Graphing cell voltages & BMS balancing action

[solarsimon]

Hi,

I see so much speculation about what/how various BMSs work & worry slightly about battery capacities. So I'm considering getting some test kit to monitor every cell voltage in a 16s LifePO4 battery during charge & discharge. There's various options for taking data from a BMS & getting into Grafana etc, but I'd rather go first principles & measure independently of any BMS, for a better level of confidence & knowledge in exactly how things are working (or not).
In an ideal world I's have 32 floating/differential inputs (2 per cell) , but things start to get pricey.

Does anyone have any recommendations?
Many thanks.

 

[upnorthandpersonal]

There are some stand-alone cell voltage monitoring/logging systems out there, but since you need a BMS anyway while using the battery, why not just use that? What exactly are you trying to find out? A BMS is a pretty simple device in principle: all it does is make sure your cell voltages stay within the safe limits. Most of the functions that people discuss about are balancing features, but even those are pretty well defined these days.

 

[RCinFLA]

LTC6813-1 analog frontend will do synchronous voltage sampling on all cells. I think there is a Chinese clone similar to this chip used is some >16S BMS's.

You need to sample cell voltage simultaneously within a few hundred usecs of each other to assure all series connected cells are measured with exactly same load or charge current.

Most inexpensive BMS's just round-robin mux the microcontroller's internal ADC. There are a few BMS's that do a synchronous capacitor sample and hold. This is easy for for up to a 8S systems but gets pricey when analog muxes capable of 65 volts must be used for 16S system.

If you want a commercial BMS that will do the synchronous sampling look for units that have cell Rs readout capability. Usually doing Rs computation requires synchronous sampling so sinewave inverter 120 Hz ripple current peaks can be used to synchonously measure cell voltages and compute delta V slump cause by cell 120 Hz ripple current. 120 Hz ripple current peak value is over twice the average DC current. to inverter

DIY'er seldom have truely matched cells so the most important thing is balancing to avoid charging overvoltage BMS shutdown. Many think having same 1 KHz AC cell impedance and measuring cell capacity at low discharge current means matched cell. It does not. Matched cells have same cell terminal voltage slump, from no-load rested voltage, for same moderate (0.2 to 0.4 CA) load current, at same uniform cell temp.

Even with truely matched cells, packing prismatic cells together can cause center of pack cells to get warmer at moderate discharge currents that causes some cell mismatching due to cell temp variation.

 

[upnorthandpersonal]

If you want to do use what @RCinFLA talks about, you can get the DC2350B evaluation board for LTC6813-1 which is essentially an 18 channel monitor. It can connect to a PC together with a DC2026C board. Both are available at e.g. Mouser:

https://www.mouser.fi/ProductDetail/Analog-Devices-Inc/DC2350B?qs=sGAEpiMZZMv0NwlthflBi%252B%2Fu0oxoLo5y%2FRw%2FzT5ThW4%3D

https://www.mouser.fi/ProductDetail/Analog-Devices-Inc/DC2026C-KIT?qs=sGAEpiMZZMuqBwn8WqcFUj2aNd7i9W7uthUq4n7o1t6H1ePWQrV3pA%3D%3D

More info:

DC2350B Evaluation Board | Analog Devices

Demonstration circuit 2350B features the LTC6813-1, 18-channel battery-stack monitor. Multiple boards can be linked through a 2-wire isolated serial interface (isoSPI™) to monitor a long series of cells in a stack. The DC2350B demo board also features reversible isoSPI enabling a redundant...

www.analog.com

Linduino DC2026C Evaluation Board | Analog Devices

Linduino One controller board (commonly referred to as Linduino or DC2026C) is Analog Device's Arduino compatible system for developing and distributing firmware libraries and example code for Linear Technology’s integrated circuits. The code is designed to be highly portable to other...

www.analog.com

 

[Cal]

Probably the simplest and cheapest method is to use two MBS. The second BMS is used for measurement verification. I've used the Chargery 4S BMS as a data logger. There have been some issues with their 16S cell voltage accuracy though.

You could also build a measurement system using Arduino from scratch. I've played around with several cell measurements designs and like @BiduleOhm multiplexer/adc design.
https://diysolarforum.com/threads/diy-bms-design-and-reflection.4065/