Help Request! Anyone familiar with such 3S BMS designs?

[soubitos]

Hello all
I want to incorporate a BMS circuit to my charger projects.
I have the schematic of a 3S BMS circuit and i am wondering...
The design is based on 2 protection IC's for each battery, one is over/under voltage protection (in this case FS312F-G, originally was some HYCON chip with different voltage limits) the other (HY2213-BB3A) is the balancing IC.

Will it work as 1S if you only connect a battery between 0V-3.7V or 2S with the batteries in 0-3.7V & 3.7-7.4V?

Also, if i want to convert it to a 2S design to save on parts, do i simply skip the topmost part of the circuit?
There are 2S boards out there which are in fact 3S boards with what appears as a 3rd battery unpopulated part.
And i know they work as 3S if you fill in the missing parts so, just to make sure before i waste prototype pcbs and a couple of months before i got this ready, will it work? what you think?

[Daixiwen]

It will probably work as a 2S if you unpopulate the top part and connect the Vbatt to the 7.4V point. I'd strongly advise to test your BMS thoroughly before using it though, overcharged cells are no fun to deal with
You can test them by connecting cells, about at mid point (3.7V), putting the battery pack under load/charge and injecting a voltage using a power supply directly between the VSS and VDD pins of the protection circuits, over the limit. As you have a 100 ohm resistor you can do this without actually bringing the cell to a dangerous voltage level. Then check that the BMS works by stopping the load/charge.

For a 1S check the Vgs of the power mosfets that you use. On some Mosfets the voltage from one cell could be too low to control it properly, or at least to get a low Rdson. But for 1S you can do a much simpler circuit, you don't need the balancing part of course and can probably connect the protection IC directly to Mosfets. Check the datasheet for the chip, there is probably an example design for 1S.

[soubitos]

It will probably work as a 2S if you unpopulate the top part and connect the Vbatt to the 7.4V point. I'd strongly advise to test your BMS thoroughly before using it though, overcharged cells are no fun to deal with
You can test them by connecting cells, about at mid point (3.7V), putting the battery pack under load/charge and injecting a voltage using a power supply directly between the VSS and VDD pins of the protection circuits, over the limit. As you have a 100 ohm resistor you can do this without actually bringing the cell to a dangerous voltage level. Then check that the BMS works by stopping the load/charge.

Thank you for taking the time to respond... i did some further research for this schematic..
First of all, in this i replaced the original Hycon protection chips with FS312F-G which i already use in 1S protection circuits together with FS8205A dual mosfets and work like a charm. And i verified it will work as the attached design in 2S configuration without problems.
At least for the protection chips i am pretty confident they work properly.

[soubitos]

For a 1S check the Vgs of the power mosfets that you use. On some Mosfets the voltage from one cell could be too low to control it properly, or at least to get a low Rdson. But for 1S you can do a much simpler circuit, you don't need the balancing part of course and can probably connect the protection IC directly to Mosfets. Check the datasheet for the chip, there is probably an example design for 1S.

Now, due to pcb limitations i might not be able to fit the balancing part of the circuit..
In the first schematic, i have the standard 1S protection circuit (as you will find in most TP4056 charging boards these days), using FS312F-G protection circuit (also used in the above BMS) and a dual mosfet FS8205A. Bellow it I have a cascade version of this to form a 2S protection...

My problem though is, i need to be able to select between 2S and 1S...  One of the chips i will use this with is TP5100 (http://www.tpwic.com/uploadfile/2013/0410/20130410110217263.pdf) which needs a High to its CS (Cell Select) pin, to switch to 2S configuration. Since i want to keep this ability i added a simple toggle switch connected to a 5V rail. In position 1 the green LED is ON to indicate 1S configuration CS is left floating so the chip works as 1S. In position 2 a red LED is turning ON and the CS pin goes high, enabling 2S configurationn... until this point i think it is perfectly fine... The problem is, how to change the cascaded protection circuit from 1S to 2S using the same switch?

I thought to add a MOSFET from the positive of the second battery to the positive of the first battery so if it works as i have in mind, when the switch goes in 1S position, the mosfet should be fully on bypassing the second battery and driving the already lower voltage from the charger to the first battery....

Do you think this will work as intended? 

