Multi Cell Lithium Charger from a single 5V supply.

[beanflying]

Just finished a prototype of a simple multicell LiPo charger which is part of a larger project. Might be of some use to others and the DC-DC 1W modules are worth a look for other projects too. There is certainly lots of alternate solutions that work some of which I own but this design was because I had the modules in stock. Also the cost of bits are around $10 AUD for a 4 cell charger.

There is no circuit diagram for it at this stage nor do I really want to rehash BS of BMS boards it is a low current charger that is all!

However for those who are not aware due to the normal series connection of the cells for charging some form of isolation or floating ground is needed on the charging circuit to enable cells to be charged individually and maintain less than 4.2V/cell (puff voltage) as the others finish topping up.

If you want a Rapid charger for multicell packs then there is plenty on the market already this is a 'float' style charger for circuits under 50mA draw.

Basic Specs.

Maximum charge current 180mA with other steps of 120, 60 and 30mA. The 180mA max is governed by the 1W DC-DC.
Input voltage range of 4.5-5.5V DC.

Basic Blocks and Components.

5x7cm Protoboard (I had a stack of rubbish ones already)
B0505S 1W Mornsun 5V DC-DC isolated converter. http://www.mornsun-power.com/uploads/pdf/B_D-1W.pdf
TP 4056 based single cell Lithium charger board (cutdown). https://dlnmh9ip6v2uc.cloudfront.net/datasheets/Prototyping/TP4056.pdf
3 Pole DIP switch
6.8K, 3.3K, 10k and 20K 1% 1206 Resistors but through hole would be fine.
Stash of wire, connectors etc as per the photos.

Construction Notes.

Take the evilbay TP4056 board and chop it as per the first photo also remove the current set resistor. The board is double sided with the lower surface being a ground plane for most of it. One side of the 0.4 ohm resistor is Vcc and In- and Bat- are common. Bat+ is obvious. Using some cutdown angled header pins I effectively staked the board to the 5x7 board. If you use double sided protoboard raising the board a touch or some kapton tape would be in order.

The surface mount resistors fitted nicely between the bent over DIP switch pins and while it won't win any beauty contests it works nicely. The switches short or bypass the resistors to limit current and the 6.8K is unswitched. The formula for calculating the resistors is in the datasheet above but 6.8k = 180mA, +3.3K = 120mA, +10k = 60mA and +20K = 30mA

The pad the normal current set resistor goes to runs through the resistor chain to ground so 6.8k to 40k max.

Output wiring/header loops or shorts the Bat+ of the first cell to the Bat- of the second cell. In the case of the shown 3 cell Bat+ of the second board loops to Bat- on the third board.

Minor Issues!

Output Ripple of the Mornsun modules varies a lot so some playing with Caps to improve that. Output Ripple on the chargers in use is virtually Nil.
The Mornsun Modules generate a little heat so standing vertically will help dissipation. 70% efficiency but still up to 0.3W.
 
If anyone is really keen I can throw a wiring diagram together but the App notes above and photos should be ok to have a go for most.

Only 7 or 8 more boards to go 

[Ian.M]

That seems like a rather complex way to do that, although you get 11/10 for modularity and creative use of EBAY eCrud.

Personally, if I had to do that, I'd start with a boost converter that does CC and CV modes, and add a MCU to supervise the cell voltages and shift the boost CV setpoint and/or trigger converter shutdown at full charge.   Balancing would be done by monitoring each cell voltage and, for each cell PWMing a MOSFET + series resistor to shunt charging current round the cell when its reached its full charge voltage (or to initially discharge all cells to the same voltage).  When all cells have reached the desired voltage,  either drop the boost output to a safe float voltage or shut it down.   

[beanflying]

I fail to see how programming a Micro to do this job and raising the discrete parts count by several times is less complex, cheaper or easier? 

My larger job entail currently 15+ LiPo cells of different capacities and over 20 (possibly toward 30 later in the year) in the next few months. As part of this job there will be battery switching and reduced voltage charge points to extend the LiPo's lives so a micro may be useful then.

Going into a production job or if space is an issue then dedicated PCB and discrete components might make sense but for a one off modular charger this made sense to me and as my nearest electronics store is 3 hours (or 3 weeks) away use what you have 

[soubitos]

Would this https://www.tindie.com/products/11347/ cover the functionality demanded?

