Solar Powered Garden Light - Hack

Thread Starter

StampedePete

Joined Aug 31, 2015
6
Hi Guys,

I'm thinking about hacking a simple solar powered garden light to power my Arduino based sensor array. The light has a single AAA battery that provides 600mA and 1.28V (tested values). The solar panel provides 1.15v (tested value based on indoor light).
Ultimately I am looking to power an Arduino micro (3.3v) about 3 times a day for roughly 2 minutes a pop to transmit temperature and moisture/humidity data. The fact that the battery provides 600mA was very interesting as that is ample power for this setup, and to boot there is ample space in the light body to hide my Arduino and sensors.

The low voltage was my greatest concern, rather than using a boost converter I was thinking of putting three AAA batteries (in series) in there but I would need a new charge controller. Can anyone recommend an IC that would be able to take a 1v input and charge these three 1.2v batteries?

If anyone has built anything similar I'd love advice on the build.

Thanks!,
Peter
 

ronv

Joined Nov 12, 2008
3,770
You might just try it. Most of the lights just have a little boost converter with no regulation except for the current draw from the battery, so odds are fair it will work just be slower to charge.
 

Thread Starter

StampedePete

Joined Aug 31, 2015
6
Ok MikeML, what do you think about two batteries and then a boost converter for the Arduino. Obviously I don't want to do this if this is more work than it's worth.

Thanks ronv, I was hoping that would be the case as I really don't need a lot of amperage for very long. I had originally run the calculations with a 3.3v 500mAh battery and I calculated something like 11 months of possible run time from one fully charged battery. So if I can have the solar panel slowly charge the battery enough to offset the minor draw of the device, that would be great.
 

ronv

Joined Nov 12, 2008
3,770
Sorry, I didn't put that very well. Your 3 batteries would go where the LED normally goes. Easy enough to try. Remove the led and replace it with a 1k resistor then measure the voltage across the resistor.
 

Bernard

Joined Aug 7, 2008
5,784
Garden light with 3 X 3 cm SP gives 2.55V OC after one diode drop, 22 mA SC full sun.
Is there a signal available to announce operation of sensors ?
What is current draw of operating sensors ?
A little later: One k resistor sub for LED gives an output of 11 V pulse .3 us at base connected to smaller pulse of 1.2V , .7 us wide.
Still later: Rectified pulse with 1N4148, filtered with 1 uF ceramic cap. charging 3 Ni-MH AA batteries, 4 V, @ 1.6 mA.
Strange but might work: Keep AAA battery in normal place to charge up during day, replace LED with rectifier & cap to charge 3 AAA's all night; connect load thru a diode to give an output of 3.3 V ??
 
Last edited:

ian field

Joined Oct 27, 2012
6,536
The open-circuit voltage from the panel is nowhere near high enough to charge three 1.2V cells in series.
Might be worth trying with a supercapacitor - but it'd take a pretty big one for a 2 minute run.

The blocking oscillator from a garden light should have better luck charging those than a nickel battery pack.
 

ErnieM

Joined Apr 24, 2011
8,377
The latest generation of garden lights have inside a custom IC that has the reverse charge protection diode, a switching controller, and a dark detector using the panel all in one 4 pin device.

When used in conjunction with a LED they definitely do have voltage regulation, but that is a side effect of driving a LED directly: they regulate at the LED voltage. One could get most any voltage out by removing the LED and adding a zener, along with a bypass cap.

Should you re purpose one of these you need a way to turn the panel off so the device sees darkness t turn on, but while leaving the panel connected to the battery.

