# Solar battery charger

#### stupidlogic

Joined Aug 10, 2010
39
I am working on a design project where I need to use power generated from a solar panel to charge a battery and then in turn power LEDs after dark. I am on a team and my assignment is to design the controller to regulate the battery charging and driving the LEDs.

I am just curious if anyone has done a project similar to this or components of it. Such as an LED driver, battery charging methods, or regulating input from the solar cells.

Efficiency is of utmost importance for this project. Would using a microcontroller to run a charging program be appropriate or is there something 'smart' enough to do it on its own? We haven't decided whag type of battery we are using yet. We are leaning towards Li-Ion/Polymer or Ni-Cd.

I guess any input on designing a controller that can do something like this would be helpful. Right now Im working on the research phase but will be building a protoype in a couple months.

Thanks for any help!

#### BJT_user

Joined Oct 9, 2011
35
Hello 'Logic. This is an interesting project. Firs could you tell us what kind of voltages you are looking at. What is the voltage output of the solar cell(s) and what is the voltage and type of batteries you wish to charge?

I'm aware of many outdoor LED lamps that use solar power to charge them during the day, then turn on at night. I've disassembled one before and all I find is a few discrete components, IE: transistors, diodes, resistors and a photo detector. If the voltages in volved are high enough, the circuit could be built with a 8-pin DIP dual op-amp. But for lower voltates, you may require more discrete circuits.

#### russpatterson

Joined Feb 1, 2010
353
Hello,

Sounds like a fun project. I started down this one a couple of years ago (in my limited spare time). I used 12V AGM (sealed lead acid) batteries because they are very recyclable, cheap and relatively easy to use.

I documented most of the project on my blog:
http://backyardsolar.blogspot.com/

If you look back in the blog entries I provide the whole schematic then go through each section of the solar controller. The controller has a light sensor so you can run your LED lighting at night. I recently added a PIR motion sensor to it as well. Maybe some of this will be of use to you. Let me know if you have any questions about any of the info there.

I think a microcontroller is key for this kind of stuff. I can make the PCB available to you at BatchPCB.com if you're interested.

#### stupidlogic

Joined Aug 10, 2010
39
As a team we haven't decided what sort of voltage levels we are going to use yet... or the batteries. Which I know makes a big difference.

We don't have a lot of room for the solar panel and have fairly harsh energy requirements (it must be able to light LEDs to read for 4hrs/day). We are probably going to build a custom solar array from individual cells to give us the voltage/current requirements we need; so there is some play here. Basically this post is to try to get some ideas of where to look for more information.

Batteries... probably going with LI-Ion/Polymer or Ni-Cd, but our "battery expert" doesn't have all his research done yet.

I am working on the controller for this project so I'm a little held up waiting for my team mates to get some of their research done so I know what kind of voltage/current supply we will have on a good day/bad day of sunlight; what kind of battery we are using; and what power requirements do the LEDs have.

We know that this project is impractical from a cost standpoint but the object is more to do a feasibility study/sub-system design.

Areas of importance are to develop a high efficiency system that requires very small amounts of energy to run. The LEDs are kind of stuck at whatever it is for the given LED and the amount of energy available per day is variable due to weather, seasonal changes, and latitude. (We are going to be doing an analysis of solar intensity with the poynting vector, etc. to get an idea of how much energy we have depending on how close to the equator we are.) To simplify, we need HIGH EFFICIENCY CHARGING and VOLTAGE REGULATION.

I've been looking at using buck/boost DC/DC converters vs a linear regulator for regulating voltage from the panel. Also I read an article about variable frequency DC/DC converters.

Am I looking at getting a microprocessor to run this sort of control system or would it be easier/more efficient to use discrete parts, etc?

Any information is good hear, I'm completely open to any suggestions until I research out and we decide on something.

#### k7elp60

Joined Nov 4, 2008
562
I just finished a similiar project. I use a solar panel to charge a 12V AGM battery and LED's to light 7 segment displays to display my address at night. I use the solar panel as the light and dark detector and use a PWM circuit driving a P-chan mosfet to power the LED's. The PWM controls the intensity of the LED's. I live in a desert community where the temperature variation is great and everyting is outside and is waterproof.

#### russpatterson

Joined Feb 1, 2010
353
@stupidlogic, if you're doing the controller you should be involved with selecting the type of battery to use. The different battery chemistrys demand very different charging techniques and circuits.

#### stupidlogic

Joined Aug 10, 2010
39
@stupidlogic, if you're doing the controller you should be involved with selecting the type of battery to use. The different battery chemistrys demand very different charging techniques and circuits.
I realize this and we are all apart of each other's research giving input and ideas, but we each have an area that we are primarily responsible for in order to divide up the work load.

