For one reason or another, a lot of us have a bunch of 18650 cells sitting around. Whether they’re for flashlights, our fancy new vape pen, remote controlled toys, or something more obscure, there is a need to charge a bunch of lithium ion cells all at once. This project, by [Daren Schwenke], is the way to do it. It’ll charge ten 18650 cells quickly using a stock ATX power supply and less than twenty bucks in Amazon Prime parts.
The idea began when [Daren] realized his desktop lithium ion charger took between 4-6 hours to fully charge two 18650 cells. With a Mountainboard project, or a big ‘ol electric skateboard waiting in the wings, [Daren] realized there had to be a better solution to charging a bunch of 18650 cells. There is, and it’s those twenty bucks at Amazon and a few 3D printed parts.
The relevant parts are just a ten-pack of 18650 cell holders (with PC pins) and a ten-pack of 5V, 1A charging modules (non-referral Amazon link, support truly independent journalism) meant to be the brains of a small USB power bank. These parts were wired up to the 5V rail of a discarded ATX power supply (free, because you can scavenge these anywhere, and everything was wrapped up with a neat little 3D printed mount.
Is this the safest way to charge lithium ion cells? No, because you can build a similar project with bailing wire. There is no reverse polarity protection, and if there’s one thing you never want to do, it’s reverse the polarity. This is, however, a very effective and very cheap solution to charging a bunch of batteries. It does what it says it’ll do, nothing more.
It is actually a scrap piece of polycarbonate hacked up with an angle grinder, bent to shape, and then spray painted on the inside. :)
I recorded this build from start to finish, but the back of my head occupies half of it.
Made from 10x TP4056 linear Li-Ion charging boards with OVP. The article left out that Daren built this for charging cells he was cycle testing, and presumably binning, not the whole pack. Simple, but overall decent hack!
You aren’t kidding about reverse polarity..
I hacked a few of those USB power banks into chargers of varying refinement
The chargers work pretty well especially considering I picked them up for $3 each and they came with two batteries as well.
I made the mistake of reversing the polarity of the battery on accident one day and it fried every transistor on the thing
NOW I put a black band on the negative side of all my bare cells with a paint pen before I charge them the first time
Well at least you didn’t have a second cell plugged in already when you did it. That would have been a good show!
I liked the idea of keeping each cell separate. It keeps the bad ones from preventing the good ones from charging, and already charged cells don’t dump half their charge into a discharged cell at high rate then either.
Marking the cells prior to charging the first time is a pretty good idea. I’ll have to steal that.
I have a 2 cell charger from China. In recessed lettering black on black it says “forbids to invert” where the cells snap in. It does not say who or in what position. The plus and minus are hard to read. Needs red paint or tape to mark plus.
I have thought about a series of cheap single cell chargers to charge a battery of e-bike size.
I built a large LiPo charging bank from similar 5V2A charging boards – they have 2 TP4056 in parallel, no needless uUSB socket or discharge protection circuit and usually come with bicolour 3mm-TH red/green indicator LEDs so you can easily poke them through a case. 30 of them on a 5V 300W PSU, little heatsinks attached to each and a fan blowing across them all (they can desolder themselves if shorted without active cooling)
I have built my own charges that has reverse polarity protection, so it is not rocket science. It checks the polarity before engaging and connecting the battery to the charging circuit.
I have a cell phone/camera battery pack charger from China. It has movable contact for lining up different batteries packages. It automatically change its polarity to match the pack. The AC to DC circuit is weak and no where near the current as it claims. It can be powered from its USB output with a USB A to USB A cable and that helps a bit.
Shit happens. If it is a Li battery, then really bad shit can happen. Charger designs based on the TP4056 should include at least a NRTL-recognized PTC for each battery, but this will not always prevent a fire for all single-fault conditions. And the battery case should be at least V-2 rated plastic, but metal is better.
There are better (more reliable) charge control ICs, but they cost 2x to 4x more.
PTC is the wrong solution. It might seem right UNTIL you understand the problem and the requirements.