I changed the schematic to correct switch wiring

[Daixiwen]

From what I can understand from the TP5100 datasheet you shouldn't keep the CS pin floating, so you should at least put a pull-down on it.
As for Q1, are you sure it should be on that branch of the switch? It should be turned on in 1S mode. You should also have a pull down on its grid to make sure it is turned off when the switch is in 2S mode.
You need to be careful with the Vgs on that transistor too. With a fully charged cell (4.2V) and with a 5V control signal, you only have 0.8V between the grid and the source of Q1. You must be sure that this is enough to turn it on with a Rds that is low enough. If not you may have to add more control logic to drive it with a higher voltage or use a PMOS instead, source connected to the charger and grid to ground to turn it on.
A problem with this approach though is that if you have a cell connected to the 2S pins and have the switch in the 1S position (and according to Murphy's law, this will happen) then Q1 is putting the top portion in short circuit. The short circuit protection will kick in and disconnect the cell, but it has a reaction delay, and you need to be sure the high current pulse  is within the safe operating area of the Mosfet. (This applies of course also to the two FS8205A)

[soubitos]

I am afraid the overall design is turning more complex than one would like...
The original idea with one Mosfet (design considerations you mention are noted, i just laid a schematic to see what goes where  ) is NOT going to work well for more than one reasons ..
I then thought of two mosfets, both connected to the Charger (+BAT) with each one driving the voltage to either the topmost cell for 2S or directly to the first cell for 1S (the second cell left floating, even if present it would not be affected at all).
But this approach is single direction only.. so perhaps i can charge 1-2S but i will not be able to drive the charged battery current to the rest of the power path allowing for discharging of the batteries when no external power is present...

CS pin on TP5100 is left floating for single cell use, datasheet recommends connecting it to VREG pin if (permanently) in 2S configuration...

Right now i am working on an even more complex yet i think much better approach using 4 DML3006LFDS (Single channel smart load switch https://www.diodes.com/assets/Datasheets/DML3006LFDS.pdf)... i will most probably add a few extras to the project like a QC2.0/3.0 "charger" IC and a uC (might be attiny24 but most probably an atmega8). The uC will take over the "logic" for the power path switches in 1 or 2S to and from the batteries (DML3006 has both an ENable and Power Good pins which the atmega will handle combined with the TP5100 CS and PWR_ON pins. This is going way beyond the cheap/easy project anybody would be hoping for but i think, withing its specs it will be a complete tool for anybody involved with lithium batteries.. with the atmega8 present it can get the smarts to propagate battery status to "slave" devices, offer battery monitoring etc.
I have even seen several BMS solutions based on "arduino" ... makes me thinking   

[Daixiwen]

Why do you need to be able to switch between 1S and 2S in the first place? I agree it make things a lot more complicated.

[soubitos]

Why do you need to be able to switch between 1S and 2S in the first place? I agree it make things a lot more complicated.

The idea is to have "one charger" to suit more needs and be able to do more with it!
I already make a 1S module based on TP5100 which an adjustable booster output.
I take it one (or more) step further making full use of the chips potential and then adding some!
I already changed the output booster to SEPIC configuration allowing to adjust from 1.5-24V...
If you want higher output voltages (and currents) you can go to 2S without changing anything else... for under 5V output you can switch to 1S configuration.. Then it offers flexibility in input voltages too.. got 5V charge 1S, got more than 9, you can go 2S... You will be able to use common single cells (like 18650) without need for special connectors, dedicated packs etc.. you simply take two cells and make a 2S pack for this project, then take one cell out for something else maintaining almost the same functionality of the "charger".. With a uC onboard it will also be able to "communicate" with an arduino or a RPi, be used by them as their single power source enabling and disabling output booster etc... it is a project in progress and eventually, i might just make 5-10pcs just to say I did it but i really hope it will find many others interested to something like this!

If it works out, i already think of doing the same thing with a 2/3S charging IC...

It is obviously no replacement for compact dedicated charge modules but rather something a hobbyist or DIY'er would like to have on his arsenal! Or as i think of it, a charger for control freaks....

I swapped the Diodes part with perhaps a more practical one from TI.
This is how far i have taken it by now... the rest is probably on the uC side

[soubitos]

I thought they wouldn't fit ... 88x88mm just started taking shape
Went for an STM32F103 and OLED.. because.. why not?
Still a lot of work and i expect changes upon changes but that is about it
the only thing missing is the BMS....
Other than that.. it will have all the "extras" to play with 

[Daixiwen]

In principle I think this should work. But as I said, be sure to test at least that the overcharge protection does stop the charging if you inject a voltage higher than the limit between VSS and VDD on the protection chips. Even if it works on paper you never know if the components are actually working or have been damaged during soldering.