View Full Version : Solar Battery Charger
Does anyone have knowledge/experience they would like to share regarding what type of rechargeable batteries one would use in a small solar powered system. The system currently includes a propeller with temperature, humidity, and pressure sensors and two stepper motors used to align the panel towards the sun. This would be contained in an outdoor enclosure.
Here are the operating specifications of the solar panel: 16.5 Volts, 850 mAmps, 14.4 Watts; blocking diode is included on solar panel.
I have a charge controller and a small lead acid battery (12v - 1.3AH). I believe the charge controller is designed for use with lead acid batteries only, but not positive on that.
I figured lead acid due to its use in the automotive/marine industry under harsh environments, so I guess I'm curious as to what any of you would recommend, e.g. Ni-Cd, Ni-MH, Lithium, etc. ...
01-10-2008, 03:05 AM
Hey, I'm working (very slowly) on something very similar. But, I'm planning on using solar cells from some broken solar yard lights to do it... Those use AA size Ni-Cd batteries. I think these are easiest to use, but don't hold a lot of energy. I think I'm going to try using NiMh instead...
Only thing wrong with lead-acid is that it freezes in winter time (and I think D cell is the smallest they come in)...
I think stepper motors use a lot of power though, you may need a big battery to drive those. I plan on using small servos to do it...
01-10-2008, 04:01 AM
I am very interested in the field of renewable energy and have designed my own solar controller units (analog and recently with a HC08 MPU) i also have done a lot of work for calculating (and understanding the movement) the position of the sun. I also have built
a simple positioning unit for my Stirling motor (see attachment) and so on....
My last project is to build a universal motion controller (MoCo) for positioning and other purposes (see the second attachment)
So maybe working together can be very useful!
01-10-2008, 08:27 AM
Bryce, if weight isn't an issue I would stick with SLA (sealed lead acid), they are generally quite forgiving, have a good price to capacity ratio and can supply large currents when needed. Freezing is only a concern if you deep cycle the battery (drain it to the point of being nearly empty), and that can be detected and prevented. And to some degree designed out by choosing a battery capacity a certain percentage higher than whats minimally needed (For example, if you choose a battery capacity capable of supplying 3 days of power without recharge and the solar panel is capable of supplying more than what is required to operate 24 hours on a sunny winter solstice, you should greatly reduce the likelyhood of "running dry").
Paul Baker (mailto:email@example.com)
Propeller Applications Engineer
Parallax, Inc. (http://www.parallax.com)
Beau Schwabe (Parallax)
01-10-2008, 08:42 AM
If your solar cells are amorphous as opposed to crystalline they will be more sensitive to lower levels of light.· So much so, that even on an overcast·day, they will still operate at about 10-15 % of their full rating where as a crystalline cell will produce next to·nothing under the same conditions. ...· So even at 85mA to 127mA you can do "some" charging.
Beau Schwabe (mailto:firstname.lastname@example.org)
IC Layout Engineer
01-10-2008, 05:29 PM
here come "numbers":
- on a sunny day you can charge about 3.3Ah into your battery with your 15Wp panel. I have taken a 1000kWh/m2 per yaer sun power.
- Your acid battery contains 16.6Wh (12V x 1.3Ah) so you could draw 0.65W over 24h (no charging!).
- with a series regulator you could draw about 50mA but with a stepdown converter (efficiency 70%) it will be about 130mA!
Thank you all for your suggestions!
Paul: I believe the charge controller I have also protects against "running dry", but I will be sure to include the proper battery capacity in my design.
Beau: The cells are amorphous, and I was surprised to see how sensitive they are to lower light. On a moderately overcast day the cells achieved 90-95% of their full rating. Nice!
