Can I drain a battery at the same time as charging it?

[crazedrobot]

Sorry if this question seems noobish, but I want to have solar panels charge a battery (probably LiPo) at the same time as I am drawing power from it. Would this work? I don't see why it wouldn't but I've never done it before. Thanks.

[definitionofis]

Yes it will work.
Do you have a way to regulate charging so the solar panel disconnects when the voltage gets too high, which indicates the battery is fully charged? Batteries are damaged if they are over charged. Some can explode or catch fire too.


MCP73811 looks simple for LiPO

[crazedrobot]

Thanks for the response. Yeah I was thinking of using this charger which should connect between my solar panel and LiPo and then stop the current flow once the battery is charged.
http://www.sparkfun.com/products/8293

[Daanii]

That's a good way to do it. You lose some current charging the battery. And you lose some current discharging the battery. So hooking the load and the charger both up to the solar panel will give you the best efficiency.

[crazedrobot]

So just to make sure I understand correctly, you're saying I should connect the load to both the solar panel and the battery? (And then connect the solar panel also to the charger connected to the battery). And then when there is no sunlight the load can draw power from the battery and if there is sunlight, it can draw power from the solar panel, right? Thanks!

[Soeren]

Hi,

Sorry if this question seems noobish, but I want to have solar panels charge a battery (probably LiPo) at the same time as I am drawing power from it. Would this work? I don't see why it wouldn't but I've never done it before. Thanks.

No, it's impossible to charge and discharge a battery at the same time (it would be like telling the electrons to go in opposite directions at the same instant - and a schizoid elektron is a very ugly sight )


What happens when you connect both a battery and a charger/solar panel/whatever is, one of the following 3 scenarios...

Battery voltage >> whatever voltage => load takes current solely from the battery and the whatever is either insignificant or a load (depending on construction). No charging takes place.

Battery voltage << whatever voltage => load takes current from whatever and the battery is charged with the excess power. If load takes more current than whatever can give, the load will pull the voltage down to the point where...

Battery voltage = whatever voltage => Battery + whatever shares the load - no charging takes place.


- A load will draw its current from the highest voltage.
- If the battery voltage is lower than the panel/charger, it will charge.
- You need a diode minimum, to make sure you don't send current into the panel when its voltage is lower than the battery, which it will be when it's shaded and at night.

[definitionofis]

So just to make sure I understand correctly, you're saying I should connect the load to both the solar panel and the battery? (And then connect the solar panel also to the charger connected to the battery). And then when there is no sunlight the load can draw power from the battery and if there is sunlight, it can draw power from the solar panel, right? Thanks!

Your last sentence is correct, but the way you said it in the first two sentences sounds confusing.

The solar panel only connects to the charger input. The load only goes across the battery. The output of the charger, of course, goes to the battery.  Then your last sentence is correct:

And then when there is no sunlight the load can draw power from the battery and if there is sunlight, it can draw power from the solar panel, right? Thanks!

[Soeren]

Hi,

If the solar panel goes into a charger, the load will always draw from the battery first and the charger (assuming at least a semi-intelligent charger) will kick in when the battery voltage goes below a certain limit.

Since most solar panels used on 'bots are unable to supply what the 'bot takes, it's better to have it (via a rectifier) directly to the battery.

It would have to be a very large and panel with an outrageously large output voltage and a very small battery to damage the latter (and forget explosions, won't happen).


A solar panel chosen with care (and understanding) and a suitable rectifier (eg. a Schottky diode) is really all it takes, as the internal resistance of solar panels are reasonable high.

[crazedrobot]

Thanks for all the responses, guys!

So if I understand correctly Soeren, you're saying that I don't actually really need a special battery charging circuit, I should just connect the solar panel to a diode and then connect that to the battery?

So it would be like this: solar panel -> diode or charger -> battery -> load

[Soeren]

Hi,

So if I understand correctly Soeren, you're saying that I don't actually really need a special battery charging circuit, I should just connect the solar panel to a diode and then connect that to the battery?

