item4d510372cd:g:yG8AAOSwJ7RYXkVQ -
a small solar panel.
The harvester is done on BQ25570 with 3.3v converter for sensors. It is harvesting as low as 100mV which is excellent - no need for a the sun to charge supercaps. Can also charge supercaps indoors.
In the beginning I fully charged supercaps to 5V and then disconnected the solar panel. Voltage was slowly dropping though my sensor did not sense it as it is connected to 3.3V converter. However, as soon as voltage dropped below 3.3V, the the sensor could see it dropping. The drop is around 0.2V a day, which means that from 5V to 2.9V without the solar panel disconnected, the sensor was sending voltage to controller every 5 minutes for more than 10 days!!! 10 days without a single change! Then as soon as I connected the solar panel, voltage recovered literally in minutes. See the graph below
Very impressive in my view. TO DO - I'll try to assemble a more complex weather node reporting 6-8 parameters to the controller with 10-15 minutes sleeping time to test this setup.
@alexsh1 that is awesome! Was this with signing enabled and with the receiver of your messages requesting signatures? Would be interesting to see what effect the atsha and signing algorithms would have on the consumption. The atsha is kept in dormant state unless it is needed.
-
-
@alexsh1 that is awesome! Was this with signing enabled and with the receiver of your messages requesting signatures? Would be interesting to see what effect the atsha and signing algorithms would have on the consumption. The atsha is kept in dormant state unless it is needed.
@anticimex No, this time I have use Sensebender GW Serial and have had a few issues with the computer going into a sleep mode. Therefore, I disabled signing for now. Testing signing would be the next step.
-
@alexsh1 what brand and where did you get them? I am running too a solar node but even with more capacitance than yours I am not getting that much runtime, probably because I am sending every 10 minutes temperature, humidity, tx pwr percentage, tx and rx rssi, battery % and supercap voltage
-
@alexsh1 what brand and where did you get them? I am running too a solar node but even with more capacitance than yours I am not getting that much runtime, probably because I am sending every 10 minutes temperature, humidity, tx pwr percentage, tx and rx rssi, battery % and supercap voltage
-
I noticed there isn't much attached to it, so a real case scenario with a temp hum sensor could be more useful
-
I have got my hands on 30F 2.7V supercapacitor.
With it was charged for the first time, it only lasted sending 2583 times (1 time a minute voltage and an 20ms LED pulse and then sleeping; sleeping consumption is 4-5uA). The number is a bit disappointing. It is however connected via boost MCP1640 to provide the node with a stable 3.3V. I may try to get the supercapacitor connected directly to the node to see if I can extract more packets.
-
I have got my hands on 30F 2.7V supercapacitor.
With it was charged for the first time, it only lasted sending 2583 times (1 time a minute voltage and an 20ms LED pulse and then sleeping; sleeping consumption is 4-5uA). The number is a bit disappointing. It is however connected via boost MCP1640 to provide the node with a stable 3.3V. I may try to get the supercapacitor connected directly to the node to see if I can extract more packets.
-
I have got my hands on 30F 2.7V supercapacitor.
With it was charged for the first time, it only lasted sending 2583 times (1 time a minute voltage and an 20ms LED pulse and then sleeping; sleeping consumption is 4-5uA). The number is a bit disappointing. It is however connected via boost MCP1640 to provide the node with a stable 3.3V. I may try to get the supercapacitor connected directly to the node to see if I can extract more packets.
@alexsh1 said in Node running on supercaps and a solar panel.:
I may try to get the supercapacitor connected directly to the node to see if I can extract more packets.
That's your best choice, unless you're also running a PIR or something that requires a higher minimum voltage. In that case, you could switch to this boost converter, which has only 7ua quiescent current:
https://www.openhardware.io/view/285/33v-Boost-Converter-with-Pass-Through -
@alexsh1 said in Node running on supercaps and a solar panel.:
I may try to get the supercapacitor connected directly to the node to see if I can extract more packets.
That's your best choice, unless you're also running a PIR or something that requires a higher minimum voltage. In that case, you could switch to this boost converter, which has only 7ua quiescent current:
https://www.openhardware.io/view/285/33v-Boost-Converter-with-Pass-Through@neverdie I think you are right. I have tinkered with your other boost converter - https://www.openhardware.io/view/279/Adjustable-Boost-Converter-with-Pass-Through
MCP1640's quiescent current is much higher 19uA.
Now, I'll try to run it from the supercapacitor directly. Pity that anything below 1.9V is going to be wasted (nrf24l01+ is unstable around 1.9V)
-
I bet you will see a proportional drop according to the number of values you want to report. I am still stressing it a bit, so I think there may be margin to save some power by sending less important values with a lower frequency
@gohan Just to let you know. I have been running this node with Si1132 + Si7021 + another sensor (reporting more than 6 values every 5-10 mins) for weeks now. I have just measured the caps = 4.8V (peak is around 5.2V) - it is late at night. In the morning the voltage will go up. I managed to find a very small, but efficient solar cell smaller than in the photo.
The key is to have a low sleeping consumption, i.e. all sensors must sleep properly. In this case you get 4-8uA sleeping current and can live on supercaps for days without recharge.
-
@gohan Just to let you know. I have been running this node with Si1132 + Si7021 + another sensor (reporting more than 6 values every 5-10 mins) for weeks now. I have just measured the caps = 4.8V (peak is around 5.2V) - it is late at night. In the morning the voltage will go up. I managed to find a very small, but efficient solar cell smaller than in the photo.
The key is to have a low sleeping consumption, i.e. all sensors must sleep properly. In this case you get 4-8uA sleeping current and can live on supercaps for days without recharge.
@alexsh1 said in Node running on supercaps and a solar panel.:
I managed to find a very small, but efficient solar cell smaller than in the photo.
Please do tell: what did you find? And where?
-
@neverdie I think you are right. I have tinkered with your other boost converter - https://www.openhardware.io/view/279/Adjustable-Boost-Converter-with-Pass-Through
MCP1640's quiescent current is much higher 19uA.
Now, I'll try to run it from the supercapacitor directly. Pity that anything below 1.9V is going to be wasted (nrf24l01+ is unstable around 1.9V)
@alexsh1 said in Node running on supercaps and a solar panel.:
MCP1640's quiescent current is much higher 19uA.
Or, better yet, just turn it on prior to your sensor readings, and then turn it off.
-
@alexsh1 said in Node running on supercaps and a solar panel.:
MCP1640's quiescent current is much higher 19uA.
Or, better yet, just turn it on prior to your sensor readings, and then turn it off.
-
@alexsh1 said in Node running on supercaps and a solar panel.:
I managed to find a very small, but efficient solar cell smaller than in the photo.
Please do tell: what did you find? And where?
@neverdie sure - this is the one i am very much happy with. Previous one was not supplying enough juice to charge fully supercaps.
