Let's Make Robots!

Low-power remote power-independent (solar charged) sensor node!

Well, for a while now I've been entertaining the idea of building a remote sensor node to keep track/record of my indoors "balcony orchard".

This project here will be my starting point, a Low-Power Wireless Sensor Node where most of the work is already cut out for me.

It's power consumptions are reported to be:

Sleep Consumption 0.14 mA

Operating Consumption 13.57 mA

and the author says he can squeeze out around 6 months of operating power using a 540mAh 3V Coin Cell Battery. That would be more than enough for my requirements... but I put it into my head that I want to dabble with solar charging...

after some abbundant generic lmr searches and generic google fu I found quite a few interesting things on the subject, but I grow restless thus I decided that perhaps either:

Ni-Cd 60mAh 3.6V 14h 6mA Rechargable Battery Cells


Ni-MH 80mAh 3.6V 14h 8mA Rechargable Battery Cells

should be enough to keep the module running (assuming that the sensors I plan to add do not screw up my plans) continuously provided I don't run into a long series of overcast/dark days. Given that the module will take residence near a West faced window it will have plenty of solar exposition.

Regarding charging Ni-Cd and NiMH from all that read the easiest way to charge them is trickle charging at C/10, that is at 10% of their maximum capacity. (there are some issues however regarding longevidity of the NiMH batteries when doing this and other rates are recommended).

So I have my eye on this: 4.5V 5mA Solar Panel Power DC Battery Charger

to recharge batteries amounting to 3.6V from what I've seen it's recommended to use solar panels ranging from 4.5V-6V maximum output, and the reported current output is nearly 10% of at least the NiCd battery... and well I guess less can't hurt... it just charges even slower.

Regarding, circuit to do the inline solar charging I'm still looking for the simplest one I can get away with, this because of my self imposed design size constrains:

- design the module as small as possible

- least number of components

For now what only  occurs to be as the simplest is the use of a diode to prevent battery depletion during the night/dark hours. I guess that I should also make sure that recharging is only actived if the solar panel is delivering voltage over the minimum required threshold to actual recharge the battery.

So, that's all for now and I'll be accepting suggestions and the such if have some to give, or find any flaw in my reasoning :P

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

In this application, the single diode is probably the best choice.  The rated current of the solar panel is a maximum, and will often/usually be less.  Most NiCd and NiMH batteries can take low level current continuously without harm.  There is very little danger of overcharging.

As for the last statement about only charging when the voltage reaches the threshold, that isn't a problem.  Remember, whenever there is a voltage difference current will flow from higher to lower voltage.  Any time the solar panel is generating voltage higher than the current battery voltage (plus the diode drop) current will flow into the battery.  In the reverse case, the diode prevents current from the battery throught the solar panel.

I do, however, recommend a Shottkey diode for the lower voltage drop.  At this current level it will only have about 0.2 V drop as opposed to a standard silicon diode's about 0.5V ( Yes, I know the normal answer is 0.6 or 0.7, but that depends on the current.)


Oh, one last thing.  Make sure either that the battery you use can handle the PEAK current (13.5 mA) or you use a large capacitor across it to keep the voltage from dropping during peak current times.

Yeah, after mentioning the diode I've "forgotten all about it" so it should be a good show stopper regarding unwanted decharge via solar panel :)

I've also seen the Schottky diode been mentioned several times for this kind of application I understand the advantage of the lower voltage drop, I just confused a while back regarding them when I was reading about different types of diodes. They seem to me the most confusing in respect to the cares one had to have when using them.

I'm vaguely aware of the reverse leakage current issue from what I understood/read: "As the reverse leakage current increases dramatically with lowering the forward voltage, it can not be too low; the usually employed range is about 0.5-0.7 V and P-type semiconductors are employed only rarely

this means the voltage I'm applying (solar cell output has to be over x Volts, or there will be reverse leakage anyway? or am I misunderstanding it? because if so I guess during the night the voltage output of the solar cell would be around zero, plus I've read tangencially about another issue with solar cells, thermal breakdown. Don't know the details of it, but seems like sustained high temperatures can lead to the solar cell to degrate greatly in performance :s Now that I think of it, that were the issues I had in mind when I was wondering about "extra protection"/control.

Regarding handling peak, I just have to wait/test and see but adding a capacitor is not too drastic (I think -- although I read recently that circuits operating at lower voltages should beware of undesirable capacitance effects --- and my mind spirals out of control :P) 


Using a shottky shouldn't present any real problems in this circuit.  The reverse current will probably be higher than a regular diode, but it probably won't be very high.  The reverse current goes up as the reverse voltage goes up.  3.6V is pretty low, so the reverse current will be, as well.  Probably a few microamps at most.  ALL diodes will have some.

