There has been a thread recently about a lack of USB-C enabled single cell chargers capable of going down to 500 mA for smaller cells, so I started wondering - would there be any point in developing my own?
I started mocking it up a bit in kicad, and wrote together my preliminary feature-list:
100% open source once tested, firmware and hardware
Microcontroller with USB-DFU, enabling first time flash and updates over USB without any programming tools
USB-C input, spec-compliant negotiation, up to 5V 3A
Powerbank feature, probably limited to 5V 1.5A
Simple user interface (1 button, 1-2 RGB LEDs)
500mA, 1A, 2A, 3A charge current (default to the lowest, changeable with single button taps)
Charge progress monitoring via RGB LED fading from red (empty) to green (full)
As small as possible
Focus on 18350, 18650, 14500
My current design is made around having one 18350 socket on one side, and either (pads for both are on the board) 18650 or 14500 on the other side, and to make a 3D print adapter from 18650 to 14500. Only one slot can/should be used at a time since they are all parallel.
Now, this is very much unfinished and just in a “Does it even make sense” stage. It would ofc get a 3D printed enclosure protecting the PCB from damages.
What’s your opinion on this? Does it make sense to proceed with this project (as in, is anyone interested in owning one of these)? Price will definitely be a lot steeper than for cheaper single cell off-the-shelf units from China - the battery holders alone probably cost more in small batches
Mainly trying to find a niche worth filling. I will order 5 prototypes, and maybe sell some boards at-cost to forum members - no official sale or anything is planned for now (as I do not have a webshop or registered business).
Any recommendations/features you would like me to add/things you would like me to remove?
I was going to say 3A+ is overkill for these cells, but I guess some high current 18650 cells support 1C charging.
Magnetic leads would allows charging any cell, making more use of the 3A+ charging curent, but then we need to find a simple way to make them.
I wonder if sliding contacts can be found off the shelf, instead of using two holders, which would still provide limited cell compatibility.
A ~200mA setting would allow 10440 charging.
In anycase I’m interested in the circuit and firmware details.
My original plan was to use a MAX77757 since I was planning to use that chip for other projects anyway - however it seems those chips are now unobtainium and very spendy or out of stock everywhere. Will have to switch to a different chip. Too bad, their ability to negotiate both USB-C up to 5V3A and USB-BC1.2 up to 5V 1.5A as well as older Apple/Samsung “fast charge” protocols on the input on the same chip that does the bidirectional DC-DC is pretty nice.
EDIT: Mouser has them, 4$ a chip is still steep for this application and I will not get 1000 to reach good prices ^^
I tried to find, but could not find anything so far Designing some 3d printed contraption with a spring sounds like poor reliability as well, hence the (not too great) compromise of the 2 holders.
Magnetic would be fun, but again, I struggle finding a good and scalable concept to design and make them.
I’ve been thinking about that as well, but most I can find are aimed at large cells and have limited/no support for 18350s (at least according to the not-so-reliable specs).
I used that one before and I couldn’t get 15W input an a 15W samsung charger, I couldn’t get reverse boost to work either, USB-C devices didn’t want it as source, I don’t think I did something wrong but who knows.
Oof that is good to know. Seemed like an awesome chip on paper…
Especially not having to rely on separate port controllers. Those always inflate the BOM and firmware complexity… Particularly in dual role port scenarios.
+1 on mag leads, they make for excellently versatile AND compact chargers. This could be as easy as having the holder PCB portion free of other components though mag leads work best when the charger is polarity-agnostic
Hmm, another maybe dumb idea: A tiny “core” with the electronics and an XT30 connector, and then make it modular to connect either a 18350, 18650, 14500 holder or mag leads
If all your travel flashlights are 18350 only (like mine), just pop a fixed 18350 holder module on. If you need various, make it mag.
Would also allow me to make it simple first, worry about how-to-mag-connector later (and possibly have community-developed charger bays, like a 3d printed spring adapter).
This is exactly what I have been looking for!
I only need a single slot charger for my 14500 batteries, default of 500mA is also a good choice for these smaller cells.
Currently using the free included charger from Wurkkos so this looks like a nice upgrade.
First post here, have been lurking for a while but felt I had to add my support for this project!
I think if it’s going to be a travel charger, size is a factor and when even a single cell a charger gets big enough then you may as well carry a 2 cell charger, and so on.
Personally I think magnetic is the way to go, there are a few chargers in the format you have put forward already available though may not be at the specs you state. There aren’t many out there magnetic wise and no usb c that I can see - it seems to be a format that was tried by a few companies, then given up on, not updated to usb c and so on , the olight multifuncional magnetic one looked great as it did all battery types, but for some reason they no longer seem to sell it. You couldn’t change values on it though. Maybe that type of thing but with an inline small box with the gubbins inside for charging/buttons/leds etc and usb c obviously (maybe even have 2 input wires…so if you don’t have usb c then you could use the regular one, or an adaptor?.
Nitecore had the best attempt - but no usb c again, options for charging etc if it were I’d buy one could easily have a tiny led screen like on other usb charging cables to show current etc, and maybe room for a couple of buttons and usb c to usb a rather than micro usb.
I think that is a fantastic idea. An open-source version of the Olight UC sounds fantastic to me.
Maybe like G0OSE mentioned, a small brick with a button for changing modes.
This sort of cable do with magnetic ends would solve a lot of issues with having sliders and slots and whatnot, I reckon. I would definitely be in for a few units for something like that.
Lol, when looking on Aliexpress I saw this and before realising thought what a good idea…until I saw it wasn’t a magnetic charger
Still, good idea for one!
This is an awesome idea, however with 21700 gaining popularity, I think some form of support would be great. Perhaps make a pcb compatible with 3d printed battery holders up to 10180-21700 and let the user decide which cell support to solder to each side of the board.
Magnetic options are good, but they’re also flimsy and prone to breaking wires.
Magnetic leads will require reverse polarity protection (well a normal one would as well) and idealy would work either way.
RPP can be done with these circuits : Reverse Voltage Protection for Battery Chargers | Analog Devices
For no polarity leads, like the Olight UC, the MCU would need to detect the polarity and electronically swapp the outputs if needed, then enable charging.
I’ve been messing around a bit in ltspice with 2 pairs of n-ch and 2 pairs of p-ch FETs configured as ideal bidirectional diodes
Have to do some more in-depth planning and testing but honestly that looks like it would be possible to automatically swap in hardware, at the expense of 4 n-ch and 4 p-ch FETs which do not add much cost to the BOM.
It will probably end up needing active control by the MCU, but it works better than expected already.
This sounds fantastic!
Count me in for a 14500 version.
As you are asking for suggestions: I would love a double bay version - when traveling I will often charge two cells at once and would like to have only have to use one USB cable to charge both.