2P4S vs 4S2P, and paralleling question

[Ryban]

I'm pricing and sizing out our new house bank. I like the idea of using the EVE Grade A 280AH cells, as they are affordable (0.11 $/Wh) and seem to be good quality.
https://www.18650batterystore.com/products/eve-lf280k

I'd like to build a 560AH bank with these, but am torn between making two batteries and putting them in parallel, or making one battery by putting the cells in parallel (which I've mocked up in CAD below).

The advantages of building the 2P4S bank:

• Simplicity
• < 1/2 the cost, since I would only need 1 BMS/Contactor/associated hardware (especially since I'm leaning toward the REC BMS)
• Probably more compact
• Potentially less issue with load imbalance if two batteries were used instead of one -- though this can be mostly addressed by making sure all external cables are equal length and the BMSes are identical.

The disadvantages maybe:

• Lack of redundancy, relying on a single "battery"
• Lack of redundancy, since only 1 BMS/Contactor will be present -- but I'd likely carry an extra BMS onboard anyway.
• Potential issues with cells not being loaded equally?

The last bullet point I have the most questions about. Ideally the resistance between the parallel cell terminals should be as low as possible.

Should I have the battery cables contact in the middle of the bus bar?
At the bottom of the photo, I have the positives of 2 cells connected linearly to the negatives of the next 2 cells, such that two terminals are closer to one another. Would it better to use equal-length cables between each of these terminals instead? Am I over thinking this?

[fxykty]

You are over thinking, but that’s OK with your first LFP build. Refer to Jedi’s posts of model system architectures for more in depth information.

2P4S is much less complex and as long as all the cells are evenly matched (consecutive serial numbers is one way to tell) you shouldn’t need to worry about imbalances between the 2P cells. Your mockup using copper bars (multiple sheets, or solid) as shown looks fine to me. You definitely do not need to take off from the middle of the 2P terminal bar.

4S2P is much more complex beyond just needing a second BMS. And note that the 2P will be to the positive and negative (or positive charge, postive load, and negative) bus bars, not to each other. Other than redundancy there’s not much benefit.

[Ryban]

The thing that was sort of a shocker was when I started reading about how different LFP batteries in parallel will potentially discharge unevenly. But the more I thought about it, the more I realized almost certainly the same thing happens with our LA batteries in parallel (if cable runs are different lengths/contact resistances are different/internal resistance is different), but we don't have BMSes on these and don't notice the uneven discharge.

I think I'll build the 2P4S battery because it just makes a lot more sense. I can maybe carry a spare cell or two on board if I'm really worried about a single cell bringing the bank down, but we also have a 230AH LA starter battery to fall back on, so we're not wholly dependent on the LFP batt. We could also re-assemble the bank into a 280AH 4s bank too without too much trouble.

[mlydon]

With 4p2s, you won’t be able to isolate each cell in the ‘p’, so if one cell is crapped out you can ruin the other cell….with 2p4s, you can monitor and adjust every cell in the pack.

Matt

[lateral]

Originally Posted by Ryban

I'd like to build a 560AH bank with these, but am torn between making two batteries and putting them in parallel, or making one battery by putting the cells in parallel (which I've mocked up in CAD below).

The advantages of building the 2P4S bank:

• Simplicity
• < 1/2 the cost, since I would only need 1 BMS/Contactor/associated hardware (especially since I'm leaning toward the REC BMS)
• Probably more compact
• Potentially less issue with load imbalance if two batteries were used instead of one -- though this can be mostly addressed by making sure all external cables are equal length and the BMSes are identical.

The disadvantages maybe:

• Lack of redundancy, relying on a single "battery"
• Lack of redundancy, since only 1 BMS/Contactor will be present -- but I'd likely carry an extra BMS onboard anyway.
• Potential issues with cells not being loaded equally?

The last bullet point I have the most questions about. Ideally the resistance between the parallel cell terminals should be as low as possible.

