Electrically speaking, parallel is parallel. Whether they all parallel at a bus bar, or separate cables, or whatever, it is still electrically the same (sans some slight wiring resistance differences).
House burned down
- Thread starter Jejochen
- Start date
robbob2112
Doing more research, mosty harmless
Parallel batteries are parallel batteries if you connect them to a bus bar or to each other. If there is not some form of circuit protection device they can dump to each other.
With just two of the car/golfcart format batteries MRBF on the post should be enough. But a class T would be better.
If you have more than 2 parallel batteries then if one has an issue ALL others can dump into them. For server rack batteries there is typically a breaker on the unit, but I have seen some that have a switch that looks like a breaker and the doc says to only use it to turn them on and off, not a breaker.
For instance - if you have 10 parallel batteries connected to a set of bus bars. If one has a cell short the other 9 will dump into it. So, if the other 9 limit their current to 100amps you have 900amps dumping through a cable presumably rated at 125amps. The BMS might stop it or it might have been shorted by the bad cell.
Assuming you have Class T on all the wires connected right at the bus bar then when the 900amps flows it happens for 0.01 to 0.1 seconds as it blows. In that time the wire between the class T and battery could still potentially melt if it is long enough to have more than token resistance. This is why wires to batteries should be kept as short as possible.
Another certainty is when the fuse blows it is going to be red hot from the current * resistance for that fraction of a second. This is why the fuse holders need to be non-flamable and have a cover over them. The hope is to contain any flying red hot bits and also allow things to cool down without a fire. This is why a plain plastic cheap knockoff fuse holder is a bad thing, it will melt/deform/catch fire?
And this in IMO where the spark came from, the Mega fuse melted and caught and possibly the wire sheath as well. (yes I have said it a few times, but it is worth repeating)
Brucey
Solar Wizard
Generally we would be using a BMS with disconnecting capability per string. So the mosfets or contactor would disconnect the string in case it detects an issue. Whether this be 5kWh or 15kWh.
In the OPs case he had a single disconnect at the far end beyond the seven megafuses. It's not a common configuration and it could be argued is unsupported based on Batrium documentation. Basically he had 100kWh of strings with only a single means of disconnection.
robbob2112
Doing more research, mosty harmless
robby
Photon Vampire
- Joined
- May 1, 2021
- Messages
- 4,215
If all the batteries are connected together at a Buss-Bar and then to a Fuse, It could happen that one pack shorts out and all the other packs start dumping current into it.
That is why I do not Trust a BMS to Shutdown Power as MosFets will typically fail in a Drain-Source Short and at that point nothing is going to shut them down until they physically burn out. The Breakers on most low end battery packs are more like On/Off switches than Breakers.
IMO all Packs should have a Class-T fuse either built in (preferably) or external to the pack. Then after they connect to the Buss-bar there should be another Class-T fuse sized for the Inverters Max Current draw x 1.25 that will determine the value of the fuse needed going to the Inverter.
Last edited:
The contributing strings should be lower on the curve, so they below after the shorted string; 7 strings really does put you at a critical point compared to say 2-3 strings in that you hit instantaneous range on the damaged string faster. Once you start arcing, the fault switches to a high impedance as well, reducing the chances of being "saved" by the contributing string fuses.
We also don't know for sure if a single cell failed short or if multiple cells had dendrite issues which could escalate the issue under heavy load-- higher resistance but more cells impacted. Possible it is related to the original damaged string with one cell replaced.
That is how I saw it as well; it is possible there were two failures before the fuse blew (and we will never know for sure): inverter fault, current surge on the battery stresses one (or more) marginal cell(s) fails leading to a short, and finally the fuse tries to clear the fault.
I guess the thing needing additional information is the frequency of catastrophic cell failures. I had assumed for LFP it was less than 1:10,000,000 over a 10-year calendar life.
Where did you get those figures from?
Yes, common for MOSFETs to fail short circuit, IME.Agree on all points. A mosfet is really just a control device and not a protection device, the "switches" are unlikely to be able to break anything more than 2-3x rated current. I like the idea of a mid-string fuse to manage fault risk; worst case scenario it cuts a bolted fault voltage in half.
SWAG, but anything lower than that makes them grossly inappropriate for an automobile with >100 LFP cells in it.
robbob2112
Doing more research, mosty harmless
The AIC of the Mega fuse in question - 125a - is 2000a. So is well within the capability of it to break the short. The problem is the resistance of the fuse causes it to be red hot with 750amps in it before it blows in 0.2 to 0.01s. (fuse time/current chart as source - figure is 6x rated current)
Other strings _should_ only dump up to the BMS allow limit. Now how does the batrium control that? Seems like the disconnect was at the shunt which is above the fuses. Which means instead of 125a each they would just go full tilt and be dumping 9000amp or so that ohms law says is the max with the resistances involved with wire and all fuses. I dunno if I think things went that high or if it was someplace in between. If we figure 125a each it is still 750a.
750amps assumes only one cell shorted so the string resistance lowered by the resistance of a single cell and voltage dropped by only 3.4v. I figured conservatively.
OP said only the one fuse showed issues with being blown. So presumably it blew into a ball of melted plastic and the short stopped, but damage done.
Last edited:
I interprete this as each paralleled battery was fused before the common busbar, a shorted cell in one battery pack should have cause one or more fuse to pop
Oh, sure. I was just trying to find an upper bound in this case, with a lot of assumptions of course.
Having said that (and I don't disagree) I have shorted out packs with a JK BMS, and the BMS was able to shut off the pack without damage. I mentioned this in #245.
robbob2112
Doing more research, mosty harmless
In the other recent fire the OP had mounted breakers on each pack and only one blew out of the 6 packs.
Go read this post - it has details you may be interested in
House burned down
Fuse also needs to break only 3 volt potential, again very easy task for even something less capable than class-T. Good point here as well 3v doesn't like to arc
diysolarforum.com
Last edited:
robby
Photon Vampire
- Joined
- May 1, 2021
- Messages
- 4,215
MosFets with a proper design circuit around them are excellent devices at shutting down before they get overloaded.
What I have found out with my own designs is that they tend to have the most risk of D-S shorting when the Mosfet is driven near it's limits and starts to heat up the junctions to the point where they breakdown. For example if you design the system to shutdown at 125A and the MosFets are passing 120A for a long period of time the Junctions will heat up and eventually fuse together or blow the junctions apart (The famous pow sound followed by magic smoke). The last time I got a reality check on this was a couple of years ago when I burned out two ARF1500 Mosfets ($700 worth) by keeping the power levels high for too long while tweaking the upper cutoff point on the reflected power protection circuit.
robbob2112
Doing more research, mosty harmless
Most frequently we see pictures of BMS with the fets destroyed and possibly parts of the board missing.
Since you deal with these things from a design perspective - What does that tell you about how they are installed?
robbob2112
Doing more research, mosty harmless
One thing it seems people are missing here is the batrium design - the BMS doesn't have any ability to shutdown or limit current other than the shutmon module.
In this case the shuntmon was installed after the different battery banks were brought together. Which seems to be exactly how batrium intended things to work. This means the only thing standing in the way of the different banks dumping into each other were the 7 fuses.
The image is for a 2 battery setup.
In this case the shuntmon was installed after the different battery banks were brought together. Which seems to be exactly how batrium intended things to work. This means the only thing standing in the way of the different banks dumping into each other were the 7 fuses.
The image is for a 2 battery setup.
Last edited:
robby
Photon Vampire
- Joined
- May 1, 2021
- Messages
- 4,215
I cannot really put a definitive answer on it. Short of having a schematic for the circuit (which still might not be helpful) I would have to test the BMS with a variable load and see how it behaves at different current levels.
Zwy
Emperor Of Solar
BMS can fail stuck closed circuit.
I would not rely on a BMS to break high current.
I've read your posts about Batrium repeatedly in this thread. From what you have posted, you do not know what you are posting about.
My suggestion to you is to do some actual research instead of putting out posts that are not accurate. Or, buy a Batrium and learn how it operates, how it is set up and how a system should be designed using it.
Just stop it.
There is nothing wrong with how the OP setup his system. Each parallel battery is to have circuit protection. The OP did have this in place. The shunt trip breaker is for system shutdown in case of a critical fault where a cell has low/high voltage or other parameters that are set to trip the shunt trip.
Your posts are inaccurate.