I'm going to be doing a lot of research this weekend and if I have the time I'll post some of what I find on here. I appreciate all the ideas/suggestions so far!

And on a side note with LEDs... this was brought up to me. If you pulse the LEDs fast enough that the human eye can't tell, would that use less energy than to have a constantly on?

#### russpatterson

Joined Feb 1, 2010
353
Yes on the LED's. I've heard that the eye can't tell a brightness/illumination difference between PWM at 100% or 80% (anything over 80hz is hard to detect, I usually run them at 200hz) but I haven't done any real tests to verify.

Look at LEDSupply.com for some modern, high output LED boards, heat sinks and drivers (get the wired buckpuck, the non-wired are a hassle). Be sure to have some optics (a lense that fits the board). as that will improve the performance for reading quite a bit. (safety tip, don't look directly at the LED's, never power one w/out a proper heat sink or you will shorten the life span considerably, use a metal core PCB).

#### stupidlogic

Joined Aug 10, 2010
39
Thanks for the input. I'll direct my LED "expert" to that website as well as check it out myself. Just off hand do you have any idea what sort of power requirements the high output boards have? Power consumption is a primary concern of this project. Also we already have plans to use diffusion lenses and stuff to help disperse the LEDs (one of our guys is a photonics major so he knows enough about light/lenses and stuff).

Once again thanks for the help.

#### russpatterson

Joined Feb 1, 2010
353
I would use a one-up board like this one: http://ledsupply.com/creexml.php

Looks like from 2-10 watts. The data sheet is available from the product page.

If the requirement is for one person to be able to read by the light then one LED with a wide angle lens should do it.

#### wayneh

Joined Sep 9, 2010
17,172
You haven't mentioned whether cost is an issue. If so, I'd look for panels pulled from those cheap solar landscape lights. Practically free if you shop around. They come with batteries, but if you want more, find a spent laptop battery pack and open it up. You'll find several Li-ion cells, and in my experience most of the cells will be OK. When one dies, the whole pack goes. Anyway, these give good energy density and since they're 3.7V each, it's easier to get to higher voltages if you want.

For the sake of efficiency, it'd be best to avoid bucking or boosting if you can. While 90% efficiency or more is possible, it's a loss you can avoid by careful pairing of battery to panel.

You'll definitely need to think about charging strategy. Nicads are lead-acid are fairly easy, since they can tolerate a higher current at full charge without damage from overcharging. So all you need to do is drain off any excessive current from the panel once the full-chrage voltage is reached. NiMH batteries need to see essentially zero current once they've been taken to full charge (which is also harder to detect than with lead or nicad). I believe Li-ion are similar but you'll need to study up on it.

One design issue I've never seen addressed is how to design around state-of-charge, or SOC. What I mean is that charging efficiency is maximal when a battery is at, say, 60% of full charge. So ideally - for efficiency - the panel would always be charging batteries that are at the opimal SOC. But normally the battery is at some lower state in the morning and is probably near or past full charge by sundown. I think designers look for maximum hours of LED lighting (favors fully charged battery) rather than maximum efficiency per se. I don't know where the sweet spot is.

