Solar panel and light-activated power LED

Using Arduino General Electronics
#1

Hi everyone, I know similar questions have been asked on the forum before but, differently from other topics, I am going to turn on a high-power LED trying to use the smallest amount of solar cells for weight and space constraints reasons.

I have saved four small high efficiency solar panels and a 18650 Li-ion battery from two power banks and I would like to power a small sailboat crepuscular navigation lights using solar energy.

The light uses a 3W 700mAh (2-3V) white LED which will be turned on and off using a LDR. Solar panels produce 5V 100mA, so an array of four would give out 400mAh, still less than what the LED requires.

The wiring possibilities include: - TP4056 and a boost converter (XL6009 for instance); - the new TPxx board (see pic below); - a Mppt Controller J3L2 (see pic below); - a IC QX5252F based circuit (as described here) - a direct connection from the panel to the battery and a switch activated by on the LDR to cut off solar panels and connect the battery to the light.

Which approach do you think works best? Should I add more solar panels and batteries? Thank you

s-l1600.jpg|800x800

s-l1600-2.jpg|800x800

#2

You are confusing battery capacity, measured in mAh (milliAmpere hours) with current, measured in mA (milliAmpere).

The 3W LED may have a [u]maximum current[/u] specification of 700 mA, but you do not need to provide 700 mA.

To estimate how long a fully charged battery can run a device, divide its capacity in mAh by the average current draw in mA, to get hours. Divide that by two to account for battery aging, misleading manufacturer claims, etc.

For example, alkaline AA batteries have capacity typically 2500 mAh. At 700 mA current draw, the 3W LED would run for 2500/700/2, or roughly 2 hours.

To recharge your battery, the solar panel has to provide enough mA, for as many hours, to exceed the mAh withdrawn overnight, by at least 50%. A charge controller is required.

You MUST NOT overdischarge lithium based batteries, or you will destroy them in the first usage. A discharge controller PCB is required.

Finally you need to take into account the power required, and the weather (hours of full sunlight) when making the following decision, not weight and space.

trying to use the smallest amount of solar cells for weight and space constraints reasons.

#3

…and, here are the photos [supplied by the OP]:

11c31df57f18f5463ce107e20c6e7a9c22b8b71f.jpg

e08d40383203c0544350062c2f4e537fe343a687.jpg
And, just in case you don’t know [judging by your inclusion of the LED Forward Voltage range, I’m guessing you might not], that it’s important to current drive LEDs–especially high wattage LEDs like the one your talking about. And, a Switch Mode Current Driver is the best bet for this–which, is, infact, what that QX5252F is. But, the one you will need, will have the capability of driving a much higher wattage LED. Also, if the LED doesn’t already have one, you’ll need to attach a heatsink to that LED.

Fun fact: That QX5252F actually uses the Solar Panel to sense the presence of sunlight in order to channel power to the LED at night [the function your LDR will be performing].

BTW [to the OP}: If you want to learn how to insert images into your posts, so the rest of us don’t have to download them, to see them, check this out :wink: : https://forum.arduino.cc/index.php?topic=364156.msg3749026#msg3749026

#4

Would the QX5252F alone be able to drive the LED or do I need to add a boost converter on the QX output? From the schematics it seems to be limited to 200mA in output.

#5

You need to rethink your project completely, not worry about driving the LED at this point.

#6

jremington: You need to rethink your project completely, not worry about driving the LED at this point.

Why? It's a simple crepuscular light with a battery charger using solar panels

#7

Why? See reply #1.

Good luck with your project!

#8

SeaWalker: Why? It's a simple crepuscular light with a battery charger using solar panels

Have you looked at the datasheet for the *QX5252F *: https://www.mikrocontroller.net/attachment/158139/QX5252.pdf It can only drive up to 300mA -- which may be OK, but if your LED normally runs at 700mA, then it may be a bit dim. But, maybe "dim" is bright enough for you. I don't think you can harm anything by trying [unless don't put a heatsink on the LED [if it needs one]].

