I wanted to make my own solar lamp too... but gave up after confronting reality one too many times in the last couple years. Initially it was meant to build a security light for the utility shed that illuminates all four sides using 4 LEDs total in a 4S1P configuration and operated at 350ma. Five PV purchases [from 5W to 50W], three charge controllers [from 3A to 10A], and ten lead acid batteries later [ruined quite a few 7s, 12s, and 35AHs] I gave up the idea of running anything on PV as a permanent fixture that draws in excess of 1A direct drive. Every year is the same, one too many hurricane/rain days and the batteries go kaput. If I'm not in Florida, those batteries might still be alive :shrug:
The most successful circuit I've came up with is driving a single CREE XPG at 700mA, using a LuxDrive buckpuck, switched by a relay triggered using a LTR-4206E IR detector, only drawback is that is relay is always "ON" [turns the buckpuck off] when daylights out, which siphons some power of of the battery when it should be recharging.
Sun Hours "I have sunshine all day, but why is my PV not producing?"
Simple, depending on where you are located, sunshine intensity varies alot! As opposed to measuring "how many hours of sunshine there are" Theres a unit called Sun-Hours. Where I am at, It can be sunny all day, but on average I many only receive gain from only about 4 productive hours of sunshine. PV placement and angle will vary this as well.
What this means: If you insist on increasing your gain [by an okay margin], you'll need to invest in a solar tracking device that will align your PV for you. Otherwise, come up with an average for your PV's output. Buy that PV panel, tack it on a datalogger and a near depleted battery with a charge controller, record the gains at your planned installation site for about a week per month across the four seasons. Come up with a chart for your location, then take that information and redesign your plan with respect to the load you intend to operate. You'll find the numbers will be way off than you initially hoped for.
PV Limitation
Many manufacturers do not use real sunlight to test their panels, simply because of its inconsistency when subjected to weather. Therefore, a massive xenon lamp array is used that simulates 1 sun condition as close as possible, run it for several hours, then take an average. You cannot, should not, and must not base your load calculations solely on the values printed on the solar panel. Depending on surface contamination, wiring resistance, etc. you'll be lucky to get maybe 70% of that rating if you average the gain you would receive in a day.
What this means: Your solar panel should be larger than your existing calculations, which in term may require you to upgrade your charge controller, wiring, and mounting design as needed
Lead acid Limitation: ugh... where should I start..
Regardless of their capacity, you should de-rate their performance in regards to temperature. Sitting outside means there maybe drastic temperature fluctuations from dusk to dawn. Batteries like to cozy up, and will get lazy when cold. My experiments show that if its really cold their voltages reduce and may even refuse to charge until their temperature returns to a reasonable level. AGM SLAs are considered general purpose/stand by batteries. If it doesn't say Deep cycle on them, do not use it for PV applications... unless you prefer to operate your loads to consume only 40% capacity at max. [That is, say you lug a 35AH group U1 battery home, if its not a "deep cycle" battery, you would only be able to use 40%, or 35x.4=14AH on the C/20 scale, increase or decrease according to your load accordingly.] Battery contacts subject to galvanic corrosion if you use the wrong connectors, and subject to electrolytic corrosion if placed in a humid environment without protection. They tend to sulfate if partially discharged and your PV cannot recharge it back up in time [PV surface blocked by something or the sun blocked in clouds] you you may have to do some battery maintenance on the bad days. Some Charge controllers for PVs nominally float batteries at 14.4-15V... this is nice for FLA Lead Acids. but for SLA it will cause the small ones to swell and bubble. SLAs should not float above 2.35V/cell or approx 14.1V. FLAs are cheaper, but you'll have to vent hydrogen and go at it with a hydrometer every once awhile. You also cannot lean FLAs at an angle or any angle other than perpendicular to the ground. Battery is full at 13.8V, completely dead at 10.5V, since we are running only 40% out of the battery, your usable window is even smaller: between 12V and 13.8V.
