Top Balancing a large bank

[codfish]

Balancing the sets the way you are suggesting would make things much easier , and also have the batteries in service much sooner . Would you suggest using the generator to supplement power to make sure I'm hitting the target voltage daily, or just let them cycle normally with the solar ?

Just let them cycle. You are only losing a small % of capacity at the top above 3.4v/cell. Balancing just lets you use more of that. I assume your battery KW are sized so that you don't anticipate running them to a low SOC normally, so that lost % really doesn't impact you.

 

[DerpsyDoodler]

The reason there is not a lot of information is because to do it quickly, requires a lot of Amps and the cost of high current power supplies is not cost effective for something that might be done once or twice over the life of the cells. Follow the advice above with patience and you should be okay.

he could purchase 2 or 3 cheap power supplies for use in parallel. 20a is better than 10, 30a is better than both.

 

[Barbuda Cottages]

Just let them cycle. You are only losing a small % of capacity at the top above 3.4v/cell. Balancing just lets you use more of that. I assume your battery KW are sized so that you don't anticipate running them to a low SOC normally, so that lost % really doesn't impact you.

That's correct , worst case they may go down to 25%, my calculations have them running between 40-85 % most of the time . Trying to get the longest life out of them .

 

[codfish]

That's correct , worst case they may go down to 25%, my calculations have them running between 40-85 % most of the time . Trying to get the longest life out of them .

That's what I suspected. You could start even lower than 3.4v/cell to avoid any HV disconnects. Say 3.375v/cell. Just make sure that balancing kicks in around whatever cell voltage you pick. I assume you'd like to get them up to capacity before the "season" starts, but you've got months to creep up to that.

While you are in this process, you might want to set the "absorb" time longer than normal for LFP ( often very short). That will hold the pack at the higher voltage for a bit longer, letting the balance circuits do their work.

 

[Barbuda Cottages]

160 cells makes 10x 16s batteries.
Odds are you get a few less than awesome cells.

Here is roughly what I did with mine...
While the courier waited(gave him a tip for his trouble)
Before you accept delivery.
Inspect the boxes Including the bottoms.
Weigh the boxes.
Verify with your nose that you don't have a leaky cell(the electrolyte has a terrible chemical smell and its toxic).
If all present and correct take delivery.

Open the boxes one at a time and inventory the contents.
Make sure each individual cell is free of physical defects.
For each cell, check the voltage and record it along with a serial number on masking tape that you stick to the cell.
The cells should all be within millivolts.

Do you have an internal resistance tester?
Remember at best a battery has the capacity of its weakest cell.
Please report back on the results of the delivery check.

I received all 80 boxes and if I was in the States or Canada I probably would refused the delivery . The plastic pallets that the boxes were packed on both collapsed and more then 20 of the boxes had damage to them . I have to say that I was very impressed with the foam packaging around the batteries , luckily there wasn't a scratch on any of the batteries !

 

[John Frum]

Glad the cells pass visual inspection.

 

[snoobler]

160 cells.

Frick.

How many BMS?

 

[DerpsyDoodler]

Awesome news!

 

[Barbuda Cottages]

That's what I suspected. You could start even lower than 3.4v/cell to avoid any HV disconnects. Say 3.375v/cell. Just make sure that balancing kicks in around whatever cell voltage you pick. I assume you'd like to get them up to capacity before the "season" starts, but you've got months to creep up to that.

While you are in this process, you might want to set the "absorb" time longer than normal for LFP ( often very short). That will hold the pack at the higher voltage for a bit longer, letting the balance circuits do their work.

Just finished checking the voltage on all the cells , 159 of them are between 3.288 to 3.296 volts , one is 3.346 ? Haven't checked IR yet .
When you say to set the absorb time a little longer , what would you suggest ?

 

[Barbuda Cottages]

160 cells.

Frick.

How many BMS?

I have 10 BMS's to connect .

 

[snoobler]

Glad to hear it.

