I have made changes to the DC-DC charger and finished programming the BMS and the DC-DC charger in a first version. Both are working, however the DC-DC charger needs a few more network functions to communicate to the central system . I have also made a link between the two, the BMS can now disable the charger if required.

The change I made to the charger had to do with safety margin in terms of heat development and robustness of the power sources. I replaced the cheapish power supplies with better ones and the larger one (30A max) is fan cooled. One other reason was that the brake actuator draws around 30A when activated, which can now be delivered by the DC-DC charger without cycling the trailer battery. The DC-DC charger gets powered from the house battery.

The BMS is now finished and can balance 4 cells with a tolerance of 5mV. The low, high and variations can be programmed via touch screen or via CAN bus from the central system.

The balancing is done through charge of the low cells and can happen at any time. At the moment I have it programmed at the high end of the charge curve, when total charge current drops below 8A at 14.1V. This works very well but I don’t think that it is the most energy efficient solution. I have been experimenting and thinking a bit about this and I believe that the best solution for Lithium charging is active balancing through equalising higher to lower cells permanently. With such a concept any form of warning or cell monitoring is obsolete and the charge voltage can be even higher. The cells handle 3.8 V easy meaning one can charge with 14.4V or higher. Because the balancer will make sure none of the cells will run away this is the faster and better charging.

What one can read on a popular forum is dominated by an opinion of a group of people who follow one train of thought, which dismisses balancing as not required or too expensive or complicated. I guess this is mainly due to the lack of knowledge about products or concepts which will work properly at a reasonable cost. I found such a balancer and could not make it cheaper, however it is not a smart device and cannot talk to others, but it does its job and is very simple to install. I might replace the balancing component in my BMS with this unit.  I will also use it in the truck where I replace my LA batteries with Lithiums. With this solution I do not have to worry about charging the battery with 14.8V. The balancer will make sure that no cell runs away. Cell monitoring can be added if there is a danger of over discharge but otherwise it would not be necessary. I will connect this unit to all my banks.

dsc_2117

Above is the BMS now with all connections plugged in. I use sensor wires for the cells plus balancing wires for each cell. That makes it a bit awkward to connect. I tried without the sensor wires but got wrong readings on the cells when balancing, so I had to go back. Another disadvantage of this concept is that one needs a galvanic separated power supply for the balancing voltage with a floating ground otherwise the output cannot be connected to each cell individually. Since I use standard components and not a purpose built hardware this is all a bit more involved and needs a bit more power than a purpose build unit. However with a decent display and the CAN bus interface as well as a Real Time Clock it has a lot of features.

dsc_2116Above is the DC-DC charger. In the back one can see the new power supplies, 30A and 10A, used for bulk and an absorption like stage where the balancing happens.

dsc_2115

Both units with their remote displays. The BMS with the touch screen and the charger with a line display.

dsc_2114

BMS connected to the target battery. Lots of wires, not really very elegant but effective, better than burning off high cells. I actually get quite a bit of difference in the cell voltage due to the newly replaced cell. This is also a good reason to use active balancing. In such cases a monitor only system may not be very clever.

dsc_2118

Charge wires and sensor wires (the thinner ones) connected to each cell.

dsc_2119

This is the source battery for test at the moment. I will take it apart and build a 8 cell battery with 2 cells in parallel.

dsc_2120 Above shows the screen of a balancing process in progress. Charging with 2.3 Amps the cell will be taken to 0.01 above target before power is switched off. It will settle back a bit after a while and the then balancing will start again until the cells stay a maximum 10mV apart after a programmable waiting time. I will have to experiment with the algorithm a bit down the track. Right now it is good enough.