2P4S vs 4S2P, and paralleling question

[fxykty]

To clarify, we are running two 4S batteries (12V) in parallel. Both batteries are built from large Winston cells, one set three years younger than the other. I ignore the relative SOCs as they don’t really matter, but after three months with this setup it works just as Jedi says: the batteries have very similar voltages and current flows and whenever they differ you can see the current flow favour the higher (on discharge) or lower (on charge) voltage battery. So far I haven’t noticed them being more than 0.0x V different in voltage, even with big loads. Current flows can be quite different at the knees, particularly so when one battery has reached 100% SOC and the other hasn’t yet.

Another benefit of running two batteries in parallel is that all current flows will be halved for each battery. Lower C values are more friendly to your batteries.

If you can, go 4S using larger cells rather than 2P4S with smaller cells, regardless of whether you decide to parallel two batteries or have just one battery.

[Ryban]

Originally Posted by Pete7

I do question the need for 560Ah of LFP capacity on a 32ft yacht. We managed with electric cooking on 120Ah of LFP with 85Ah of LA in a hybrid bank. This was upped to 220Ah plus 85Ah LA, last summer to give a bit of extra capacity for rainy days, hence the different sized LFP batteries.

Partly it comes down to how little space a 560Ah bank will take up, and how cheap the cells are these days. At least for the 2P4S option, where the BMS costs are low still. Upgrading our 460Ah LA bank to a 560 Ah bank will still be freeing up 2 GC2 cells worth of space.

I've also experienced plenty of days where tropical waves come through and we have clouds for days, and had to cut back our usage. I want to ensure that we're able to continue cook/charge our outboard battery (if we decide to go electric there) on days where we're not getting sun.

Originally Posted by fxykty

Another benefit of running two batteries in parallel is that all current flows will be halved for each battery. Lower C values are more friendly to your batteries.

If you can, go 4S using larger cells rather than 2P4S with smaller cells, regardless of whether you decide to parallel two batteries or have just one battery.

Aren't the individual cells in a 2P4S pack discharging the same current as if they were in a 4S2P arrangement though (assuming cells are balanced/matched and the 4S2P arrangement is discharging equally too)? There's more current going through the FET BMS, and battery cables, but I'm planning to use a BMS that controls a contactor.

Unfortunately I haven't found a high Ah cell that I can accommodate in that location, because most cells are not built tall, and some manufacturers specifically advise you not to put cells on their side (e.g. Wisnton).

[Pete7]

Originally Posted by Ryban

Partly it comes down to how little space a 560Ah bank will take up, and how cheap the cells are these days. At least for the 2P4S option, where the BMS costs are low still. Upgrading our 460Ah LA bank to a 560 Ah bank will still be freeing up 2 GC2 cells worth of space.

I've also experienced plenty of days where tropical waves come through and we have clouds for days, and had to cut back our usage. I want to ensure that we're able to continue cook/charge our outboard battery (if we decide to go electric there) on days where we're not getting sun.

Ok, understood.

Originally Posted by Ryban

Unfortunately I haven't found a high Ah cell that I can accommodate in that location, because most cells are not built tall, and some manufacturers specifically advise you not to put cells on their side (e.g. Winston).

That's a shame, I thought Winston did, but looking at the EU distributor the range seems to have shrunk, or at least consolidated. Same thing seems to be happening in Europe with solar panels becoming almost a series of standard sizes.

https://shop.gwl.eu/LiFePO4-Single-C...00Ah-and-more/


Pete

[s/v Jedi]

Attached my typical early morning battery status. The diagram shows my setup with each battery having it’s own BMV and those shunts connected to the same SmartShunt which measures the total.

The BMV’s report 9% difference in SOC and the solar almost taking care of all the load, but the LiTime bank still discharging a little while the Winston bank is already charging.

As the sun comes up more they both start charging with the Winston bank seeming to finish charging much earlier but the LiTime bank catching up before that happens. It’s simply the rapid rise of voltage in the knee that halts the charging in the Winston bank with the LiTime bank then taking the full charge to catch up.

[s/v Jedi]

For height of cells, you can simply stack multiple layers. Here is my Winston battery: it is eight cells in series, 400Ah each. Four on the bottom, four above that, then a cover with fuses, shunt and the cell leads that are fused and 10AWG to allow in-circuit cell charging.

The second picture is the same battery installed.

[s/v Jedi]

Originally Posted by Ryban

Aren't the individual cells in a 2P4S pack discharging the same current as if they were in a 4S2P arrangement though (assuming cells are balanced/matched and the 4S2P arrangement is discharging equally too)? There's more current going through the FET BMS, and battery cables, but I'm planning to use a BMS that controls a contactor.