As for thermal breakdown, that is true.  They are semiconductors (silicon diodes, in fact) and like all semiconductors will degrade with heat.  The sun shining on them is the greatest source, but you can't avoid that anyway!  Typical cells convert about 15 percent of the light that hits them to electicity, most of the rest to heat.  That's the main reason large panels have to be replaced: they "wear out."

In your third paragraph I'm don't quite follow the qoute.  Could you give a link to that?  I would like to read it in context.

About the capacitor.  You will need to use a "good quality" one.  Ceramic or tantalum(maybe) or a good quality electrolytic.  Cheap ones have high leakage current that could take nearly as much power as the circuit when sleeping.

Good luck.  Sounds like fun!


I've read it here: http://en.wikipedia.org/wiki/Schottky_diode 4th paragraph. I just gave a quick look to try and refresh the little I had learned about them back then. I'm certain I've degraded the proper context...

...however "The reverse current goes up as the reverse voltage goes up.  3.6V is pretty low, so the reverse current will be, as well." this is re-assuring, I was suspecting this, but my poor interpretation of what I had read previously seemed to indicate contradictory things.

Furthermore, could you advise on how would I go about choosing the proper specs for a Schottky diode in this particular case? would something like a small, general purpose Schottky wouldd do? (eg. BAT46)


P.S. - If I had read the whole wikipedia article until the end :)

"Because of a Schottky diode's low forward voltage drop, less energy is wasted as heat making them the most efficient choice for applications sensitive to efficiency. For instance, they are used in stand-alone ("off-grid") photovoltaic (PV) systems to prevent batteries from discharging through the solar panels at night,"

Apologies, I claim information overwhelm and too much tabs open :)


Thanks for the link.  I see now.  In your case, that particular note most likely isn't relevant.  Most if not all common shottkys have N type instead of P type semiconductors.  

Reverse current:  As I mentioned, all diodes will have some reverse current.  Usually, the higher current the diode is rated for, the more reverse current.  It is good to use a smaller diode if possible.  BAT 46 is rated 100 mA and is a very good choice. If you look at the datasheet here http://www.vishay.com/docs/85662/bat46.pdf on the second page under electrical characteristics it says the MAX reverse current at 10 V is 0.8 uA at room temp.  That same section also shows what I was talking about with greater voltage.  The reverse current rises.  I believe it is linear.  I took a quick look at the 1N5817 data sheet for reverse current.  The 1N5817 is a 1 amp rectifier shottky.  At 20V reverse, it can have as much as 500 uA.  Smaller is better.  As a side note, not really important here, but maybe useful at some point to your or someone else reading this, semiconductor junctions (diodes, bipolar transistors (pnp and npn), etc) are essentially resistors.  BUT, they aren't linear like a resistor.  In a resistor, if you double the voltage, the current doubles.  In a junction, it is exponential, e^x.  That is where the "turn on" voltage comes from.  That is the voltage where significant current begins to flow.  As more current flows, the voltage drop increases, but slowly.  If you care you can google "diode equation"  for the gory, mathy details.  Reverse current doesn't see it as a diode, so is more linear.

As for not catching that last part in the wiki article, I wouldn't feel bad.  When reading those long, detailed explanations it is very easy to overlook the important point you came there for.  Done it many times myself.  

Good luck with this.  I look forward to seeing it.  Once you get it going, I can also offer some pointers on decreasing the power even more.  I have a bit of experience in that area with AVRs.  They can be tricky to get lowest power.  One thing I recommend up front is to use a smaller chip than 328.  A tiny85 would be good if it has enough pins.  Bigger chips use more power.

Hope this helps.

Things are more clear now :D Thanks for the assist :)

The solar/charging part I can only do once the solar/battery stuff arrives from China... so in about 3 weeks... the sensors are also coming on the "slow boat".

As I stated (I think) I want to try to near-spec the requirements as much as I can... I have a few tinies I want to try. The tiny85s are already "looking at me" for me to experiment with them, but I suspect I won't have enough pins for all the things I'm thinking, the tiny84As might do... although I will need SPI for the nRF24L01+ modules... and thus yet another library to provide that on the tinies... I hope I have enough programming space... but I rant... I shall post a specific attiny-related forum post when it comes to that :)