Great job, specially for implementing it the analog way! Thanks for sharing.

Thumbs up for doing it analog!

Hi Mads, I don't know if you're aware of this, but it looks like your dc-converter is one of those that come aplenty from china, and which uses an LM2596, similar. If that's the case, one of the (normally permanently grounded) pins on that chip is an enable pin, which means you can turn the converter off *without* switching the input power... Also, I would consider using proper comparators instead of opamps - they mostly have open collector outputs which can be driven right to ground, and easily connected together as wired-or/wired-and..

I noticed that all opamps are used as comparators. LM339 will do a better job. It has open collector output so you can even skip some transistors. For example the NPN transistor for overcharge protection could be the output transistor of the comparator. The PNP transistor must have hfe >= 300mA/6ma = 50 minimum (because minimum sink capability of the comparator is 6mA). And of course resistors must be calculated to drive the base with 6mA. Also if you put just a tiny bit of hysteresis on U1.3 - just few millivolts, it'll switch on and off instead of being half turned on and heat up. It'll need a bit rearrangement in U1.2 and U1.3. You can switch the SMPS at the high side using PNP transistor which can be driven by U1.2 output. If you use the same transistor for Q3 as Q1 then you need R6 to limit the current to 6mA. That's (Vbatt min - Vsat comp) /6mA => (3.6V-0.4V)/6mA ~= 510 Ohm, but I think you can easily get transistor with minimum hfe = 100 and double that value to reduce power loss when LED is on. The undervoltage protection can be done by connecting output of U1.3 to negative input of U1.2 which will short R7 to the ground. If the LDR's resistance get too low during the day you can add a resistor in series to reduce power loss when battery is low. But that won't be a problem because soon voltage will rise and the undervoltage protection will disengage. One more thing - it's a nice challenge to make it all analog, but 3mA is a lot when you have just a small battery and a cloudy week. One advantage of MCU is that it can sleep and opamps/comparators cannot. If you used a mcu you can make it sleep when under voltage condition detected and wake up every minute or so to measure voltage for few microseconds. That'll improve the life of the battery a lot. And last but not least there is one big problem with not using a MCU. As you know Li-Ion batteries do not like to be tickle charged, but because your circuit draws some current Q1 will always switch on and allow some current into the battery. So a mcu can continue turning on/off Q1 simulating constant voltage and after some time (let's say 1-2 hours) it can stop charging until the voltage drops under 4V for example. P.S. sorry for the long comment. I didn't mean to ramble that much.

Thank you!!!! you have definitely helped my understanding................

amazing job!!!

nice

Thumbs up for analog approach. I started getting angry at the beginning, like: "Oh great, put a uC in that; there's no other way" :D But I kept on watching and I'm glad I did. Good project, but how good is the performance with that buck regulator, which is the voltage source? It has a potential ability to run away

Hi, good job and interesting project to try. However, do you have in mind the MCU version of the same project ? It would be fantastic. I'm thinking of an ATTINY 85. or similar. Tnx.

I am at 1:50 and paused because you are in for some trouble as the printed parts will melt and sag (mostly sag and start to deform) in direct sunlight with the solar panel there collecting the heat AND those panels will age very rapidly as in one season they turn an awful yellow and by season two it has become so bad they can't even collect the sun anymore. The best ones to use are those 5/15 watt in the glass and while they cost more they don't age like that. I know these things as it happened to me on both counts and the farther south you go it only gets worse. As for the plastic not even ABS holds up too well in the southern sun on a 105f degree day.

Very interesting project, thanks for sharing, just one ask about the battery, there may be several charges over a single day, is this type of battery ideal for this project? a lead-acid battery would be a better choice, isn't?

I think your design could be improved a bit. Life time of lithum battery is greatly depenent on the time you keep it at maxinum voltage (4.2V). The charging circuit has no hysthesis but keeps the voltage at maxinum all the time it can when the sun shines. If you added some hysthesis to the charging op amp to switch charging off when it is at full voltage you'd get more life out of it. Ideally you'd have the hysthesis only when the charging current is low ie. it would turn off at the end of the CV cycle. (This could perhaps have false triggering when the sun dies because the current naturally drops.) If the battery becomes a bit lossy at the end of the life it could draw tiny constant current all the time (if no low current turn off) which could make it overheat and burts to flames in the worst case also. You could also include a cheap true charging control chip to get off the self solution with the same problems with false charging ending when the sun goes out and it is at end of the cycle.

Did you went to school to build all these things? And what degree do you have?

I would use an ANA608 solar charge IC and a AA NIMH battery. Really simple and works well.

but u can find out when its dark, from solar panel itself. so why use a ldr ?

can you link the schematics also pls?

how do you get 10 WH? wouldn't it be 3.6v(nominal) * 2.2AH(maximum real AH I know of for 18650) = ~7.92 WH? I'm genuinely curious, not trying to bash or anything.

Hej ven. længe siden! Jeg har endelig fået bygget min 3D printer færdig her i weekenden, og det spiller bare :) Jeg vil gerne sende nogle billeder hvis du har lyst til at se :) Jeg tænkte på noget, vil du hjælpe mig med at bygge en Arduino styret musik visualizer der lyser nixie bars op?? Evt selv lav et projekt med det ;) Jeg har et link her til en indlæg jeg lavede så kan du læse lidt mere om det forum.arduino.cc/index.php?topic=428330.msg2952390#msg2952390

IF you want to make it by yourself just go to InpliX page.

I did it quickly and efficiently with instructions from Avasva website.