Unfortunately I haven't found a high Ah cell that I can accommodate in that location, because most cells are not built tall, and some manufacturers specifically advise you not to put cells on their side (e.g. Wisnton).

You can’t buy the matched cells. You would need to order for millions of dollars before the manufacturer will do this for you… you don’t even get to order from a manufacturer and probably, for affordable cells, you will get used cells that are “renewed”.
Matched cells come from the same manufacturing batch, normally with sequential serial numbers but still checked for equal curves. Matching cells is more than measuring internal resistance… it must be matched over the whole range of SOC.

So you will always have one cell doing more than the other parallel cells. Only under heavy load, like hundreds of amps, and both cells in the linear part of the curve, will the current become equal… but not their soc%.

Edit: now there are ways around it… no guarantees but here are two options:

1. Buy Winston cells from Julia Yu. Mine were 8 cells all sequential serial numbers. This still doesn’t mean they are matched but there’s a very good chance.

2. Buy the cheap LiTime batteries with parallel cells (i.e. the 12V 200Ah) and those pairs are most probably matched. Simply take them out and you have them.

[wholybee]

How big of a deal is it, really? Assuming 2 or 3 parallel cells, discharging mostly at a low C rate. With a 500+Ah bank, even an electric galley is low C rate.

One cell works harder than the other. So that cell hits the theoretical 5000 cycles first, and your battery fails after 18 years instead of 20?

And your capacity isn't reduced, or shouldn't be. When the cell that is working harder gets into the steep part of the curve, the voltage difference between cells will cause the other to work harder to catch up. At least at lower C rates.

I chose my configuration (3p4s) because the size and shape of the space dictated it. I worked out different configurations with a dozen cell sizes, and all the available at the time drop ins. I could get 300Ah using 100Ah cells, 200Ah using drop ins. There wasn't enough room for 4s3p.

Now, i did spend a lot of money through an EV dealer to get Calb cells, drop shipped to me directly from Calb, sequential and very closely matched, with factory test data etc. But im not sure how much that really matters.

I argue matched cells are more important in a series arrangement. Yes, as jedi says the current is exactly the same. But IR is only one factor in a matched cell. They are also matched to self discharge, charge efficiency, etc. and those will affect cell balance even is current is the same.

In a parallel arrangement, although one cell works harder, once that cell gets into the steep part of the curve, they will come together again. Not at all different from when cells are paralleled together and charged to top balance.

Anyway, my conclusion is that it doesn't matter. Whatever configuration works. People will argue for one or the other. Meanwhile, other people run different configurations with no problems.

[s/v Jedi]

Originally Posted by wholybee

How big of a deal is it, really? Assuming 2 or 3 parallel cells, discharging mostly at a low C rate. With a 500+Ah bank, even an electric galley is low C rate.

One cell works harder than the other. So that cell hits the theoretical 5000 cycles first, and your battery fails after 18 years instead of 20?

And your capacity isn't reduced, or shouldn't be. When the cell that is working harder gets into the steep part of the curve, the voltage difference between cells will cause the other to work harder to catch up. At least at lower C rates.

I chose my configuration (3p4s) because the size and shape of the space dictated it. I worked out different configurations with a dozen cell sizes, and all the available at the time drop ins. I could get 300Ah using 100Ah cells, 200Ah using drop ins. There wasn't enough room for 4s3p.

Now, i did spend a lot of money through an EV dealer to get Calb cells, drop shipped to me directly from Calb, sequential and very closely matched, with factory test data etc. But im not sure how much that really matters.

I argue matched cells are more important in a series arrangement. Yes, as jedi says the current is exactly the same. But IR is only one factor in a matched cell. They are also matched to self discharge, charge efficiency, etc. and those will affect cell balance even is current is the same.

In a parallel arrangement, although one cell works harder, once that cell gets into the steep part of the curve, they will come together again. Not at all different from when cells are paralleled together and charged to top balance.

Anyway, my conclusion is that it doesn't matter. Whatever configuration works. People will argue for one or the other. Meanwhile, other people run different configurations with no problems.

Really? So what happens with 3 or 4 parallel cells when one develops an internal short?

Can you give data on different self discharge from the same type and age cells?

[lateral]

Yip, Being able to turn half the bank off was the decider for me.
Custom bus bars easily made out of 15mm copper pipe squashed flat
in a vice. Diagonal fold for 90° bends & heat shrink makes setup more do-able.

[wholybee]

Originally Posted by s/v Jedi

Really? So what happens with 3 or 4 parallel cells when one develops an internal short?

Can you give data on different self discharge from the same type and age cells?

Data, no. And I can't even provide a link to the video by an engineer at EVE that explained it.

He was talking about different failures and reasons a cell would be rejected and not sold as a grade A cell. We all know about differences in IR or differences in capacity. He said the most important factor to know with the grade b cells was self-discharge.

