DSG MOS fault on SOK48V100 Battery
- Thread starter stromy
- Start date
eleceng1979
New Member
- Joined
- Jun 19, 2022
- Messages
- 40
That fault tends to state that the discharge mosfet has failed. Meaning, it might now be shorted and always allowing discharge. The BMS will not be able to stop discharging due to a low V/other conditions.
I would guess arcing while discharging or turning the batteries on, then connecting large inverters while the battery is energized causing a large current is to blame. Not using a pre charge resistor/circuit.
I would guess the BMS is bad. More reading
https://diysolarforum.com/threads/bms-mosfets-explained.29318/
I would guess arcing while discharging or turning the batteries on, then connecting large inverters while the battery is energized causing a large current is to blame. Not using a pre charge resistor/circuit.
I would guess the BMS is bad. More reading
https://diysolarforum.com/threads/bms-mosfets-explained.29318/
The strange thing is,
- both charge and discharge seem to work normal
- the error goes away after some time, but comes back
Also no arcing or uncontrolled connection/disconnection or anything like that happened. The battery is in constant operation at a Victron Multiplus-II, for months without a change. Suddenly this error came up.
eleceng1979
New Member
- Joined
- Jun 19, 2022
- Messages
- 40
As a EE, if the BMS logic is written the way I think, then…
1. A shorted discharge mos would only be detected if I command to shut it off. When off it should be open circuit, and I measure such current flow to stop while measuring the voltage across its source/drain regions. Open/off = voltage present, closed/on = 0 V
2. When I request the discharge mos to open/off, it is because there is a problem with discharging. Cell too low, pack too low, too hot, too cold, current too high, etc…
3. Operation would remain normal, on the surface to the untrained eye. There are bypass diodes inside the mos which would still allow charging, and due to possibly being shorted, discharge is possible.
4. Inside these BMS’s there are two groups of mosfets, in series, reversed. One group discharge and one group charge. Each group controls one function or both off/open for battery off/lockdown. A single group failure doesn’t render the battery pack useless, it simple renders the BMS not in control of the current.
5. Either way, no matter what I or anyone says, your BMS is faulty. This is not normal behavior.
6. The fault is going away and coming back due to 2 reasons. 1, random stupid firmware glitch or 2. It is due to the discharge mos being commanded on (no fault) like normal, and then is commanded off but does not shut off due to being shorted.
7. Your BMS is faulty and possibly not protecting your batteries while being discharged which is concerning. Over discharging is just as bad as over charging.
8. Your picture states the following info. BMS has commanded discharge MOS to open, current is still flowing, therefore discharge MOS fault, alarm led on, alarm horn is sounding, yet current flows. I cannot guess your parameters, but the BMS is reacting to a parameter to protect the pack, and not able to due to being faulty.
9. All of these types of BMS use solid state mosfets to switch current off/on. These can be problematic as a single device failure takes the whole group down. More expensive BMS systems use this + an additional backup contactor. Yours does not.
1. A shorted discharge mos would only be detected if I command to shut it off. When off it should be open circuit, and I measure such current flow to stop while measuring the voltage across its source/drain regions. Open/off = voltage present, closed/on = 0 V
2. When I request the discharge mos to open/off, it is because there is a problem with discharging. Cell too low, pack too low, too hot, too cold, current too high, etc…
3. Operation would remain normal, on the surface to the untrained eye. There are bypass diodes inside the mos which would still allow charging, and due to possibly being shorted, discharge is possible.
4. Inside these BMS’s there are two groups of mosfets, in series, reversed. One group discharge and one group charge. Each group controls one function or both off/open for battery off/lockdown. A single group failure doesn’t render the battery pack useless, it simple renders the BMS not in control of the current.
5. Either way, no matter what I or anyone says, your BMS is faulty. This is not normal behavior.
6. The fault is going away and coming back due to 2 reasons. 1, random stupid firmware glitch or 2. It is due to the discharge mos being commanded on (no fault) like normal, and then is commanded off but does not shut off due to being shorted.
7. Your BMS is faulty and possibly not protecting your batteries while being discharged which is concerning. Over discharging is just as bad as over charging.
8. Your picture states the following info. BMS has commanded discharge MOS to open, current is still flowing, therefore discharge MOS fault, alarm led on, alarm horn is sounding, yet current flows. I cannot guess your parameters, but the BMS is reacting to a parameter to protect the pack, and not able to due to being faulty.
9. All of these types of BMS use solid state mosfets to switch current off/on. These can be problematic as a single device failure takes the whole group down. More expensive BMS systems use this + an additional backup contactor. Yours does not.
Hi eleceng1979,
thanks for looking into this.
The status picture is during normal (light) discharge. So at the time of the picture, the discharge MOS is supposed to be closed/conducting?
However i do not know the behaviour of the BMS. The fault could be still there from something that happened much earlier.
I cannot see anything unusual in the parameters or status that would trigger a BMS disconnect action. The Victron Multiplus-II inverter can only draw max 3000W, which is well below the 100A rating of the battery. According to my metering, it never went above 2000W.
So it seems, i need a new BMS.
thanks for looking into this.
The status picture is during normal (light) discharge. So at the time of the picture, the discharge MOS is supposed to be closed/conducting?
