Home › Forums › Mayfly Data Logger › Dying battery in winter solved by 3.5 W solar panel
- This topic has 5 replies, 4 voices, and was last updated 2017-02-23 at 10:10 AM by SGFulton.
2017-01-03 at 2:23 PM #1893Anthony AufdenkampeParticipant
Some early adopters of our wireless sensor stations, such as those shown in our post: EnviroDIY Mayfly logger stations deployed in PA, DE and MN!, have seen their battery voltage plummet and systems go offline over periods doing the winter. One assumption is that the cold temperatures caused the problem, but the Mayfly board is good to -40C and the LiPo batteries we use have worked well down to -20F (see Mayfly Operating Temperature Limits forum response). Rather, in our experience, low battery voltage in the winter is almost always due to decreased solar charging that comes with shorter days and lower light intensity of winter.
Lucy Rose, our partner in Minnesota, recently reported “There doesn’t seem to be a strong relationship between air temperature and battery voltage. There does seem to be some coincidence between sunny/cloudy conditions and datalogger performance, though. I looked at the hourly met data from the Chisholm-Hibbing airport (about 10 miles from our sites) and it does appear that the periods when the West Swan Upper datalogger cuts out coincide with stretches of overcast conditions; the battery voltage increases and the datalogger begins reporting data again when the sky becomes clear.”
We had similar issues with our stations deployed last year in a deep canyon in Bhutan.
Luckily, the fix for not enough solar charging is easy: replace the 2 W solar panel with a 3.5. W solar panel (https://www.voltaicsystems.com/3-5-watt-panel or https://www.adafruit.com/products/500). Since we did this in Bhutan, we haven’t had a single winter dropout.
2017-01-07 at 3:24 PM #1922ArnoudParticipant
Thanks for the info. Does the cable of the Adafruit solar panel need some adjustment (or additional jack), or one can plug it in directly?
With friendly greetings,
2017-01-07 at 11:08 PM #1923Shannon HicksModerator
It’s not mentioned above, but equally important to surviving the cold weather and reduction in sunlight is to increase the capacity of the LiPo battery being charged by the panel. Depending on the type of sensors used and their current draw, a smaller battery might have trouble making it through a long stormy winter week, no matter what size panel is connected to it. So I recommend increasing the battery capacity to as large as you can comfortably get in your enclosure.
In our loggers, I use 2000mAh batteries in sunny spots, a 4400mAh battery in shady spots, and a 6600mAh battery in really problematic areas. This, combined with the larger panel, and reducing the transmit rate of the radio/cell module (this is the biggest power hog) should get you though anything.
And to answer the question above, the Voltaic Systems panels come with a small barrel-type connector. If you want to use those panels with a Mayfly, you’ll need to connect it to the JST header on the Mayfly, by either cutting and splicing your own homemade cable to the panel, or by ordering one of their barrel-socket-to-bare-lead cables, available on their website.
2017-02-22 at 11:40 AM #2049SGFultonParticipant
Hi, I note that the specs for the Mayfly state they only support a 2.5W solar panel. Here I see that you’re able to connect a 3.2W panel. How about a 6W panel (e.g., https://www.adafruit.com/products/1525) or 9W panel (https://www.adafruit.com/products/2747) or up to a 20-100W? We’re running Arduino Megas in the field hooked up to 12V marine grade batteries recharged via a larger 20-100W panel (e.g., https://www.amazon.com/ECO-WORTHY-20W-Solar-Panel-Polycrystalline/dp/B00PFGP0EA) and solar panel charge controller.
I should add that we’re powering a lot of sensors and components (Atlas Scientific specific conductance, BME280, Dallas 1-wire temp, XBee, FONA 808, Druck pressure transducer, LCD screen, real time clock, SD card, etc.).
2017-02-22 at 11:55 PM #2051Shannon HicksModerator
I’m not sure where you saw something that said the Mayfly only supports a 2.5W panel, because that’s not accurate. The only limiting thing is the voltage that the panel produces — the Mayfly only works with panels with an output voltage of 6 volts. If you use a solar panel with a 9v or 12v output, you will damage the Mayfly.
The reason for the voltage limitation is the MCP73831 charge controller chip on the Mayfly board. It can’t handle input voltage above 6 volts. It also has a maximum useful input current of 500ma, so if you do the math (Ohm’s law) you can calculate that a 3-watt panel will provide 500ma of current. So for sunny locations, any panel that supplies more than 3 watts is overkill. If you are in a shady location, or in areas that don’t get at least an hour of direct sunlight, then using a larger panel like the 3.5-watt (https://www.voltaicsystems.com/3-5-watt-panel), 6-watt (https://www.voltaicsystems.com/6-watt-panel), or even 9-watt (https://www.voltaicsystems.com/9-watt-panel) might be useful. The 9-watt panel puts out 3 times more current than the Mayfly can use (1500ma) but if your logger is really shaded, then the efficiency of the panel drops significantly, but even at 33% output, it would give same power as a 3W panel in full sunlight, and the maximum useful current the Mayfly will accept.
We usually use the 2-watt panel (https://www.voltaicsystems.com/2-watt-panel) in almost all of our locations because they are sunny enough to keep the battery charged. But in a few areas we use the 3.5-watt and have had great results, so we’ve never needed to use anything higher than that.
If you’ve already got a 12-volt deep cycle battery and separate 12v panel and charge controller, then it probably makes more sense to power your Mayfly board from the 12v battery and not use a 3.7v LiPo to power the Mayfly, and then you don’t have to use a separate 6v panel just for the Mayfly. To power the Mayfly from 12v, you have to use a 12v-to-5v adapter like this one (https://www.amazon.com/HitCar-Inverter-Converter-Recorder-Straight/dp/B00U2DGKOK) and plug it into the microUSB port on the Mayfly. There is an external battery header (unpopulated) on the Mayfly (version 0.3 and 0.4 boards), but it has a maximum input voltage of 5v. The upcoming Mayfly version 0.5 will be able to accept a connection directly 12v batteries, but for the older boards, you’ll need to use the 12v-to-5v adapter. It’s easiest to use the ones that have the microUSB connector on them, but you could also use one with bare wires and connect that to the external battery header using a polarized 2-pin connector with 0.1″ pitch.
2017-02-23 at 10:10 AM #2055SGFultonParticipant
Hmmm, now I don’t see where I saw the 2.5W reference. Perhaps I was mixing it up with the Sodaq Mbili. That’s good to know. At least one of the locations I’m intending to sample in the eastern KY coalfields, and possibly others at a different site here in GA, are very forested. I’ll have to try to run the solar panel out to a sunnier spot to ensure it’ll work long-term.
Are you using a cellular modem or XBee type device on any of your sites? If so, how often are you uploading, and how is the batter supply in those cases?
Thanks for responding so quikly!
- You must be logged in to reply to this topic.