#### stupidlogic

Joined Aug 10, 2010
39
You bring up some very good points. Cost is a factor to some extent. This is a senior design project for my EE degree and we have an industry sponsor paying for our prototypes. The final product is supposed to be VERY cheap to manufacture (i.e., < $50 in quantities of 100,000+), but we know that this cost requirement is going to be hard to meet so it's not that strict of a requirement. Just something to aim for while developing efficient systems. The other part is that we are supposed to test them up here in the Upper Peninsula of Michigan in the winter. Basically no sunlight, so herein lies some of the difficulty. Even in days where there might be a week or more with no direct sunlight this is suppose to function properly. I guess that was just some more details about it. I really appreciate all the pointers and information provided. It's very useful and as I start my search. #### Smoke_Maker Joined Sep 24, 2007 126 Logic, Been here done this. Key points are NiCd beats Lead acid, NiMh beats NiCd and Li-ion beats NiMh. Look at cost and cycle life, set up a spread sheet listing all battery choices and see which one beats out the price per cycle. My guess is NiMh, also look at environmental factors. You do not want to use the battery's full capacity, cycle life of the battery will go up with a shallow discharge, your graphs should show you what percent of charge to use. When you decide on which battery to use you will find specific chips designed for that battery chemistry, IMHO you do not want or need a PIC. Goggle ITC4060 LED burn time, how long do you want the LED to be on before the battery has to be recharged, hours, days, weeks, local weather will give you a idea of how big a battery (AH) you need. With only using a small percentage of the battery capacity LED current regulation will not be as big a deal any anymore. Solar cell, you do not want any bigger solar cell than needed, this project could be a very small foot print depending on how much light you want and how long you want the light to burn with out sunshine. A nuclear winter will make this project very difficult, but then again may be your light will be the last one burning #### #12 Joined Nov 30, 2010 18,223 My recommendation is to think backwards. Start with how much light is required, then what LED or combination thereof will make that amount of light, then you can figure the amp-hours and voltage to feed the LED(s), then the battery and its charger, then the solar cells. #### iONic Joined Nov 16, 2007 1,654 My recommendation is to think backwards. Start with how much light is required, then what LED or combination thereof will make that amount of light, then you can figure the amp-hours and voltage to feed the LED(s), then the battery and its charger, then the solar cells. I agree. This is a very good way to think about this project. Even the statement "we need enough light to read for 4hrs" is vague. To read what and from what distance? 1 LED with a angle of dispersion of 90 Deg at 12 inches or less might work, but it won't at 3ft. Adding optics to change the angle will drop its efficiency down. As far as solar efficiency is concerned there are three big hitters: 1) The quality of the cells themselves 2) Tracking the Sun 3) MPPT Maximum Power Point Tracking No's 2 and 3 definitely fall under your category. In general you get the biggest gains with Solar tracking in the summer and the biggest gains from MPPT in the winter, especially in warmer climates. As mentioned above, customizing the solar array for the battery (keeping the voltage as close as possible) will aide overall efficiency. #### Audioguru Joined Dec 20, 2007 11,248 One minute ago I was in the dark and was reading a newspaper with the light from a$2.00 solar garden light. It was in the sun all day (the sun is past halfway from summer to winter and was low in the sky at noon today). The light works on one charge today for about 2 hours. In the middle of summer the sun is almost directly straight up and one charge lasts all night (8 hours).The battery is a single Ni-MH AA cell. The circuit steps-up the battery voltage to light the 3.5V white LED.

They sell a solar garden floodlight for only ten bucks and it is much brighter.
The solar garden lights advertised as being "12 times brighter" are very bright but are also 12 times more expensive.

#### wayneh

Joined Sep 9, 2010
17,172
Another factor to not ignore here is temperature. A solar panel is a wee bit more efficient at lower T, but that doesn't matter so much. What matters is that a cold panel can survive using a passive concentrator - a mirror - to double the illumination and output of the cells. Very cheap way to get more juice and lower the effect tracking, or NOT tracking. In hot locations, panels need sinks and heat dissipation to survive a concentrator strategy, but heat removal may be a smaller concern for yoopers.

I guess your sponsor wants to sell outside the U.P., where people actually live , so you'd need all the heat mediation. But that might be an easier and cheaper approach than active tracking.

#### stupidlogic

Joined Aug 10, 2010
39
... but heat removal may be a smaller concern for yoopers.

I guess your sponsor wants to sell outside the U.P., where people actually live ...
You must live somewhere fairly close to the U.P.... because half of the people I know don't even know it exists. haha.

Thank you everyone for the replies. I really appreciate it and it has helped out by hearing others' ideas.

At the moment, it looks like we are going to go with NiMH batteries primarily due to the fact that overall they have a longer life if taken care of properly. I may have forgotten to mention that this device has a desired mean time to failure of greater than or equal to five years.

Anyway, after doing more research about charging methods and driving LEDs I came across two possible IC solutions that have some possible application for this project. I'm curious if anyone has used either before.

LED Driver IC: TPS61042 (http://www.ti.com/lit/an/slyt084/slyt084.pdf)

One other important factor I'm still not quite sure about is whether or not I'll need a microcontroller to run all of this or if I can just "string" it together and have it work without the need for one.

There is one feature that I've thought of to help reduce energy consumption that I believe would REQUIRE a microcontroller, but I need to do some analysis to see if it's worth the cost. That feature is a wake up interrupt of sorts. Basically completely shutdown the entire device until "first light." The solar panels should make it easy to detect when the device can charge the battery and then wake up the control unit to regulate and charge the battery.

Good idea? Or what do you think? Thanks.

#### Bernard

Joined Aug 7, 2008
5,784
Some time ago,' purchased a 10mm warm white LED as a reading light for motor home. Rated @ 220,000mCd @ 3.2V, 100mA. Has 5 chips which give a nice uniform light for a 8.5X 11 in. page, at about 15 in , 40 deg. The Electronic Goldmine, G18242, about US\$ 2.49; also brilliant white @ 265,000 mCd.
Average solar output, mid day about 4kW, overcast with drizzle, output down to 1 kW- surprised me.
Four hours per night-- a 4060 IC makes a good multi hour timer, if not using a microcontroller.

Last edited:

#### Evil Lurker

Joined Aug 25, 2011
116
What about using a super capacitor bank? No charging IC's to deal with, extremely efficient, no memory effects, and long life. I highly doubt anyone would get 5 years off a solar array on any sort of battery chemistry.