Also, if you notice the operating voltage of the *QX5252F * [0.9 - 1.5V], it's not really designed for a Lithium chemistry. NiCad or NiMH, more like.

#9

You need to answer one important question before doing anything , and thats how much energy will your lights consume over a 24 hour period . That will allow you to figure out the battery capacity required, and once you know the battery capacity required, you can figure out the solar panel capacity needed. The solar panel capacity required is determined by your latitude which gives the worst case number of sun hours in the middle of winter, not allowing for cloudy days, of which you will have to estimate.

#10

I can get up to 10 hours of light, this means that the LED will be powered on for at least 10 hours, possibly not dimmed. I prefer to use a Li-ion rather than an alkaline together with the QX5252F, whose output current is definitely not enough to properly power the LED I need. Should I look for another solution?

#11

Something doesnt add up here. If your led is a 3 watt type, and works for 11 hours, that requires 33 watt hours of energy, which simply cant be delivered by an 18650 cell. The best 18650 cells in existance are only 3.4 AH , which gives approx 12.5 watt hours. Can you actually measure the leds power consumption.

#12

Something doesnt add up here.

Yep, the OP clearly doesn't think mathematical calculations and project planning are interesting or important.

One can always learn by doing and failing, albeit slowly.

#13

jremington: Yep, the OP clearly doesn't think mathematical calculations and project planning are interesting or important.

One can always learn by doing and failing, albeit slowly.

I surely did my calculations, I am just unsure which approach works best for Li-ion batteries since most solar lamps use NiCd ones. Considering three 18650 batteries joined in series I will get 6600mAh and 9.5h of autonomy which is quite sufficient, as the average sunlight hours in the period I expect to launch the boat is 12h.

I was checking out this simple circuit, as for the QX IC it doesn't require a light sensor/LDR and it could help on saving battery life blinking the LED or checking activation hours vs a RTC.

#14

Considering three 18650 batteries joined in series I will get 6600mAh

Nope. Battery capacities do not add in series.

#15

Okay then in parallel. The C rating is 1 so it’s enough for the LED.

#16

SeaWalker: I was checking out this simple circuit, as for the QX IC it doesn't require a light sensor/LDR and it could help on saving battery life blinking the LED or checking activation hours vs a RTC.

Yup...that's a cute circuit, and it looks like it might actually work --- BUT, the missing piece is [as said before], if you plan to use lithium chemistry batteries, you have to manage them properly, or you will damage them and/or cause a fire. You might get away with using the AVR µC [plus support circuitry] to manage the batteries, but you should know what you're doing.

Some battery chemistries are damaged by deep discharge -- Li being one. So, some sort of battery level detection should be added, with, a routine to not only turn off the PWM output, but also, to put the µC to sleep [with something like a periodic watchdog timer wake-up that runs a test of the battery voltage and decides whether to go back to sleep, or stay awake (this can be done on a PIC µC, but I haven't investigated AVR µCs for this sort of thing)].

Also important is to not over-charge, or charge too rapidly, the batteries [also a chemistry dependant thing -- but ESPECIALLY IMPORTANT for Li chemistry!]. Unless the Solar array's internal resistance and power output is perfectly balanced to the charging needs of the batteries-of-choice [which is unlikely because of the variability of solar power], a mere diode is inadequate for most charging scenarios. NiCads are more forgiving, in this respect--but it's still important to not over charge, or "trickle-charge" at too high a current.

#17

ReverseEMF:
Yup…that’s a cute circuit, and it looks like it might actually work — BUT, the missing piece is [as said before], if you plan to use lithium chemistry batteries, you have to manage them properly, or you will damage them and/or cause a fire. You might get away with using the AVR µC [plus support circuitry] to manage the batteries, but you should know what you’re doing.

I plan to let the AVR sleep and use the two resistors in parallel to the solar panel to generate an interrupt to wake up the uC.

Can I use this small circuit? It is designed for Li-ion and seems to have low-voltage cutoff too.