Last week the kit including the BMS arrived. No circuit or wiring diagrams, just a pile of parts. Hmmm...
The BMS consists of a module for each battery cell and a "control unit" if you can call it that.
The BMS components are from EV-Power. Here is a link: EV Power cell module.
The "control unit" seems to consist of a few relays and a butchered plugpack which floats around in the box. It seems that I am supposed to connect the mains to the Vectrix charger through this box. As far as I can determine without tracing the tracks on the PCB, it is supposed to switch off the mains when the battery is charged.
BMS Module Testing
Since I am not going to entrust a $4000 battery to a pile of components which I don't understand, I started by testing the BMS modules. I hooked up the modules to a Lab power supply. A Fluke multimeter was connected via separate leads directly to the module and a second multimeter was used to indicate the state of the Solid State Relay (SSR) on the module. Like this:
It took about 2 hours to measure all 42 modules. Here is the data:
- VLL is the cell voltage lower limit (LL). Below this voltage, the SSR goes open circuit.
- Vth is the threshold at which the module starts to shunt some of the charging current through itself. I measured the voltage at which the module starts to draw more than 10mA.
- VUL is the cell voltage upper limit (UL). Above this voltage, the SSR goes open circuit as well.
- IUL is the shunt current drawn by the module at VUL.
In retrospect, I should have also recorded the shunt current at maximum allowable battery voltage which is 3.65V for the CALB CA66Fi cells. However, the current rises rapidly from Vth and reaches around 0.5A at Vth+10mV.
The only parameter which approaches a normal distribution is Vth. For the other parameters the production process seems to be less well controlled.
A significant worry is VUL. As can be seen from the measured values, VUL is 3.965V (average). This is 0.315V higher than the maximum allowed as per the battery datasheet. This is also the voltage at which the SSR switches to open circuit (OC).
However, the real worry is that the control unit does not know when the first cell has reached it's maximum voltage and the BMS module starts shunting current. By the time the SSR switches to open circuit, at least one cell will have been cooked for a while.
The maximum shunt current achieved by the BMS modules is 0.8A, some only get to 0.73A.
I don't think that shunting 0.8A when charging with 10A or 15A is going to make much of a difference to a cell.
How it ought to work
To properly manage the battery and to be able to equalize the pack, the charging system needs to have at least 3 states:
- Nominal charging, e.g. C/4 until the first cell reaches the maximum allowable charging voltage and the BMS module starts to shunt current. The chain of SSRs must go open circuit at this time.
- Equalization phase. The battery charger reduces the charging current to a value not much higher than what the BMS modules can shunt, in this case 0.8A or maybe 1A.
- Charge complete. When the battery pack reaches VBatt = N × 3.65V, all BMS modules should be shunting current and all cells should be at SOC = 100%. The charger reduces the current to 0A.
A "real" charger should of course also have a state for the case where the initial VBatt is below the lower limit, timeouts, alarms, etc...
I have the wrong BMS modules. They might be designed for cells which have a Vmax=4.0V (maybe Thundersky?). They also switch at the wrong time.
The modules don't seem to be suitable for the cells I have because:
- The modules don't indicate when the first cell has reached Vmax and shunting starts.
- The SSR switches at the wrong woltage for my cells.
- By the time the module's SSR switches, at least one cell (but probably many) will have been cooked for a while with nominal charging current minus 0.8A