Joerg: I would be happy to help. I'm an aerospace engineer during the day so dynamics and celestial tracking are right up my alley. The only minor problem I have run into with regards to solar tracking with the propeller deals with the numerical computations forced to be in single precision (32 bits) where they should be computed in double precision (64 bits), but even with that limitation the azimuth/elevation errors are relatively small (< 1 deg) for solar panel orientation. Matrix and quaternion math is possible using a pointer/array method for position vectors and coordinate transformations. What hardware have you selected for your motion control, i.e. steppers, servos, or something else? Wonderful solar collector!
Rayman: I'll replace the steppers with servos and see what advantages/disadvantages there are for panel orientation.
01-11-2008, 02:07 AM
I have made my apprenticeship in a airplay factory and was skilled in mechanics and electronics then studied electronics, so i think we have a lot in commune. I worked on airplanes like Mirage III, Venom, Hunter, Aluette III and even on rockets like Ariane 3.
Concerning the hardware:
If the mechanics part of the positioning system is self locking (sorry i do not know the precise therm in English [selbsthemmend in German]) a stepper motor is the ideal choice, because the motor can be switched off between movements. I do the hardware by myself (PCB on a LPKF milling machine, mechanics on my turnlathe and my milling machine) the layouts i do with EAGLE. For development i use UMDL (Universal Micro Development System) www.systech-gmbh.ch (http://www.systech-gmbh.ch).
What kind of formulas will you use for calculating the sun position?
There are typically two methods I use: low accuracy and high accuracy. The low accuracy method is accurate to approximately 0.01 degrees and assumes purely elliptical motion and neglects perturbations by the Moon and planets. The calculations involve polynomial equations but need to be transformed to a local coordinate frame in order to simplify the azimuth/elevation angles the steppers/servos are commanded to control. Here is where the vector, matrix and quaternion math is used.
The higher accuracy method (VSOP87) is accurate to approximately 0.0006 degrees but involves significantly more calculations.
Astronomical Algorithms (Second Edition) by Jean Meeus and Fundamentals of Astrodynamics and Applications (Second Edition) by David A. Vallado are two good references to use.
As mentioned before, the propeller is a 32 bit microcontroller so double precision calculations are not built in. Therefore the accuracies listed above are meaningless when using the propeller. I've found my calculations are accurate to approximately 1 degree or slightly less when using the low accuracy method.
Let me know if that helps.
01-11-2008, 11:20 PM
Why are you calculating it at all. Wouldn't you use a tube/cell combination to get the exact position ?
If you are referring to a tube and photocell to keep the sun within line of sight, I decided against that approach due to cloud cover affecting line of sight. In the end I was looking for a more robust design.
01-12-2008, 12:38 AM
I recall visiting the Univ of Colorado/Harvard solar observatory in the Rockies; some 40-50 years ago, so details are vague. It used 'photo cells' to detect the sun; if cloud cover, the telescope ran slightly faster than normal sun movement. When direct sun light returned, it knew to correct the tracking by slowing down. Movement was intermittent, if I recall (maybe every minute or so). Not high accuracy, but worked for their needs. Wasn't like needing long observation times like Hubble when observing distant galaxies for hours.
Just a thought.
01-12-2008, 01:41 AM
Have you considered a Maximum Power Point Tracking scheme in this? I know you'd like to keep it simple, but if you extract a certain current from your panels and keep them at their maximum power output voltage, you can increase the power output, especially on cloudy days, by about 15-20%. Nearly all of this is done in software with minimal external components. The main external components are a couple of Analog-to-Digital chips. I'm currently working on building a large(50amp or higher) unit that I plan on selling.
I've seen this label used with professional solar charge controllers. I didn't realize the MPPT scheme was that effective; thought it was just a marketing gimmick! I will have to look into this further. Thank you for the info.
01-12-2008, 04:12 AM
I think calculating the position makes sense, because this way a lot of energy can be saved! There is non need for continuous adjusting the system. A positioning system in combination with a MPP tracker can improve the system performance dramatically.
But i think there is no need for a super precise calculation method. Have a look at the attachment where i edited a letter with my tutor about the positioning problem. It is in German, but if somebody is interested i will translate it.