LFP mobile House bank - xSyP serial vs parallel design thoughts

[john61ct]

At a low voltage like 4S, variations in voltage would be easy to spot at the string-unit level.

No, because the strings are all in parallel, so all the *string* voltages are exactly the same.

> The ONLY way you can tell by voltage what is going on is to separately monitor every single cell in a pack made of series strings with no cell-level parallel connections.

But there are plenty of protective devices that monitor mid-point battery bank voltage, given two equal paralleled blocks on both sides.

You can set the delta sensitivity, and if the difference goes beyond that, it takes that bank / pack offline, or raises an alarm, or starts the shore power charger, or whatever you want.


> Way harder to do than just keeping them all in parallel.

Now I'm confused.

There is no choice between all S or all P here.

The target bank voltage determines how many Series, so far just 4S for 12V.

Just as there needs to be 6P .

I believe you're saying first parallel is better, so 6P at the lower level to get to 150Ah, **then** 4S to combine those 3.2V blocks into 12V nominal.

Right?

Nothing is getting unhooked in normal usage, whether bulk charging the whole 12V block at 50A,

or using 4x BC168s to balance-charge each cell at 30W, say 700W in total.

 

[john61ct]

When cells are paralleled, current will flow until the voltages in each cell are identical.

OK thanks, so the LA vs LFP difference here is one of degree in that regard.

1A: 6S lead - NO
2A: 4S LFP - NO
1B: 6P lead - YES
2B: 4P LFP - YES

So, for both types, units connected serially, need to be actively, explicitly balanced, whether done manually or automated.

And again both types, only parallel connections leads to anything like self balancing.

To prevent inrush issues, best to get units at least "pretty close" balanced before connecting in parallel.

 

[john61ct]

So, even if a given use case requires 24 or even 48V for total bank voltage,

I **really** want to keep to set standard 150Ah @12V pack modules.

These may well each include their own pack-level protections - first alarms, then automated isolation - temperature, LVD/HVD, maybe high current rate alarm.

To the extent both higher AH capacity **and** the higher voltage is required, parallel the packs first, then series, as with within the packs?

Some of those protections may be also applied at the full-bank level, redundancy being a good investment in this arena.

If per-cell charging is included in all the packs, then maybe don't also need additional 24/48V chargers as well, everything powered off 12V for that.

Feedback welcome. . .

 

[billvon]

So, for both types, units connected serially, need to be actively, explicitly balanced, whether done manually or automated.

Not really. For serial connections of lead acid you just have to charge them. You can equalize charge if you want, but usually that's not needed due to the self-balancing aspect of lead-acids.

To prevent inrush issues, best to get units at least "pretty close" balanced before connecting in parallel.

Yes. Generally for LiFePO4, you want to connect them all in parallel at first. You can then top balance them (balance them at 3.5V or so) or bottom balance them (at 3 volts.) There are pluses and minuses of both. It takes about a week depending on the loads/sources you use. That way when you assemble the battery everything is in balance and at the same voltage.

 

[999zip999]

LTO have you looked at this cell ?

 

[jonescg]

If you parallel all 16 cells and have a many-hundred amphour, 3.2 V cell, it will need to be charged up to 3.55 V in order to drive any balancing. If left to sit, the currents flowing into the lower cells will be in the microamps, and take years to eventuate. At least by charging up to Vmax it will be driven by a higher potential difference.

So, in short, if you want a 48 volt battery of decent capacity, buy something like a CALB 180 Ah cell. Put 16 of these in series, get a decent battery management system which monitors and balances all 16 cells, and be done with.

 

[amberwolf]

So in parallel. With both lead and LFP?

Any battery will eventually become the same voltage if you parallel them.

It doesn't necessarily mean they will be balanced for capacity, as that depends on the SoC curve of the specific cell/chemistry and the voltage they're at. For example, 3.2v on LiFePO4 covers a wide portion of it's capacity, though the exact voltage varies a bit.


You would have to look at the specs from the cell manufacturer for any specific cell, for their SoC curve to see what voltage is equivalent to what state of charge, to see if you could actually balance them capacity-wise at any particular voltage.

> LFP will not self-balance, period.
Except in Parallel, right?

No, it will not self-balance at all. Self-balancing is a mechanism by which cells will alter their own voltage, like the way lead-acid cells decrease their voltage by boiling off electrolyte, or NiXX cells decrease their voltage by heating up significantly.

It *will* become the same votlage across all the paralleled cells, but that's just normal basic electrical circuit behavior, and is not self-balancing.

Also, no 4/6/12V lead here, only single cell 2V units, as is common for large banks in marine & industrial use cases.

So, if one lead cell is at 1.85V, and the other is at 2.05V, hooking them up **in parallel** (nominal 2V) will cause no current to flow?

Of course it will. Any time you parallel two different voltage batteries (or anything else with two different voltages across them), current flows out of the higher voltage into the lower; if the difference is significant and the internal resistance is low enough, and capacity is high enough, you can even get a fire from it.

But if the voltages are nearly equal already, the current flow will be small, and it could take a long time for the cells to become the very same voltage.

And if one LFP cell is at 2.8V, and the other is at 3.3V, hooking them up in **series** (nominal 6.4V) will cause no current to flow?

It can't--there is nowhere for it to flow. Current can only flow from a higher potential to a lower one. If there is no connection between potentials, there can be no current.


You can draw these out on paper and see how they work. If you're not sure, there are places like https://www.allaboutcircuits.com/textbook/ that have great articles to learn the basics of circuits.