However i do not know the behaviour of the BMS. The fault could be still there from something that happened much earlier.
I cannot see anything unusual in the parameters or status that would trigger a BMS disconnect action. The Victron Multiplus-II inverter can only draw max 3000W, which is well below the 100A rating of the battery. According to my metering, it never went above 2000W.
So it seems, i need a new BMS.
Attachments
eleceng1979
New Member
- Joined
- Jun 19, 2022
- Messages
- 40
So, a few things.
1. A victron 48/3000/35-50 inverter manual states min battery size is 100ah, which you are at, for the reason stated below
2. It can provide a peak power output of 5500W, a max efficiency of 95%, which could worse case put you about 5789W, when around battery volt of 52 would put you roughly at 111A, which is over the BMS first stage shutdown. Not sure if you have such loads, but is a possibility.
3. Due to tolerance %/calibration issues, it is plausible that operation near such limits could happen prematurely
4. Knowing the fault comes and goes seems to indicate that an event is trying to stop discharge, and then returns to normal conditions, which hints at possible settings being exceeded
5. Knowing that the fault does come back seems to indicate shorted discharge mos/faulty BMS
6. Sorry, many assumptions about loads/usage that I don’t know. Having the right situation to cause such high currents doesn’t mean it is happening in reality. Being at min aH size, plus the BMS fault alarm, plus a potential shorted mosfet seems to summarize up to quickly rising currents, over the BMS rating (don’t know what it is). Your settings show a 110A for 1S alarm and 150A for 0.5S. That is more than enough current/time to kill a bank of 100A mosfets IF BMS is only 100A rated.
7. A general rule of thumb is 100aH per 1KW of inverter for this reason. Most inverter manufacturers state a range, but this doesn’t consider the battery chemistry. An old LA could care less on such possible abuse, LI BMS that is scanning at 1uS intervals would self destruct/alarm. I have literally, purposely used 2 LA batteries in series to stick weld a repair in the field with no welder. LI would have zero chance. Just showing the difference in sensitivities and how using LA thinking of robustness doesn’t apply to LI due to all of the solid state BMS failure modes.
8. Or it is a defective BMS that didn’t survive normal operation
1. A victron 48/3000/35-50 inverter manual states min battery size is 100ah, which you are at, for the reason stated below
2. It can provide a peak power output of 5500W, a max efficiency of 95%, which could worse case put you about 5789W, when around battery volt of 52 would put you roughly at 111A, which is over the BMS first stage shutdown. Not sure if you have such loads, but is a possibility.
3. Due to tolerance %/calibration issues, it is plausible that operation near such limits could happen prematurely
4. Knowing the fault comes and goes seems to indicate that an event is trying to stop discharge, and then returns to normal conditions, which hints at possible settings being exceeded
5. Knowing that the fault does come back seems to indicate shorted discharge mos/faulty BMS
6. Sorry, many assumptions about loads/usage that I don’t know. Having the right situation to cause such high currents doesn’t mean it is happening in reality. Being at min aH size, plus the BMS fault alarm, plus a potential shorted mosfet seems to summarize up to quickly rising currents, over the BMS rating (don’t know what it is). Your settings show a 110A for 1S alarm and 150A for 0.5S. That is more than enough current/time to kill a bank of 100A mosfets IF BMS is only 100A rated.
7. A general rule of thumb is 100aH per 1KW of inverter for this reason. Most inverter manufacturers state a range, but this doesn’t consider the battery chemistry. An old LA could care less on such possible abuse, LI BMS that is scanning at 1uS intervals would self destruct/alarm. I have literally, purposely used 2 LA batteries in series to stick weld a repair in the field with no welder. LI would have zero chance. Just showing the difference in sensitivities and how using LA thinking of robustness doesn’t apply to LI due to all of the solid state BMS failure modes.
8. Or it is a defective BMS that didn’t survive normal operation
Re 1: The Victron Multiplus-II 48/3000 is used in ESS (grid tie) mode (ACin). The output above 3000W is only possible in other modes (via ACout). These are not in use. The whole system is and was never nowhere near its max capacity.
Re 6: These are the settings from the vendor SOK / Current Connected. If these currents were exceeded, i would expect an actual warning or cutoff of the BMS. That never happened. The fault event seems not related to an actual overcurrent event, but to some failed internal selftest? Also the BMS is meant to protect from overcurrent events, so its supposed to not get permanently damaged, even if an overcurrent event would have happened.
Re 7: The sizing of the battery was for the needed capacity. Inverter sizing was the smallest available model.
Re 8: Just got the info a replacement BMS is sent out to me.
Thanks
Re 6: These are the settings from the vendor SOK / Current Connected. If these currents were exceeded, i would expect an actual warning or cutoff of the BMS. That never happened. The fault event seems not related to an actual overcurrent event, but to some failed internal selftest? Also the BMS is meant to protect from overcurrent events, so its supposed to not get permanently damaged, even if an overcurrent event would have happened.
Re 7: The sizing of the battery was for the needed capacity. Inverter sizing was the smallest available model.
Re 8: Just got the info a replacement BMS is sent out to me.
Thanks
Similar threads