Should I have the battery cables contact in the middle of the bus bar?
At the bottom of the photo, I have the positives of 2 cells connected linearly to the negatives of the next 2 cells, such that two terminals are closer to one another. Would it better to use equal-length cables between each of these terminals instead? Am I over thinking this?

Doing that hasn't worked in practise with my setup of 4s2p, 2x bms's.
One battery charges/discharges quicker than other but as it is it works in my favour as I get an early warning before audible disconnect on trial max charge/discharge. 230Ahr CaLb x8. In 12mnths now and pleased with my decision. I have tried to trace the reason with internal resistances/ volt drops and thermal junction check. Can't be bothered any more, works great. Pic of initial install, now several additions. (Audible warnings LV, HV)
Oh, I ended up with BMS's with different FW versions, so not surprised.
Oh^2, 115A bosch alternator, to Argofet splitter to house bank and SLA start @14.2V All Works fine.
Have DC-DC 18A iso to supply start battery but never needed so will use for electronics protection.

[Ryban]

Thanks for sharing a photo of your bank, that really helps.

I'm surprised they don't discharge evenly based on what I can see in your photo -- how big of an imbalance is there when discharging high currents? Is it like 10%?

[s/v Jedi]

Don’t do it. Instead, build two batteries, each one 280Ah.

Yes, two BMS’s, two fuses and two main battery switches.

You should be able to make do with just one of the two. If not, build a third.

Batteries that have cells in parallel need matched pairs. Battery manufacturers can order these from cell manufacturers (i.e. LiTime from Eve or whichever cell they use) but you will not be able to get anything better than “yes, sure, you will be okay” from the seller without anything to back that up. A small difference in internal resistance means imbalance and loss of capacity.

[wholybee]

In the end it doesn't matter. I have a 3p4s bank that is fine for several years. Still perfectly balanced, still get more than rated capacity from it. When I discharge "high" currents it is still only .2C and for a relatively short time, maybe 20 minutes to cook a meal. Unless you are running 1C for long periods, you shouldn't see any issue either way.

[Ryban]

Originally Posted by s/v Jedi

A small difference in internal resistance means imbalance and loss of capacity.

Isn't this true with two batteries in parallel too? I actually don't know what will happen when one LFP battery is fully depleted and the other still has 10-20% left, and whether that's an acceptable thing to do. If it isn't, then that's 5-10% of the total capacity you don't have access to.

[Jenta]

I made a 24v bank with those cells and did 8s4p for a few reasons.
1. I can monitor every cell individually.
2. I have essentially 4 battery banks for massive redundancy.
3. I ordered 32 cells for redundancy and spares and then figured out a way to make them all fit once I had them in hand so - why not!?!

Plus - I'm a bit of an electronics nerd and I love my gizmos.

I started with 4 Daly 30amp BMS units and was only going to use them to balance the cells, monitor the cells through Bluetooth and trigger my BlueSea remote battery switches. (one for each bank)
However, that posed new issues even after all my custom circuits I made for triggering relays and such. One being the inconvenience of monitoring, but the big one was that they will only balance when charging and since I was not running current through the Dalys they never knew the banks were charging.
I probably could have worked around that - but I ended up going with REC-BMS for my solution and love it.
Mind you, I paid as much for the full BMS solution as I did for the batteries.
But if it is in your budget, the Master/Slave configurations for the REC-BMS are really, really nice. They also have good and active support.
I went through https://ogm-energy.com/

I bought my cells almost 3 years ago now, and only paid about $90 a cell from Alibaba including freight. The REC-BMS with a master, 4 slave units and 4 shunts and wireless unit ran around $3k at the time.

But, it works with all my Victron gear and controls the charging voltages, current, inverters - everything - for any charge/discharge device on the network. So my Quattro, Multiplus, all the victron MPPT's - and even the Wakespeed WS500 I hope to install later this year - they are all "managed" by the BMS.

