EcoRenovator (
-   Appliances & Gadgets (
-   -   DIY solar power bank 5V 230Wh (

Daox 07-31-19 09:46 AM

DIY solar power bank 5V 230Wh
3 Attachment(s)
I was talking with MetroMPG the other day about the solar powered alarm clock I made a few years ago. It has really worked great, and I am very happy with it. Whilst talking with him I got the itch to do another similar project with a small scale solar something-or-another.

I thought a 5V power bank for phone or other misc USB charging would be a nice, fun and easy project. I know you can buy those gimmick solar power banks with the tiny solar panels. But, I am thinking an actual functional setup that can charge my phone at 15W (3A) nearly every day of the year. I also want it to be able to charge my wife's phone daily as well. This means it'll have to have a not tiny solar panel, plus decent battery capacity.

I first started looking at our phone's battery capacity which is 3Ah @ 3.7V or 11.1 Wh. My aim was to get at least 8 charges out of this pack. So, it needs to have a capacity of at least around 90Wh. I started looking at battery solutions. Lithium 18650 or 26650 cells seemed like a good solution. However, finding reputable cells with realistic capacities at good prices was challenging. So, I looked at nimh. Since I had such good luck with the alarm clock, and knowing they don't even need a charge controller, this seemed like a good solution. I found a nice deal on D cells that would make a 135Wh pack for $25.

Then, I remembered years ago I had bought four A123 pouch cells. These were lifepo4 cells the manufacturer rejected that didn't achieve their rated 20Ah capacity. I never had a use for them, so they've just been sitting. When I bought them I capacity tested them and most are around 18Ah. So, I have four 3.2V 18Ah batteries good for right around 230Wh. Bingo.

So, I am going to look for a solar panel to charge these up. I think I'll run them all in parallel to keep charging easy (no balancing). However, I am going to need a charge controller as well as a dc-dc boost converter to take the 3.2V up to 5V for the USB output. I'm very open to suggestions on these parts.

NiHaoMike 08-01-19 01:03 AM

From a quick search, it looks like 1S LiFePO4 BMS boards are hard to come by, but 4S LiFePO4 BMS boards are really common. If the 4 cells are closely matched, you can just go for a 4S pack, would also make the rest of the setup easier to design. Otherwise, you can use a standard 1S BMS for overdischarge protection but the trip point will be too high for overcharge protection.

For a boost converter, just find one that can supply well above what you need (I suggest 4A or more) with a low idle current. I'm not sure of any ready made modules to recommend since I generally build that kind of circuit from individual components. You can try a board for a portable USB battery, but the built in charger will not be usable without modification to limit the charging voltage.

For a charger, you'll need a MPPT step down if you're planning on using a standard 12V solar panel. (Or look into building your own with about 3.6-4V MPP voltage - that's surprisingly economical for your use case since the cost per watt of small panels is way high.) A shortcut that helps us is the fact that LiFePO4 has a pretty constant voltage over a good part of its cycle, so a 555 timer driving a buck stage with a fixed duty cycle combined with another circuit to switch it off if the battery voltage reaches 3.6V will do the trick.

Daox 08-01-19 08:57 AM

Thanks for the info Mike! I looked around a bit. I still like the idea of a 1s4p design for simplicity and forgoing any balancing issues. The batteries vary from 17.6ah to 18.6 ah.

Here is what I found on single cell charge controller chips:

I really like the idea of using a 12v panel and mppt charger because it will charge at least somewhat even during not ideal conditions. If I am going to use this year round I'll want to charge whenever possible. The LT3652 chip seems to do this and it looks like I can order a board from China for $20 or a bit less. However, finding info on the board is proving difficult.

Daox 08-01-19 01:56 PM

After a bit more looking, I am thinking more about using the CN3722 chip. It can handle a max of 5A charge current vs the LT3562's 2a. I was hoping to use a 20w panel, and this gets me close. However, I have never heard of the manufacturer Consonance. This makes me question that 5A rating. Another thing is that the controller doesn't start charging until the panel voltage gets to the set MPPT voltage. It seems to me that his would leave some power on the table, especially in my case where my battery voltage is going to be so much lower than my panel voltage.

The only other alternative is the BQ24650 which seems more complicated and all the boards are more expensive because they have more external components.

NiHaoMike 08-01-19 09:14 PM

From my experience, that sort of simplistic MPPT controller is not very good and a 555 will be way better.

At 20W and low voltage, it would probably make sense to build your own solar panel. Go for 10 or 11 cells in series (or 12 if you want the option of using conventional lithium cells later on), and then a LDO regulator with reverse current blocking set to about 3.6V would be all you need for a charge controller.

Go higher power, then just buying a premade panel would make more sense.

Daox 08-02-19 08:42 AM

Could you link me to a 555 control circuit? I'm not exactly sure how that would work. I know the 555 is a timer chip, but I have never used one.

