I’m helping my neighbor’s little boy with his science fair project. He want’s to set up a little radio powered by solar energy. So, we’ve gotten a little solar panel (this one), and a little radio.
So here’s what I’m wondering. The little radio runs on three AA batteries, 1.5 volts each – four and a half volts. The solar panel provides 12 volts, 50 milliamps. Since batteries aren’t measured in amp’s, how do I figure out what kind of resistor array I’ll need to make to tone down the voltage to the appropriate level?
I don’t need an explanation, really (not that I’d mind), I’m just looking for the right search terms or phrases.
A purely resistive voltage divider is probably going to get you into trouble. The radio does not act like a constant resistance, so the actual voltage out of your divider is going to vary. You can compensate for this by adding a filter capacitor, which will tend to smooth variations in the voltage output due to varying current.
Resistive voltage dividers also have the very undesirable characteristic that they constantly draw a bunch of power that they throw away as heat, just to do the voltage dividing.
You would be better off using a switch mode step down voltage regulator. This is going to be a lot more efficient, which means you’ll draw less power from your solar panel just to do the voltage step down. Maxim makes a bunch of cheap step down regulators. They aren’t very expensive, maybe a few dollars each.
The LM317 is a linear regulator. It will work, since unlike the voltage divider (resistor network) it will compensate for varying loads and will keep the voltage output constant (that’s what a regulator does).
The difference between linear and switching regulators is that linear regulators are always on, where switching regulators basically turn themselves on and off very fast so that on average they have the right output. Linear regulators have a lot less electrical noise, and switching regulators are a lot more efficient. Switching regulators also don’t work properly if the current drops too low.
The LM317 is going to be easier for a novice to deal with, but if it gives you problems from not having enough current to power the radio, you may need to use the switching regulator instead.
Why not expand the science fair aspect?
Use different voltage reduction schemes such as:
[list=1]
[li]A resistor[/li][li]Linear Voltage Regulator[/li][li]Switching Voltage Regulator[/li][li]Zener Diode(s)[/li][li]Occilator/stepdown transformer/rectifier[/li][/list=1]
Then compare how well they worked (under varying light/load conditions), how much they cost, size requirements, etc
Even write a paragraph on when you would use each method.
A couple things that ought to be checked out as well:
Make sure the batteries really are series connected. Oops, I see it uses three batteries, so this doesn’t seem likely, though it could put them all in parallel, even if running everything off 1.5 V seems unlikely too. I assume below that running off 5V should be okay, though it is barely possible this will overload it if they really designed it for low voltage.
Make sure you’ve got enough current to run the radio. If possible, put a ammeter in to get an estimate of the draw on the batteries. Or make a guess – how long does the radio run continuously with the batteries? Usually this is specified with whatever manual the radio had, or sometimes written on the radio itself. If it runs for about 5 hours, you could be needing all of that 50 mA you can get. (AAA batteries have something like 700 mAH rated at 10 mA, so if they’re in series, that’s what you’ll get.)
Though if the design is marginal, it might be fun to try different light sources to see which ones make the radio run and which don’t. If there is a problem getting it to run, you could expand the project to make the solar panel continously recharge some batteries while using those to run the radio. You could probably then get it to go for days before the batteries go out.
Purely resistive networks would only work well if you didn’t mind being really inefficient. If your device only had a peak power draw of a mA of less, you could come up with something adequate with just two series resistors, in the ratio of 5:7, set so that total current in them is 50 mA or more. Then you’ve got your 5V where they’re connected together. The draw from your device would be insignificant compared to the amount of current in the resistors. Though obviously you’re wasting your max current in the resistors all the time – very inefficient.
If you go with a linear regulator (as already said, the easiest option), you will probably do fine to get a fixed +5V one, just make sure the package will dissipate at least 500 mW (I get this from 12-5 = 7V at 50mA, plus some margin of safety.) Switchers are more expensive, and while the noise probably won’t be a factor, it could be.
Just another angle, odd as it might be, I’ve used diodes in series to drop small voltages like this. 1 Diode drops .7 volts, (Silicon type, 1n400x) so 10 would be needed here. However, a problem still exists with the supply voltage varying due to sun conditions. IMHO The LM317 is still the best design solution. It’s obvious, IANAEE.
Diodes don’t produce a constant voltage drop, even though we often like to model them that way. It’s actually an exponential curve. The simplest model (and the one most often used) is a constant voltage drop. A slightly more accurate model is a voltage drop plus a resistor.
You should only use diodes to drop voltage when the current is fairly constant, and somewhat close to the maximum current of the diode.
I did about 30 seconds worth of research and came up with the following:
An LM317 is going to cost about 75 cents. It’s the least expensive and easiest solution, and also one of the least efficient.
A MAX638 is going to cost about 6 dollars. It will be significantly more efficient. The external components required are only a diode and an inductor, and a sample circuit is provided on the first page of the data sheet. The internal oscillator is 65 kHz, so expect to see a little bit of electrical noise at this frequency and all of its harmonics (130 kHz, 260kHz, etc). They show a filter capacitor in the sample circuit. I’d recommend a small electrolytic in parallel with a small ceramic disk capacitor.
50 mA isn’t a lot of current to play around with. AA batteries are only good for a couple hundred mA max, and I figure the radio doesn’t kill them extremely quickly, so it can’t be pushing them too close to their limits. I think you will probably be ok just using the linear regulator.