[mikerj]

Would this https://www.tindie.com/products/11347/ cover the functionality demanded?

That is just a single cell charger, based on the same IC used in the eBay charger modules the OP used.  It doesn't help with charging series connected cells.

[soubitos]

The whole approach is somehow off then... first of all, i thought OP wanted to charge 4 cells i didnt understand he wanted to use them at the same time so it would matter if they are in series or parallel...
If you want to charge a 4S battery pack, instead of hassling with single cell chargers with dedicated isolated power supplies, you'd be way better off using a 4S charger and a BMS/protection board, especially if you are going to be using the cells while charging them more or less like a laptop battery is charged/used.
The total cost would possibly be a little higher(or about the same really) but the result much safer and better.

[beanflying]

I will be charging several packs from 1S to 5S and maybe a 6S pack totaling 20-30 cells. The batteries are LiPo (350-5000mAh) and made into packs so the balance leads are permanently connected to them. The major requirements are continuous power with no mains connected (budgeting 7 days) for transportation and extremely low noise supply hence the battery supply and isolation from the mains.

BMS is not generally needed and is one of the BS non facts of LiPo battery use created for the consumer market. LiPo's correctly charged don't smoke and unless you are a dill and run them over flat BMS is a bit of junk getting in the way in a lot of cases and another thing to potentially go wrong. Flat batteries don't magically go up in flames for the lack of a BMS they just die quietly.

Currently playing with the chip enable pin on the TP4056 boards with a voltage sense circuit so I can monitor each cell voltage and float them around the 3.6-3.7V range which helps extend their useful life with an override to allow 4.2V charge for extended use off charge.

I have several Hyperion multicell chargers already and they won't do what I need and they are used for some of my other toys

[uer166]

BMS is not generally needed and is one of the BS non facts of LiPo battery use created for the consumer market.

In the context of multicell chargers that balance during charge cycle? Or do you mean in the general case of single-voltage charging of whole pack? If latter please share source of that statement, I'm very curious.

[beanflying]

Chargers with reliable cutoffs at 4.2V/cell don't blow up LiPo's so no need for a BMS. If you don't use more than the capacity of the LiPo ie don't take it below 3V/cell then you don't need a BMS.

The reason for BMS boards is for consumer devices (idiot proof) IMO and for those who can't manage their cells within 'normal' operating parameters.

In a lot of years of using LiPo's at 5-100A+ in R/C I am yet to have one catch fire or come down dead (yes my ESC's have adjustable low voltage cutouts some are set to 2.9V/cell) from running them to low. The couple I have had puff were samples of early crap Chinese (8-10 years ago) and a couple of my good quality 6S 5000's in my Trex 600's where the peak currents are huge. No BMS will stop a battery puffing!

So BMS for a lot of uses are BS.

[uer166]

Ok so you are talking about multicell chargers that fill in the BMS's usual job of balancing, which seems correct.

If, however it is a series pack that's charged from one voltage like say in laptops or EVs, a BMS is absolutely necessary, since cell unbalance will eventually cause some to drift higher than 4.2V, and others lower over many cycles (even if overall voltage is number_of_cells*4.2V), damaging cells and potentially creating a fire hazard.

[beanflying]

Ok so you are talking about multicell chargers that fill in the BMS's usual job of balancing, which seems correct.

If, however it is a series pack that's charged from one voltage like say in laptops or EVs, a BMS is absolutely necessary, since cell unbalance will eventually cause some to drift higher than 4.2V, and others lower over many cycles (even if overall voltage is number_of_cells*4.2V), damaging cells and potentially creating a fire hazard.

Wrong!

A BMS's job is not cell balancing or charging it is cell protection only.

There is no circuit diagram for it at this stage nor do I really want to rehash BS of BMS boards it is a low current charger that is all!

BMS boards may lockout a cell at 4.2 (or under voltage) but they give zero indication of the other cells under that charge point which remain below full capacity when a cell hits those limits. Dumb Dumbs on Youtube  using BMS boards to charge LiPo's are idiots and have helped partly to fuel this misconception about them being 'balancing chargers'. LiPo's are simple to charge to a point so while a raised input voltage passing through a BMS adds charge to the cell/cells it is not generally controlling the charge or fully end point balancing the pack in any way.

https://youtu.be/KYnxmXvkMYE