I did order a CN3722 board for $9 off ebay. It will be here in 2-5 weeks, haha. I am a little concerned that I will burn it up working at it's max output. We will see. I will explore other options in the mean time.

I was able to find some inexpensive ($5) 6A dc-dc step up converters for the output. They even have undervoltage protection which is a great thing. I was thinking I would use two of them.

jjackstone 08-02-19 09:48 AM

Lots of ckts.

Daox 08-02-19 08:02 PM

6 Attachment(s)
I worked on the battery cells and got the pack assembled. As mentioned, these are A123 Amp20 pouch cells. They measure about 6.25" x 9" plus the connector tabs on top which bring them to about 10" tall. They are roughly a 1/4" thick.

To start, I got the connection tabs all cleaned up and oriented correctly. Then, I used shipping shrink wrap to hold them together.

As you can see, the tabs now need to be connected to create a 4p (4 cells in parallel) battery pack.

To fill the gaps, I chose to use 1/4" thick aluminum bar stock.

Two of the pieces got tapped for an M3 bolt so I can connect a wire to them easily. There are also two end plates made up with holes so it can all be clamped together with threaded rod.

And here is the pack fully assembled. Its quite the solid sturdy thing. For reference, the pack weighs just under 5 lbs.

NiHaoMike 08-02-19 08:33 PM


Originally Posted by Daox (Post 61206)
Could you link me to a 555 control circuit? I'm not exactly sure how that would work. I know the 555 is a timer chip, but I have never used one.

It's a 555 in an adjustable duty cycle configuration driving a buck converter.

Daox 08-02-19 08:42 PM

I'll have to read up on that. Like I said, I haven't used one before. I've never built a dc-dc converter either. Thats a bit over my head at this point.

Daox 08-04-19 06:36 PM

I went ahead and ordered two of the ebay dc-dc converters listed above. These are also from China. I searched for a lot of alternatives, but I was unable to find one here in the US that had the power I wanted plus the undervoltage protection that this board has.

Daox 08-04-19 07:20 PM

I am looking for one final piece before I continue with the mechanical part of this project. I would like some sort of state of charge meter. I'd like to keep it simple to eliminate phantom load on the power bank. I think something that activates with a push button would be good so it completely disconnects when not in use. It doesn't need to be an active system. However, the only other type of system I'm aware of is a voltmeter. This really isn't the greatest way of measuring state of charge, especially with LiFePo4 batteries. You guys got any suggestions for this?

NiHaoMike 08-04-19 08:43 PM

Just build a dual comparator circuit with one output to trigger around 3.15V and another around 3.45V. That will give an easy indication of low and high charge. If you want more accuracy, you'll have to use a BMS with that feature which is hard to find for one cell LiFePO4.

Daox 08-05-19 09:07 AM

1 Attachment(s)
Thats pretty much what I was thinking. Thanks for confirming Mike! You've been quite helpful on this project and I appreciate it. :thumbup:

I was able to find a nice LM3914 board which drives an LED bar graph with a push button right on it. It seems pretty perfect for the job. They have two pots, one to adjust the 'full' voltage, and one to adjust the 'empty' voltage. That is also ordered and now on its way from China.

Daox 08-06-19 02:19 PM

1 Attachment(s)
Here is a rough electrical diagram of the power bank. It's pretty simple and we have talked about lost of the parts thus far. The one we haven't yet is the switch. I'm not confident that the dc-dc converters will not suck power while not in use, so I added the switch to power them off when not in use. I imagine the charge controller is made to not drawn down the battery it is made to charge.

I think it would be nice to probably put a led on the dc-dc output as well just to make sure it is actually charging. I'll probably add that.

Daox 08-07-19 05:38 PM

2 Attachment(s)
With the electrical stuff mostly figured out for now, I am going to switch over to making the enclosure for the power bank. Thankfully, my 3d printer is big enough to handle this task. Here is my initial design. It still needs some work, but I think it looks pretty nice.

Here is an exterior shot. The USB plugs are on the right with the LED meter on the left. Still going to have to stick a switch in there somewhere.

Here is an inside shot. The larger grey boxy thing is the charge controller, and the two other smaller ones are the dc-dc converters.

Daox 08-09-19 07:49 AM

1 Attachment(s)
To keep things moving, and be able to print items before I get the electronics I had to change the design a little. I added a bunch of mounting holes to the main housing for flexibility. Once I get the electronics, I'll make up some small brackets to hold them to the housing. This will allow me to get the housing printed up. Its going to take approximately 17 hours, so I'd like to get it done sooner than later. This also allows future changes and upgrades or parts replacement with something other than the original parts without having to reprint the large housing.

Here is a picture.