So it would be like this: solar panel -> diode or charger -> battery -> load

Yes.
You should select a panel with the right output voltage compared to your battery voltage and capacity and you need to take the amount of sun vs. clouds in your area into account, as you'll need a higher voltage panel if you have less hours of sun.

The reason this can be held simple is, that a robot able to drive around with a certain area of solar panel will need a battery of a certain capacity and as long as the panel output current is less than what the battery can withstand "forever", there's no risk involved.
One exception is lithium based batteries, as they don't like to be trickle charged (wears them down fast), but lead-acid and nickel based batteries will be good.


Solar panels have a theoretical max. of 1kW/sqr-meter multiplied with ~0.15 (efficiency) leaves you with less than 150W/sqr-meter on a normal day, probably a lot less.
Unless you can keep your panel below 25°C (which you cannot with passive cooling), the output will be much lower

A solar panel rated eg. "12V"/100mA (around 15V/100mA) will probably be a bit too large for a robot (depending on size of course) and these numbers takes the rare case of a merciless sun in a clear sky, shining perpendicular onto the panel to achieve the 1.5W and 100mAh will be fine with a lead-acid or NiMH/NiCd battery with a capacity of  at least 3Ah and I cannot imagine a robot with that size of panel and less battery capacity.

If lithiums are used, a very simple circuit can switch the panel off if/when the battery voltage reach the fully charged state, but don't bother until you know for a fact that you need it - solar power ain't gonna keep your cells charged - prolonging the run-time a little is what you should opt for.

The only place where you really need a clever regulator is when you want to maximize the power output from the panel, but since this involves a switch mode regulator and more, you'll probably find the net gain too small for the outlay (cash, time and learning how it all works and goes together).


If you post the battery type (chemistry), voltage and capacity and what panel you have in mind (links/datasheets help), it would be possible to give more exact advice.

[definitionofis]

I still like MCP73811 a lot because it only costs 63 cents and is as easy to wire as a diode.

[Soeren]

Hi,

I still like MCP73811 a lot because it only costs 63 cents and is as easy to wire as a diode.

I didn't mean to say that the MCP73811 is a bad circuit in any way!

But... It has got options for either 85mA or 450mA, so it's only suited for fairly slow charging (this will be quite fine with a solar panel of course) and only for a single cell, which may be a stumbling point - but I don't recall the OP mentioning the number of cells he want to use.


In general use, the MCP7384x as used by SparkFun would perhaps be a better choice, as it can handle 1 or 2 cells (depending on the last digit in its name) and up to 1A of charge current, although it's more expensive - $1.29 in one off from MicrochipDirect, where the other is $0.51 although it would be insane to buy single chips from MicrochipDirect, as they have a standard shipping fee which would be huge for just a few bits and bobs.

No matter if you buy a single chip or a thousand of one kind, 345 of another and a few of this and that, perhaps topped up with some assemblies/boards and whether it all comes in one batch from one place, or from around the world at different dates, as is the usual modus, the fee is always the same, so I try to foresee my needs a good bit ahead and order a lot to spread the shipping fee).

I do this because the prices in Danish electronics stores are bordering illegal, with eg. a slow dinosaur like the PIC16F84A-04I/P is priced at slightly above $10, the cheapest PIC they have and a slightly newer PIC16F871-I/P, which is about to be phased out, is above $20 (and goes for $3.19 from MicrochipDirect in one off), so with the amount of controllers I use, I'd go broke from shopping locally (not to mention that I'd only have access to more or less outdated chips).

But they "all" seem to make charge controller chips these days, so it's just a matter of determining what you need.


A homemade cut-off, with a comparator as the main ingredient, will be easy to understand and very easy to adapt to any number of cells.
The proper charging needs to be done from the mains anyway, so there's no need for balancing cells when the panel is in action, as this is taken care of by the real charger (hopefully )

[cagiva]

My apologies for using this old thread, but I want to ask Soeren (hopefully still around) a question.

Hi Soeren, I'm planning to use this solar panels with this lithium ion battery so are you saying that it will be okay to connect them in parallel using this sequence?

solar panel -> 10A diode -> battery -> load