So far I believe you're saying (please correct, and of course looking from input from anyone who knows here, not just amberwolf)

1A: 6S lead - UNKNOWN

No, it's not unknown; if there is no connection from one potential to another, there is no current, so there can be no balancing.

2A: 4S LFP - NO

Correct.

1B: 6P lead - YES
2B: 4P LFP - YES

They will equalize voltage. Balance by capacity is different and depends on the SoC curve and voltage they are actually at, as noted earlier in this post.

 

[john61ct]

If you parallel all 16 cells and have a many-hundred amphour, 3.2 V cell, it will need to be charged up to 3.55 V in order to drive any balancing.

Yes, again, this was just to clarify my understanding, not about balancing as such.

And CALB 180s are a great example of what I will use for very big banks, but need minimum 2S for redundancy.

I will explore building up from the smaller 20-25AH units for more normal banks, specifically to take advantage of per-cell balance charging with the cheap 6-8A units like BC168.

 

[john61ct]

I really appreciate you guys sticking with me here, has been very helpful.

> LFP will not self-balance, period.
Except in Parallel, right?

No, it will not self-balance at all. Self-balancing is a mechanism by which cells will alter their own voltage

Aha. I apologize, I meant between cells, in the absence of any other devices.

> Any time you parallel two different voltage batteries (or anything else with two different voltages across them), current flows out of the higher voltage into the lower

Exactly, hence my confusion.

And if one LFP cell is at 2.8V, and the other is at 3.3V, hooking them up in **series** (nominal 6.4V) will cause no current to flow?

It can't--there is nowhere for it to flow. Current can only flow from a higher potential to a lower one.

So a serial connection does not present a difference in potential. Lesson learned.

Will try to make time for some https://www.allaboutcircuits.com/textbook/

For example, if my ICE charging session of 3 hours only gets my 200AH lead bank to 90% SoC, not worth burning dino juice for 4+ more hours to get to 100% as per endAmps stop-charge @ .5A

I want to hook up from my 800AH fully charged LFP bank to complete the process, resting V of the LFP bank over 13.2V, lead below 12.6V.

Say the natural current flow at those capacities+SoC starts at below 20A, then my usual #2 AWG jumper wires should be plenty over say 8-10 feet.

Would a resistor even serve a purpose here?

Obviously a DC-DC charger would be called for, but say one wasn't available. . .

 

[999zip999]

You are talking about 180ah this size cell has alot of stored energy and can be very dangerous when you're hooking them up in series. They must be of equal voltage and then all parallel to equalize. Frist so all are of same voltage. Let rest for a day or more and see if you have a leaker.Then can be series. A fire can happen with this much stored energy. Be careful.
LTO may be safer then lifepo4 and longer lasting then lifepo4 my lifepo4 72v pack is 5 yr old and 1,380 cycles is heavy. But can last many years with care. No battery problems.

 

[999zip999]

That's a great link https://www.allaboutcircuits.com/textbook/
I have a like book huge.

 

[amberwolf]

For example, if my ICE charging session of 3 hours only gets my 200AH lead bank to 90% SoC, not worth burning dino juice for 4+ more hours to get to 100% as per endAmps stop-charge @ .5A

I want to hook up from my 800AH fully charged LFP bank to complete the process, resting V of the LFP bank over 13.2V, lead below 12.6V.

The lead may not be at just 12.6v if it's 90% charged; could be higher than that. Depends on a bunch of things, from what exact chemistry / cell model it is, to temperature/etc, to current flow at the time, etc.

One article about the difficulty of determining SoC:
https://batteryuniversity.com/learn/article/how_to_measure_state_of_charge
There's other stuff out there too
https://www.google.com/search?q=lead+acid+SoC+curve

However...it's kind of a waste to use one battery bank to charge the other, or to use the ICE to do it, rather than just plugging in somewhere. But I guess if you *have* to charge one bank from the other, the lead is going to be the one needing it so it doesn't sulfate while at a lower SoC, since the LFP won't be damaged by being at a lower SoC.

Say the natural current flow at those capacities+SoC starts at below 20A, then my usual #2 AWG jumper wires should be plenty over say 8-10 feet.

Would a resistor even serve a purpose here?

:? Where?

If you mean to limit current...it would just waste power as heat, and potentially interfere with charging by slowing current down more than you want it to depending on the voltage difference. Might have to be a very large resistor, or bank of smaller ones, to dissipate the power, or be fan-cooled, etc.

Definitely a better idea to use a DC-DC that has adjustable current-limiting; it'll waste power, too, but not nearly as much as a resistor would, and it will allow you to get a full charge (in a reasonable time) where you may not with a resistor.

 

[john61ct]

Yes I guess technically won't get to "hold Absorb at 14.6V until amps taper to .005C" full without boosting the voltage a bit.

But then a low-amp boost converter's cheap.

That degree of Full only needs hitting a few times a week for longevity

 

[john61ct]

Please someone check my charge rate math?

Given a block of 6 cells, target V of 3.45Vpc and a fixed 6A current limit per cell-wire.

BC168 charging each cell independently, same as whole-string-charging wired serially, 6A @ 20.4V

--> 20.7W input per cell, 124W going into the block as a whole.

While block-charging wired in parallel, 36A @ 3.45V

--> 36W per cell, 216W going into the block as a whole.

I guess the 60% higher charge power comes from higher voltage pushing through without heating the wiring as much?

Another big plus for P wiring at the lowest cell level.

 

[john61ct]

Please someone check my charge rate math?

bump?