Daox 08-09-19 08:29 PM

2 Attachment(s)
I also ordered my solar panel for this project the other day. I went with a 20W 12V panel from Newpowa. I got it on Amazon for just over $30. It measures about 18.5" x 14", and puts out 1.16A at 17V.

I am still thinking about how to plug this into the power bank. I'm thinking a barrel connector would probably be nice and work well. Suggestions are welcome.

Newpowa 20w Watts 12v Poly Solar Panel

Daox 08-11-19 10:00 AM

1 Attachment(s)
I got the second largest piece of this build 3d printed up. It took 10 hours to print, but its on the more dense side. For reference, it is around 7.5 x 3.5 x 1.75 inches.

Daox 08-12-19 10:12 AM

3 Attachment(s)
The largest print is now done. This one took 23.5 hours to complete. I'm pretty sure that is the longest single piece I've printed.

So, here we have the two halves of the battery enclosure box.

Daox 08-13-19 12:52 PM

2 Attachment(s)
I got the handle printed. The handle is printed a fair amount more dense to take the load it will see.

I also got it installed. It was a bit tricky getting the screws all the way in there as the main housing is pretty long, but I managed. Thankfully the inside is big enough to get the screwdriver handle into.

pinballlooking 08-13-19 01:46 PM

It is looking good so far.

Daox 08-13-19 11:11 PM

Thanks! Its been a very fun project thus far.

Daox 08-14-19 09:13 AM

6 Attachment(s)
I ended up having to reprint the bottom piece. The two end holes were 1/16" off and it just didn't line up nicely. Cut twice measure once, right? :confused: I'm still working on that. Anyway, it now assembles nicely and things line up great. It even looks pretty good IMO. :)

I put some rubber D-shaped weather stripping in the bottom piece to cushion the battery pack. Its darn near perfect. It starts mashing the rubber with a ~1/8" gap.

With the handle attached, it was time to slide in the battery pack. It fits with a little resistance. There is a bit of a gap side to side, but not bad.

Next up I put the wires on the batteries. I ended up having to drill out the holes that go from the battery compartment to the electronics compartment. I apparently didn't look up what gauge wire to use to handle the possible 10A draw. I think I settled on 14 gauge.

Here we have the unit mostly assembled and ready for electronics.

Daox 08-15-19 08:20 AM

What would you guys recommend for fusing on this circuit?

Daox 08-16-19 01:01 PM

4 Attachment(s)
I got the charge controller yesterday. This was a nice surprise as it wasn't due for another week or two.

I got a little bracket printed up for it.

And got it clipped in place.

I haven't hooked it up yet. I am still wondering what, if anything, I should put a fuse on.

Daox 08-19-19 12:00 PM

4 Attachment(s)
While waiting on the other electronic bits, I ordered and received mysolar panel. It measures about 18X14".

It is a 20W 12V panel sold by Newpowa. Operating voltage is 17.2V and 1.16A. This is what I will set the 'MPPT' controller to. It will maintain the 17.2V at whatever amperage the panel can supply.

I was concerned that achieving the 17.2V on a cloudy day was going to be an issue. The whole reason I went with the 12V panel was so I could get some charging done on cloudy days as I am only charging a 3.3V battery. But, my worries were quickly dismissed as I tested the open circuit voltage on the panel on an overcast day. I edited the brightness of this picture to accurately show what kind of light the panel was seeing. I hooked up the wire leads to the voltmeter and it was showing a nice 19.8V! So, even on cloudy days it should have no problem hitting the 17.2V. How much amperage we can get is obviously going to be low, but at least it will be something.

This is roughly how the panel will bit sitting for now. I may 3d print some sort of mount to hold it at a bit of an angle or something. We will see how it does.

Daox 08-22-19 06:04 PM

The DC-DC converters are getting close to being here. I hope they get here before the weekend so I can get this thing up and running.

I am still wondering if you guys have any thoughts on putting fuses into this setup?

NiHaoMike 08-22-19 08:35 PM

Just one fuse at the battery is all you need for your setup. If it's still 15W, a 7.5A or 10A fuse will give plenty of margin.

Daox 08-23-19 01:19 PM

Thanks Mike. The max draw will be a bit over 30w (15W per USB plug) plus the converter inefficiency.

Daox 08-24-19 12:05 PM

5 Attachment(s)
I got my dc-dc converters last night. So, I printed up a little bracket for them. It went quick and easy and they fit great.

I started tinkering with them this morning. I started by connecting the dc-dc converters up to the battery bank with alligator clips. I then set the open circuit output voltage to 5V for the USB.

I was delightfully surprised that the dc-dc converter power draw with no load is a very low ~.35 mA. This will eliminate the need to switch them off and keep things simpler. It will loose .014 Ah per day with the two dc-dc converters hooked up. I think this is acceptable.

I also hooked up an old phone to the circuit. As expected, it started charging at about 5W.

