Any idea or spec info with 15Kva quattro capacitive load, I have a burnt BMS with MOSFET failed but I am trying to work out the root cause, I have no abnormal load, charging was also fine, the only thing that I can think of is about the inrush by the Quattro, is there any specification that I can work out how to strengthen my BMS design?
Make sure it is able to do the full discharge required by the inverter. (Inclusing its 100% overload ability.
In rush? With a load?
You said it was charging fine which means the system precharged and was live at some point. Maybe it was damaged in precharge?The way some bms limit inrush for precharge is voltage ramp up, others it engages a separate circuit with a large resistor.
Thanks LX, I assume highly likely the MOSFET was damaged in precharge, that was the reason why I was looking for the capacitive load value for the 15KVA quattro, I have tested the battery with 100% overload ability and it was all fine as well. I am already using a separate circuit with a 50W resistor to limit the current but I am carefully thinking to build the resistor in the BMS？
Sorry for the non-technical speak.
Not even 2 days ago I spoke to a local expert (BMS’es/Victron) who warned me about this.
He said he also lost BMS’es on larger than 5kva inverters. Actually mentioned on a Quattro.
One must pre-charge the inverters capacitors, he uses a non-BMS controller battery for this purpose as some BMS’es can be damaged.
Its a few thousand μF. The one bank has 22 * 3900 μF. So close to 123 many farrads.
Never pulled one apart so totally to count and add. 
The size you use will need to precharge in 5-7seconds.
Longer is bad (too slow) the inverter responds and voltage takes a dive as it tries to start up.
How many ohms is the 50W resistor?
when I was testing a second battery with the same setup and same procedure, the battery was in SCP but without burning out, so I assume the battery was probably sitting on a threshold or so, so if I could get a value from the capacitive load, that should be very much helpful? a 50W resistor with 47Ω so far.
1 Like
so need some mature resolution for people
1 Like
How are you determining the precharge is successful?
The 47Ω may be too high. I will go aee what i have in my toolbox tomorrow if it still has a marking on it. Fairly certain its not that high.
the battery should not be in SCP constantly or at least should be able to release from SCP very quickly? SCP is a good sign to indicate a high current but if the inrush current is way too high, the MOSFET probably would be burned straightway， is that right
SCP means something has triggered protection. So no not a normal run condition.
You don’t want that as your sign.
Usually when i precharge with a resistor i have a volt meter on the dc terminals of the inverter and hold it there untill the voltage there is at the same voltage as the battery, (5-7seconds depends on the inverter) then make the permanent connection to the battery
Obviously the bought bms do not have the knowlege of the inverter battery terminal voltage.
So for precharge time is important and so not to high a resistance to do it.
No marks on my precharge buddy, but fairly certain it is a 10/12Ω 30W
Thanks again LX, full agree with you, every time when I start up the inverter, battery will have SCP, so quite obviously the SCP came from the surge power. My set up is to use 10pcs X 48V200Ah batteries to power 3X 15KVA inverters, so that is why I am using a 47Ω resistor, sorry not being able to make it clearer before. Moving on I want to understand the exact capacitive load in 15KVA and 10KVA quattro so that I can strengthen my BMS design, including trying to improve the precharge capability of the BMS and change from MOSFET to a SSR etc
If you have your syetem design correct though, the precharge only has to happen once in a systems life.
Program the system to stop inverting before low cut off.
The other method is to use dc mppts to precharge the dc bus before restart.
Since you are designing a system that runs the precharge before closing a main contact seems the easiest to implement.
Thanks very much for your comments.
So trying to clarify, initially I had QUATTRO startup with surge power, very high inrush current(do not know this exact current value) that triggered the SCP and potentially damaged the MOSFET causing the BMS failure. As an interim solution I put in a relay and a resistor to limit the inrush current. But for a long term solution I want to evaluate how to strengthen the BMS design rather than putting in anything external, meaning I want the BMS to withstand the inrush current and not constantly triggering the SCP. From what I understood, your comment was not to go in this way but to use an external resistor to limit, is that right?
For precharge yes maybe. Since thats where you think they failed. Otherwise during normal use they are ok?
If your mosfets died during normal use then that means either undersized maybe (bank or mosfet depends on how you plan them using the cells)
Another idea for precharge is a way to pwm the MOSFETs until the precharge is done? So a bms firmware change.
The design should have nothing to do with what inverter it is connected to but rather be able to handle anything (assuming correct bank size)
Are you precharging of one battery or the whole bank?
Thanks a lot, I designed the system with 10 X 48200 batteries to work with 3 X 15KVA Quattro, but during the test process only, I tested with 3 X 48200 batteries working with 3 X 10KVA Quattro. The problem happened with two of the tested batteries during startup process. I knew I should have done the test with more batteries according to the bank sizing design, but I also was not expecting the BMS burnt out when less amount of batteries working with the inverter, if the system left untreated for a good time and some of the batteries do have the chance to go offline and if the BMS are not solid enough, I will see the risk again. That is basically why I want to strengthen the BMS design.
Tough part is getting them all synchronised.
But if the precharge is robust enough one battery can do it
exactly, this is what I was trying to do which is to enhance the precharge circuit in the BMS rather than the external device