He described the manufacturing process. The 2 electrodes are made by plating the appropriate metal foil. They are stacked with a separator. Then rolled up. And then cut into the correct size for the cell being produced. All of the cells cut from the same roll should in theory be matched, as the chemical and physical composition of the plating will be identical for all of them.

But, when the foil is cut, it leaves a slightly jagged edge. Sometimes there will be a microscopic short through the separator from one or more of the burs. That is what causes excessive self discharge, and that is the worst problem to have with grade b cells. At best, you will never be able to keep the cells in balance. Worse, over time the short becomes worse. The cell might have a very short life, or might fail catastrophically.

With the same type of cells, assuming none of them have that problem, self discharge would remain low throughout the cell life, and any small changes should be the same for cells cut from the same roll.

(I wish I could find that video. It was very interesting)

Regarding a shorted cell, I think the result would be similar to if a single cell suffers a direct internal short in a 4s configuration. That cell is going to have some massive internal current, and melt or maybe catch on fire. I have seen evidence of this happening in an RV, and the results were not good. The RV survived, and the fire burnt out on its own without igniting the whole vehicle. But the RV was a total loss due to smoke damage. And all the LFP batteries were destroyed, though you could see which cell it started in.

You can argue it is much worse with other cells in parallel to the cell with a short. But is it? I have three 100Ah cells in parallel. If one of those cells develops a short, will there be more current involved than if a single 300Ah cell develops a short? I honestly don't know, but the stored energy is the same in each case, and individually my 100Ah cells have a higher IR than a single 300Ah cell. I don't think the answer can be determined without testing. But my gut is the results will be similar. (not good)

And of course there are examples of drop-in batteries that have an xP4S configuration, and cells are paralleled together in EVs and I'm sure in hundreds of other applications.

I still see no real evidence that one arrangement is better than the other. Whatever works for the person/space/boat.

[lateral]

[QUOTE=Ryban; Aren't the individual cells in a 2P4S pack discharging the same current as if they were in a 4S2P arrangement).[/QUOTE]

At a cellular level I have to agree. So no gain except for being able to isolate half the 4s and maintain power to critical devices. You could always switch over to start battery, but that should be a last resort and leaves no backup in event it’s depleted . I don’t fancy hand cranking
When the fan is rattling with lee shore beckoning.

I am somewhat perplexed on Victron manual on mid tap reading voltage . Jedi’s BMV for each series make more sense..

[s/v Jedi]

Originally Posted by lateral

I am somewhat perplexed on Victron manual on mid tap reading voltage . Jedi’s BMV for each series make more sense..

You mean that Victron assumes two 12V batteries in series? This is how things used to be in the old lead acid days… I think they never thought to update that section of the manual.

I can do midpoint monitoring on my Winston battery but not on the LiTime batteries because I decided to buy the 24V models instead of the 12V models of equal capacity. The reason is that now I don’t have parallel cells in the batteries (don’t like it even when matched) plus I can use each battery standalone as well as in parallel like they are now. I count in their BMS’s to protect them individually while they share a single fuse just like traditional battery banks.

[Ryban]

Got tired of waffling and pulled the trigger on 8x 280Ah EVE cells today. Will decide later on whether I arrange them into two packs or a single 2P4S pack.

[lateral]

Originally Posted by s/v Jedi

You mean that Victron assumes two 12V batteries in series? This is how things used to be in the old lead acid days… I think they never thought to update that section of the manual.

I can do midpoint monitoring on my Winston battery but not on the LiTime batteries because I decided to buy the 24V models instead of the 12V models of equal capacity. The reason is that now I don’t have parallel cells in the batteries (don’t like it even when matched) plus I can use each battery standalone as well as in parallel like they are now. I count in their BMS’s to protect them individually while they share a single fuse just like traditional battery banks.

Yep, just bought another BMV712 to give me one on each 4S. Thanks!
It was not optimum with one on the 4S's paralleled.

[hjohnson]

Personally, I went with a 2P4S configuration. Due to the size of the hole the batteries were going into, I elected to use 230Ah cells for a total capacity of 460Ah.

The primary reason for doing this was the simplicity of the BMS setup. I’m using one REC ABMS to manage the whole thing, with a single contactor for the high side bus. That said, I’m using a Blue Sea contactor, which has a mechanical lockout and mechanical override. If the BMS fails, I can force things on to get home.

By the same token, I was also building a fully integrated system, with the BMS managing the voltage/current using Victron’s DVCC system. The victron setup only can track one battery, so unless you’re running something like pylontech batteries which can combine their BMSs and present as a single battery to the system, you’re kinda SOL for that functionality if you have two independent BMSs.