OpenHardware Active BMS

exco · Mar 10, 2014

So me and a friend hacked together this active BMS within a couple of nightly hacksessions
and now we're thinking about doing another few iterations and putting up a kickstarter campaign or the like
yielding in a reasonable priced open hardware active BMS.

First though I wanted to check with you and your opinions on the topic.

This is our <= 6S prototype with an at32u4 controller. It can even out the cells by moving energy from the whole pack to single
cells while loading/idle/under load. Balancing current up to 2A tested.

hooked up to a dirty 5S LiPo

with some test load to check balancing while under load

throw in your wishes/suggestions/warnings for an OpenHardware Active BMS.

dmwahl · Mar 10, 2014

I assume you're using some battery management ASIC? Or perhaps level shifting down to the AVR pins? A schematic or even block diagram would be interesting.

How are you moving charge? 2A is a lot for active balancing.

liveforphysics · Mar 10, 2014

Very cool! Can you share more about the method of operation of your balance system?

We have a few very reasonable prototype batch PCB manufacturing options folks use here with good luck.

dmwahl · Mar 10, 2014

I looked at the LTC3300 a while ago, but decided it was too complicated for my taste, especially given how easy passive balancing with bleed resistors is. Very cool if you got it working.

exco · Mar 29, 2014

We were using 2p channel mosfets for the lowest (1) cell gnd and 2n channel mosfets each for cell 1 high, cells 2-n low/high like explained here as "real switch configuration".

Coupled (to get started) with a resistor array for the microcontrollers adc pins and a galvanic isolated dc-dc step down regulator for the active balancing part.

The tests were promising - unfortunately at the time we didn't know of active balancing and chips like the LTC3300.

The battery management sytems (apart from the founding power one) I got to play with were really disappointing - most the time not even having a status button (kinda like fire-and-forget).

The active systems better yet intelligent ones are in my opinion far superior to the passive ones just wasting energy as heat.
That's why I'm trying to figure out at the moment how others think about it and build an open hardware one if there's the need.

dmwahl · Mar 31, 2014

exco said:
We were using 2p channel mosfets for the lowest (1) cell gnd and 2n channel mosfets each for cell 1 high, cells 2-n low/high like explained here as "real switch configuration".

Coupled (to get started) with a resistor array for the microcontrollers adc pins and a galvanic isolated dc-dc step down regulator for the active balancing part.

The tests were promising - unfortunately at the time we didn't know of active balancing and chips like the LTC3300.

The battery management sytems (apart from the founding power one) I got to play with were really disappointing - most the time not even having a status button (kinda like fire-and-forget).

The active systems better yet intelligent ones are in my opinion far superior to the passive ones just wasting energy as heat.
That's why I'm trying to figure out at the moment how others think about it and build an open hardware one if there's the need.

Making it open source is a great idea and will be welcomed by everyone here. I just have a couple of comments/questions...

"Wasting" energy as heat (passive balancing) is not as bad as it sounds. Active balancing systems are complex and more expensive. Consider that the losses involved in shedding a few percent of the battery's energy as heat is almost certainly less than is lost in the charger.

Regarding balancing, it sounds like you've set it up to connect the input of the DC-DC converter across the whole pack and then the output across the low cell... is that correct?

What is your standby current and how is it distributed across the individual cells? That will be one of the questions that people on ES will ask, as it's a big problem with some of the cheaper BMSs and the reason they have the nickname "battery murdering system".

OpenHardware Active BMS

exco

10 W

dmwahl

10 kW

liveforphysics

100 TW

dmwahl

10 kW

exco

10 W

dmwahl

10 kW

Similar threads