There should be no problem getting 3.3 or 5V a few times a day from just one panel and one battery.
 

ronv

Joined Nov 12, 2008
3,770
Garden light with 3 X 3 cm SP gives 2.55V OC after one diode drop, 22 mA SC full sun.
Is there a signal available to announce operation of sensors ?
What is current draw of operating sensors ?
A little later: One k resistor sub for LED gives an output of 11 V pulse .3 us at base connected to smaller pulse of 1.2V , .7 us wide.
Still later: Rectified pulse with 1N4148, filtered with 1 uF ceramic cap. charging 3 Ni-MH AA batteries, 4 V, @ 1.6 mA.
Strange but might work: Keep AAA battery in normal place to charge up during day, replace LED with rectifier & cap to charge 3 AAA's all night; connect load thru a diode to give an output of 3.3 V ??
Bernard,
Can you try it without the cap and diode? Then just measure charge current and voltage?
Nice to have some hardware around.:D
Low budget pulse charger.......
edit:
Might backfeed thru boost Ic.
 

ian field

Joined Oct 27, 2012
6,536
The latest generation of garden lights have inside a custom IC that has the reverse charge protection diode, a switching controller, and a dark detector using the panel all in one 4 pin device.

When used in conjunction with a LED they definitely do have voltage regulation, but that is a side effect of driving a LED directly: they regulate at the LED voltage. One could get most any voltage out by removing the LED and adding a zener, along with a bypass cap.

.

The datasheet I found for a colour changing RGB LED says it wants 4V.

Regardless of whether the oscillator uses discrete components or a custom chip, its basically the LED that clamps the output voltage - you can get pretty impressive output voltage at no load, but it can hardly supply any current.
 

Thread Starter

StampedePete

Joined Aug 31, 2015
6
ErnieM Thanks for your reply. That's what I was hoping. I didn't even think about utilizing the IC that is already there, but that's not a bad idea. Thank you for the IC details. Maybe I can do a 555 timer attached to a relay and cut the panel connectivity every X amount of hours in order to transmit data. Once I get a data sent confirmation then I could switch off the relay in order to turn the panel back on. Not a bad hack.

What about just swapping out the IC? What I've been searching the web for is a charge controller like this one: http://www.adafruit.com/products/390 But one that will function with the 1 - 2V output of the garden light solar panel. The advantage of a charge controller like this would be that I can run my project at the same time as I charge my battery. I need very little current (something like 2mA) to keep the microcontroller in sleep mode, and use an internal clock to send out the data every day at the same time.

I imagine if I tried boosting the solar panel output to meet the demands of the MCP73871 chip I would probably be getting very little current and a very very slow charge rate. It may still be possible due to my project's low power use requirements.
 

Bernard

Joined Aug 7, 2008
5,784
Can you define " project's low power use requirements". What is mA draw of sensor ckt. when transmitting?
Might replace internal boost with a Joule thief battery charger, possibly with over V control. With 2 batteries, one AAA primary & secondary of 3 AAA cells, power available 24 HR a day, no need to disconnect SP, or use relay.
 

Thread Starter

StampedePete

Joined Aug 31, 2015
6
The total power draw of the entire circuit is only 130 mA. So if we are utilizing this power in a maximum burst of 2 minutes (more likely only 30 seconds at a time) then I think our battery can trickle charge very slowly and still stay above water, so to speak.

So I'll design a custom charging circuit for a lithium ion battery rated at 3.3v and 750mAH. Any suggestions or recommendations? I've never made a circuit like this before.
 

Bernard

Joined Aug 7, 2008
5,784
Reading this might be a good start: http:// batteryuniversity.com/learn/article/charging__lithium__ion__batteries.
At 2 mA continuous, and 130 mA for 1 min. 3 X / day gives about 17 mAh, so total about 65 mAh. A charge rate of 20 mA for 6 h should be sufficient. Stop charging at 4 V , disconnect load at 3 V ( should not ever happen ) .
 

Thread Starter

StampedePete

Joined Aug 31, 2015
6
Thanks Bernard, that is a great article.
I'm going to see if I can write up a program to log the solar output voltage every X minutes and leave the panel outside for a few days. I also wonder if the 2mA continuous charge will act as a parasitic load and make accurate charging difficult. What's nice is the power draw is minimal and the battery has ample power so I will err on the side of under charging the battery so as not to stress it.
 
Top