I'll second the recommendation to just use them. Here's what I would do:

Measure voltage (and IR if you have an impedance tester)
Sort all cells high to low voltage
Use 16 consecutive cells from the list. Build them ensuring the BMS is set to balance during charge only and at or above 3.4V.
Parallel 10 batteries.
Charge to 56V absorption.
If BMS trips, lower absorption voltage until no BMS trips.
Use and cycle them, monitoring for outliers.
Raise absorption voltage 0.1V/week until target absorption is reached.

I just skimmed this thread, so if I'm repeating stuff, please forgive me. The prospect of dealing with 160 cells is triggering some PTSD...

 

[squowse]

3.3v Switch mode power supply is cheap. I've got a 50A one I've modified so it can deliver between 3 to 4 volts adjustable. Not had a chance to use it yet, still waiting on cells

 

[snoobler]

3.3v Switch mode power supply is cheap. I've got a 50A one I've modified so it can deliver between 3 to 4 volts adjustable. Not had a chance to use it yet, still waiting on cells

160 * 280 * 50% SoC / 50A = 448 hours of charging to top balance 160 cells at the typical 50% SoC.

448 hours = 19 days assuming everything goes perfectly.

 

[John Frum]

Just finished checking the voltage on all the cells , 159 of them are between 3.288 to 3.296 volts , one is 3.346 ? Haven't checked IR yet .
When you say to set the absorb time a little longer , what would you suggest ?

That cell at 3.346 bugs me.
After the ir test do you have the ability to capacity test individual cells?
I would be interested to see the capacity of the odd one vs one that is not an outlier.

 

[codfish]

Just finished checking the voltage on all the cells , 159 of them are between 3.288 to 3.296 volts , one is 3.346 ? Haven't checked IR yet .
When you say to set the absorb time a little longer , what would you suggest ?

I'm not sure it matters. Just use them. Keep the charge voltage below the HV BMS trip and enjoy no generator noise.
(Increase the charge voltage as you can.) But... remember this whole top balance exercise is to prepare the packs for the days when you need to dig deep into the SOC. If you never need the KW, then it mostly doesn't matter.

 

[snoobler]

That cell at 3.346 bugs me.
After the ir test do you have the ability to capacity test individual cells?
I would be interested to see the capacity of the odd one vs one that is not an outlier.

+100

Suspect that one didn't get processed like all the others OR it has very high IR. If they discharge them to a target voltage, and that is high IR, I wouldn't be surprised by a voltage bounce. Cell could be fine, but if the terminal wasn't properly fastened, it hit target voltage sooner than it should and bounced.

So tempted to test that one cell....

I'm not sure it matters. Just use them. Keep the charge voltage below the HV BMS trip and enjoy no generator noise.
(Increase the charge voltage as you can.) But... remember this whole top balance exercise is to prepare the packs for the days when you need to dig deep into the SOC. If you never need the KW, then it mostly doesn't matter.

That one cell is likely at dramatically higher state of charge than the others meaning the usable capacity of that 16S battery is dramatically lower than the other due to absurd inter-cell imbalance.

I would take that one and the 15 other highest cells and set them aside. Leave those 16 wired in parallel together for an entire week. That would get them within 5-10% of each other. Build the battery and add it in parallel to the other 9 when they are at similar voltages.

 

[Luthj]

With an immediate need of the batteries , and functional BMS I would follow the advice of everyone else here and just build the packs and use conservative charge voltages until you get everything into balance. If needed several long runs of thin wire properly sized for a 3 V supply, can be used to pull down any runner cells to speed balancing.

If You are lucky the vast majority of the cells will be within 10% SOC of each other.

 

[DerpsyDoodler]

I have 10 BMS's to connect .

Before you start assembling, scuff terminals with 600 grit wet/dry and wipe with isopropyl. Before final assembly, add a thin layer of oxguard then hit with 600 grit wet/dry again.

 

[John Frum]

Suspect that one didn't get processed like all the others OR it has very high IR.

Hope its just that a cell that the vendor tested.

 

[MrM1]

Fascinating build @Barbuda Cottages . Would love to see pics of the whole system and history of it. Do you have a build threaded or a blog/vlog of the power generation for the property?

Good luck with the 160 cells. Following the advice of the collective wisdom here as I wait for my 16 cells.