What this means: You'll need to protect your battery from the environment, additional materials like foam won't hurt, but it would increase the load on your parts budget. Lead-Acids are tolerant to mediocre charger designs, but its really more of a deadweight than anything else. Just like you, they'll need a tent and a sleeping bag on camping trips.
Runtime "I thought it said it would run 12 hours?"
In the ideal world dusk sets at 8PM and dawn arrives at 6AM, which will give it a runtime around 10 hours. However, seasonal changes on top of meteorological effects may push your runtime beyond 15 hours.. if its rainy all day [or if you find a bird nest on top of your sensor] it may cause it to run 24/7. An LVD circuit may be necessary to prevent overdischarge, most LVDs do not cut off at 12V, so keep that in mind too.
What this means: This would add more stuff on top of your circuit, and turns it tricky if you want to build a DC timer to go with it
Okay, lets see about lighting your lights,
12V 2A.... ouch, so I have to assume that it has an internal driver that requires a constant voltage input of 12Vs and it has to be be capable of supplying 2A. It might not like to work with voltage fluctuations so you'll have to introduce in a DC-DC converter.
So your layout would look something like this:
[PV] --> [Charge Controller] --> [Battery] --> [DC-DC converter] --> [Load]
The cheapest DC-DC converter would probably the open frame MeanWell PSD-30A-12, ~$25/shipped
9-18Vin, 12V 2.5A out. Lets look at the efficiency numbers:
http://www.mouser.com/ds/2/260/PSD-30-spec-34930.pdf
Efficiency 77%... just to be conservative lets say 70% without any thermal derating. Electrical efficiency in its simplest form is output over input. With a battery working voltage under load of 12.6V, your input from the battery into the DC-DC converter would actually be closer to 3A, increasing as battery voltage drops.
Assuming we are still in the ideal world, 10 hour runtime, and your battery would produce exactly what the specs read. Your outside temperature will always be 77F. You'll need a 55AH battery at minimum. Why? Isn't 3x10 only 30AH?
Here, follow me:
http://www.batteryspace.com/prod-specs/SLA_12V55Ah.pdf
Scroll down to the discharge characteristics chart under the dimensions. You see the two curved lines on the far right? Those are curves using load current of 4.95A and 2.75A respectively. If we are drawing an average of 3A from the battery, then the 3A curve should fall between those two curves. Say we want the load to cut off at Vbatt=12V, you'll see the invisible line drawn for 3A falls directly on the 10 hour vertical line. In the most ideal settings, this would still only serve as a benchmark minimum. If it was me, I'd invest in a 70AH battery, where the battery voltage is nearly flat on its way out. 10 hour runtime over 3A is about 30AH before losses. So for example lets say we'll give it a charging overhead of 5AH, then from that it is determined that we will need to charge back 35AH of juice the next morning [or else your battery is X_X by tomorrow morning]. "12V" PV are actually rated with its output voltage being about 17V. So a 10W panel does not put out 830mA at 12V, instead it puts out 580mA at 17V. Assuming 5 hours of productive sunlight in one day, (35/5 = 7) your PV array must be at least [7x17=119] 120W to break even. Since days are rarely ideal, I'd go for 150 or even 200W. If your setup site is not facing south or is tree lined, you might have to go higher.
For your system if I was to build it:
[PV]: 4x Epcom 50W $500
[Controller]: Morningstar SS-20L-12V SunSaver 20 Amp with Load LVD, $100
[Battery]: Concorde PVX-690T, $250
[DC-DC Converter]: Cincon EC9B-12S12, $50
[Load]: Your light, runtime not exceeding 10 Hours
[Misc]: Manual battery charger, meters, disconnects, fuses, grounding stuff, lightning arrestor, dusk to dawn relay, hardware, etc etc... $100
Total investment I would probably plan for would be around $1000,
My advice? Go for a smaller load like <500mA:shrug:
