Charge MOSFET shuts down and won't turn back on

[theshaww]

I have 6 server rack-style batteries. 2 of them have a weak cell and the 1 cell trips the BMS charge MOSFET every day before it reaches full, so there is still charge potential left on the table for the overall pack.

Once the charge MOSFETs shut down in those 2 packs, they don't turn back on, while all other batteries fill and then work normally. Those 2 won't discharge once this has happened. I usually have to reset the BMSs for those 2 to get the charge MOSFET to function again (which triggers the high current protections for those packs as they try to catch up (discharge down) to the other batteries at a lower of charge later in the day. Is this normal?

What else can I do?

 

[sunshine_eggo]

I have 6 server rack-style batteries. 2 of them have a weak cell and the 1 cell trips the BMS charge MOSFET every day before it reaches full, so there is still charge potential left on the table for the overall pack.

While this may be true, the only thing you can accurately say is "1 cell gets full first."

Can you monitor cell data? If so, can you post it?

nce the charge MOSFETs shut down in those 2 packs, they don't turn back on, while all other batteries fill and then work normally. Those 2 won't discharge once this has happened.

That's odd. Normally BMS in charge protection still allow discharge.

I usually have to reset the BMSs for those 2 to get the charge MOSFET to function again (which triggers the high current protections for those packs as they try to catch up (discharge down) to the other batteries at a lower of charge later in the day.

Are you saying that the "good" batteries are discharging into the "bad" batteries causing over-current protection?

Is this normal?

Doesn't sound like it.

Is there a reason you've omitted the brand name?

What else can I do?

Assuming you have an AiO in communication with the battery.

Disconnect data cable.
Set to USEr type.
Set absorption to 3.45V/cell (55.2V for 16S)
Set float to 0.1V below that total voltage (55.1V for 16S)

The goal is to charge more slowly at lower voltages and hopefully avoid the BMS charge protection while still attaining > 98% SoC and allowing more time for BMS balancing.

If the voltages still trigger protection, lower by 0.1V at a time until you get to 3.40V/cell absorption (54.4V). If you can't avoid over-voltage protection at that level, you may need to manually intervene by opening the case and either discharging the high cells or charging the low.

 

[theshaww]

Can you monitor cell data? If so, can you post it?

I'll go grab it.

 

[theshaww]

Is there a reason you've omitted the brand name?

No it's the junk Taicopower that I posted about a little while back.

 

[theshaww]

Are you saying that the "good" batteries are discharging into the "bad" batteries causing over-current protection?

Yes, but this doesn't happen if I catch it early enough.

 

[theshaww]

Assuming you have an AiO in communication with the battery.

Disconnect data cable.
Set to USEr type.
Set absorption to 3.45V/cell (55.2V for 16S)
Set float to 0.1V below that total voltage (55.1V for 16S)

Growatt 5000es will not charge at all unless it's >56v. I was thinking of buying a power supply to set it at 55.2v for a while and allow it to equalize for a longer period of time.

 

[theshaww]

@sunshine_eggo Here's today's latest data. I just grabbed the latest bit due to time. Sorry the date and time are wrong in the battery, but I downloaded today's activity just up until it filled and shut down. I can post a lot more later when I have time, but this shows the cell running off and shutting the pack down. I highlighted the cell where it caused the shutdown. I have cell shutdown set to 3.65v but it didn't capture it going above it.

EDIT: That pasted terribly, so I'll screenshot it.

 

[sunshine_eggo]

The other cells in that row are at least in the high 3.3XV range, so the imbalance probably isn't horrific. If you have access to the cells, it might make sense to bleed it down with a resistor/12V light bulb.

Growatt 5000es will not charge at all unless it's >56v.

Huh? Why?

I was thinking of buying a power supply to set it at 55.2v for a while and allow it to equalize for a longer period of time.

That would be a good option. There are some decent 60V/5A adjustable supplies.

 

[theshaww]

The other cells in that row are at least in the high 3.3XV range, so the imbalance probably isn't horrific. If you have access to the cells, it might make sense to bleed it down with a resistor/12V light bulb.



Huh? Why?



That would be a good option. There are some decent 60V/5A adjustable supplies.

I have definitely thought about all of this. I even think I may try to add a 20Ah cell in parallel with the 1 weak cell. I know the cell resistance would be way off, but I wonder if I could get away with that?

I was going to bleed the top cell, but it runs off at the bottom as well and is the cause of the shutdown at the bottom as well. The cell is just trash. I may still try the power supply just for fun.

 

[sunshine_eggo]

I have definitely thought about all of this. I even think I may try to add a 20Ah cell in parallel with the 1 weak cell. I know the cell resistance would be way off, but I wonder if I could get away with that?

It's possible, but it would be ugly.

I was going to bleed the top cell, but it runs off at the bottom as well and is the cause of the shutdown at the bottom as well.

That was my next question and indicates that a completely different approach is warranted.

The cell is just trash. I may still try the power supply just for fun.

Whether you bleed of with a resistor, or you hold at elevated voltage to allow the BMS to bleed off, you're bleeding off either way, and you will shift the problem to the bottom end.

It sounds to me like you either need to reduce your utilization of the battery bank, or need to replace the cell.

Would using the two marginal packs to build one good pack and have several leftover cells result in a more functional battery bank?

 

[theshaww]

It sounds to me like you either need to reduce your utilization of the battery bank, or need to replace the cell.

Would using the two marginal packs to build one good pack and have several leftover cells result in a more functional battery bank?

Good thoughts. I've thought the same, but I'd have to stop the charge cycle much sooner to keep from shutdown at top, then watch my usage more carefully to stop it from hitting the bottom as well. I was hoping to just have the BMS do that for me, but now it's locking and is unusable from time to time when I hit the top. It's strange, because it doesn't always do this.

I have 2 marginal packs, but I really don't want to Frankenstein these packs with the way they have welded the terminals together. I'd have to dremel them apart and then use a torch type method to reflow them. You might remember the pictures.

 

[theshaww]

Whether you bleed of with a resistor, or you hold at elevated voltage to allow the BMS to bleed off, you're bleeding off either way, and you will shift the problem to the bottom end.

I'm sure it doesn't help that the bad cell has been hitting 3.8v every day for 8 months also... That was the setting they had in the BMS! Trash company.

 

[sunshine_eggo]

Those pictures bring back something I had relegated in my mind to a bad dream. Ew.

FWIW, SOK, EG4, and most of the "server rack" batteries put their OVP level at 3.9V.

Using the BMS for routine termination is a recipe for failure. Repeated protection triggers will eventually damage the BMS. They are there to protect, not regulate.

IMHO:

1. get all the batteries top balanced.
2. reduce utilization AND/OR
3. add additional batteries in parallel for additional capacity.

 

[theshaww]

FWIW, SOK, EG4, and most of the "server rack" batteries put their OVP level at 3.9V.

Wow!!! That's way too close! Will and also Andy from the Off Grid Garage say 3.65 max, but I know this is a topic for debate. Once it gets to 3.65v there is only about 1% left to the top of the tank anyway, so I'd rather not risk it any further. 4.2v is only a hair away from 3.9v when charging with 20-30A per pack. Plus, it has a small delay before it shuts down and it ALWAYS slips above the threshold a bit when charging that hard.

Trust me, I don't want the BMS as my shutoff, but I don't really have a choice. It's either high cell shutdown or overall pack voltage shutdown... Maybe I'll just set the pack overvoltage to 55v and call it a day... maybe tomorrow will be better.