The internal resistance interferes with the LiPo charging algorithms. i.e. what’s on the battery is significantly less than what your charger sees when you have a PTC in series. (Dito for a schottky diode in series)
So the battery will not be charged fully. If you opt to have the PTC on the input side, then you can still blow out the charger as the battery reverse bias the internal diodes at the charger output.
Also little known fact: PTC has significant internal resistance as it is near its rated current. If you bother to read the datasheet, you’ll also find out that it has very large tolerance ranges for the current.
A former employer makes battery chargers for the military/police/fire. We routinely used PTCs. But we used microcontrollers in lieu of the less reliable discrete charge controller ICs. While I was an employee (~6 yr), no charger failures, and the latest chargers met the CEC efficiency for battery chargers.
During this same time period, there were two customer battery packs that internally shorted and ignited. Oher than being coated black, the chargers did not fail and were ‘exonerated’ by an external lab that the customers sent the chargers. But we did do one mechanical design change as a result of these two battery failures – increased size of an internal baffle to prevent increased tracking in the haz voltage area due to battery emissions.
You are not doing it right.
Unless I’m taking pictures, I have done most of my charging in a stainless steel enclosure with active ventilation. Aka… the kitchen sink of the workshop with the window open.
I’ve probably run ~140 cells through it now. Some cells have failed, but it looks like they were actively cooled enough to prevent thermal runaway and since I was charging each cell individually, nothing bad has happened yet. I think the latter is probably the most important part. Isolating a possible issue to a single cell prevents the immediate dumping of lots of energy from one cell to another.
I went looking before I did this and saw some pretty awesome implementations of 8 cell combination charger/dischargers using 3 mosfets per channel, PTC components, etc. I also could have bought a similar commercial product for less money then though.
Definitely not idiot proof, but that wasn’t my goal. Saving time, doing it right (with regards to the charging curve and cutoffs) and not starting on fire for long enough to complete my task was. :)
I had not seen those charge controllers before. They seem neat. I wonder if the Chinese will ever start using transformers instead of instructors. If they used transformers the outputs could be totally isolated, so you could have one charger per cell, but have the string of cells in series from the same 5V power supply. Still a neat little board though. However, isn’t 2.5V too deep into the discharge curve the cutout?
These boards look like they would make neat little backups for raspberry pi’s. It would be really neat if the board had a digital output a few tenth of a volt before the cutoff so the pi could look at that and do a controlled shutdown before the battery cutout kicked in.
Switching supplies can give you isolated outputs as well, you just have to ask for it. It’s cheaper to share a leg between input/output.
The board I used isn’t appropriate for a Pi probably, as the output voltage isn’t regulated and would track between 4.2 and 2.5V, and the Pi is expecting 5V. Arduino is fine with it. Feed directly to Vcc. You can get the ones which output regulated 5V for about 3 for $10, but generally the output mosfet for the charging portion is the same as used for the switching portion. In other words, you can’t charge while discharging with those. Tried that.
Yeah, 2.5V cutoff is a bit low to prevent damage. Prevents immediate damage I suppose but leaving a cell sit at 2.5V is still bad. The lack of an ADC on the Pi kinda limits you for making up a cutoff, but I imagine a couple carefully chosen passives could scale that voltage so a digital pin starts to read low when you hit your lower cutoff.
I’ve used these boards before, Usually with the square lipo packs.
A peruse of the datasheet of the controller chip will reveal that a resistor value can be tweaked to adjust charge current and an NTC thermistor can be connected for overtemp cutout, although the details of that NTC setup are a bit scarce (maybe someone else is more knowledgable about this part?).
I usually epoxy a hestsink onto the chip too as they tend to get a bit toasty at the higher charge currents.
The holes I drilled to cool the batteries offered a mediocre improvement to the module cooling as well. Not very much, but enough that I was fine with where they ended up. You can touch them without burning yourself. I call that a win.
I was honestly more concerned with the cells staying cool. The chip would probably shut down on over temp.
Good data about the NTC and adjustable charge current.