However, when I plugged my newer phone into the charger it also charged at only 5W. I specifically got dc-dc converters that can handle more power output so I can charge at faster rates. So, I need to figure out what is required to activate the faster charging. If you guys have any info on this it would be much appreciated. :thumbup:

NiHaoMike 08-24-19 02:33 PM

You'll have to short the data lines to each other to signal fast charging and adjust the voltage to a bit above 5V, probably 5.2V or so would be ideal. Also, if you're putting USB-C sockets on the charger, the CC line pullup has to be 10k.

Daox 08-28-19 03:17 PM

3 Attachment(s)
I had some sun today, so I worked on the charging system.

Here was the setup for today. The sun was come and go, but I did have a few periods of full sun.

I got things adjusted okayish. It definitely still needs some work. Under full sun I got the charge controller to charge at 2A. It was clearly capped at 2A as no adjustment would bring it higher than that.

Apparently there is a resistor that controls the max charge current. At 2A the charge controller transistor, inductor, and diode all got too hot to keep my finger on for more than a second or two. The ebay listing does say:
"Constant current charging current is set by external resistor Default 2A (3A / 4A / 5A)"
"If the charging current is greater than 2A, it is recommended to attach the heat sink on a good aluminum or heat sink installed!"

They weren't kidding. :) So, I would love some suggestions on how to do that. I'm not familiar with how you slap heat sinks on chips. I assume there is some sort of thermal paste / glue that you can use. Suggestions are welcome.

ThomSjay 08-29-19 12:17 AM

How-to video to apply tape...

Links to possible tapes. 12mm is approximately 1/2", and 25mm is about 1"

Daox 08-29-19 09:13 AM

Tape! That sounds easier than glue. I didn't even know there was such a thing. Thanks!

As for how to go about things. Is it better practice to make individual heat sinks, or one larger one? The diode and transistor are the same height, so I could combine that heatsink. But the inductor is higher, so it will need it's own anyways.

I also checked out the datasheet for the charge controller chip. It says "In constant current charge mode, the charge current is set by the external sense resistor Rcs and an internal 200mV reference, so the charge current equals to 200mV / Rcs." I traced the pin and it looks like that big R100 resistor is our Rcs. 200/100=2 so that makes sense. I'll need to replace it with a 40 ohm resistor to get the full 5A out of it. However, I am hesitant to go the full 5A because of how hot it is already getting. Should I be concerned, or is it okay to have things run that hot?

ThomSjay 08-29-19 10:49 AM

For simplicity's sake, I would use one larger heatsink, but, make sure that it is large enough (too large is never a problem....unless there are size constraints:) )

As far as heat goes, I have read that 60 degreesC, maybe 70, is about as high as one wants to go for a Raspberry Pi when overclocking. I think the same would hold for most electronics.

Daox 09-03-19 09:39 AM

I have been using the power bank as is for a few days now. I really haven't been monitoring the battery voltage or anything. It has been charging at 2A or less during the day, and charging my phone up at night. With the current settings it actually requires very little light to start charging the pack which I think is very cool. I haven't measured amperage at low light but I can only imagine it's quite low. I just grabbed another multimeter so I can look at amperage and voltage at the same time. It will be interesting to see how it does.

I also got my LED bar graph circuit kits this weekend. I assembled one and it seems to work. However, I can not seem to get it working the way I would like. It is supposed to have two modes; bar and dot mode. Bar lights up multiple lights, and dot only one. I can seem to only get it working in dot mode. This works, but is not what I was hoping for. I also have yet to adjust the circuit to the correct minimum and maximum voltages.

Daox 09-16-19 10:21 AM

3 Attachment(s)
I am continuing on with the LED bar graph not working the way I'd like. I added a symbol on the faceplate to indicate what side is charged vs discharged.

So, I printed up a new faceplate and mounting piece for the LED gauge. It looks like it'll work out pretty well.

Daox 10-31-19 08:41 AM

1 Attachment(s)
I got back to tinkering with this project the other day. I changed up the face too plate a little. I like it a bit more.

I still need to set the charge controller correctly. I overcharged the pack the other day to just over 4V. Lifepo4 is not supposed to be charged that high. I'm thinking I should hook up the dc-dc converters to some power supply and tweak it that way. I also have to swap out the resistor so it can charge at 5A vs 2A.

Daox 11-30-19 04:49 PM

1 Attachment(s)
Disaster strikes! I told my wife she could use the battery bank, so she did. But, I didn't recharge it because I wanted to monitor the high voltage settings on the charge controller.... The DC-DC converters claimed to have a low voltage disconnect built it, but there was no adjustment so I kind of doubt that... In any case, my cells got overdischarged and puffed up sadly. :( As of now, this project is as dead as my batteries!

All times are GMT -5. The time now is 01:30 PM.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, vBulletin Solutions Inc.
Ad Management by RedTyger