So...
If it is in the budget, I would do 4s2p for redundancy and cell isolation. It won't make the battery itself any more expensive, and the layout you drew up can still be used and have two negative and two positive leads come from the same end of your pack. That also benefits because at 1C you have a single bank capable of 560 amps or two banks capable of 280 amps. The wire gauge for carrying 280 amps will be more manageable.

But - I'm guessing you're looking at about $700 all in for a single REC-BMS Active BMS with the contactors, shunt and such.
To do that with their Master and two slaves it's going to be about $2100 - so it ain't cheap. But it's only a couple boat bucks so...

If it ISN'T in the budget, what you drew up should work well also.

[lateral]

Originally Posted by Ryban

Thanks for sharing a photo of your bank, that really helps.

I'm surprised they don't discharge evenly based on what I can see in your photo -- how big of an imbalance is there when discharging high currents? Is it like 10%?

<10%
Cheapo JBD bms's, victron BMV712 , done very economically.
What Jedi sayes about matched cells bears out. I have one cell in each bank that charges /discharges first. They are the ones I monitor with no name $10 relays with audible alarms. BMV does bank whole HV/LV. Not all bells and whistles but way adequate for my budget.

[s/v Jedi]

Originally Posted by Ryban

Isn't this true with two batteries in parallel too? I actually don't know what will happen when one LFP battery is fully depleted and the other still has 10-20% left, and whether that's an acceptable thing to do. If it isn't, then that's 5-10% of the total capacity you don't have access to.

No, it isn’t. Let me try to explain:

When you connect two cells in parallel and they are not exactly the same, one will be doing more work than the other, charging and discharging faster. This means the other cell is always lagging, i.e. not used to capacity.

When two batteries are paralleled but each battery has only cells in series then there are no cells connected in parallel. Within a battery, every cell, by laws of nature, has exactly the same current passing through.

But the two batteries can go out of balance and they will. Mine are regularly 8% out of balance, but I got there on purpose because to prove my points, I got very different batteries. One is super high end self built from Winston cells and the other is the cheapest LiTime I could find.

But it isn’t a problem. While charging, the Winston charges way faster but as it starts neaeing 97-98% and the other is just at 90%, the Winston starts reducing charge current while the LiTime doesn’t and can make up the difference before both reach 100% at the same time.

Edit: so with parallel batteries the only effect is that at times one does more work than the other, but when they reach the upper or lower knee they will automatically synchronize again.

[rslifkin]

Is there any reason why we wouldn't see that same behavior (getting out of sync but then syncing back up in the knees) with paralleled cells?

[s/v Jedi]

Originally Posted by rslifkin

Is there any reason why we wouldn't see that same behavior (getting out of sync but then syncing back up in the knees) with paralleled cells?

To a point, but when a cell discharges into a parallel cell, which happens with varying load, the pair becomes unbalanced with pairs in series, leading to more trouble. Also, you lose rated power, i.e. when the normal load is 1C you can’t double that with two parallel cells anymore.

I must say that I see this in extreme situations with the parallel batteries too. When I do a very high load via inverters (say 5kW in the galley) then everything works great and I see currently balanced over the batteries. But then when I turn the load off, I sometimes see a couple amps flowing from one battery to the other.

This is never much though. When one battery has 10% higher SOC than the other, but they are in the linear part of the charge/discharge graph, they pretty much balance the load. When one is nearing a knee, that’s when the difference occurs. But when the batteries can easily handle the amperage it isn’t a problem and the cells stay balanced because they all pass the same current.

[Pete7]

Originally Posted by s/v Jedi

But the two batteries can go out of balance and they will. Mine are regularly 8% out of balance, but I got there on purpose because to prove my points, I got very different batteries. One is super high end self built from Winston cells and the other is the cheapest LiTime I could find.

Edit: so with parallel batteries the only effect is that at times one does more work than the other, but when they reach the upper or lower knee they will automatically synchronize again.

Our experience with with different sized drop in batteries is similar, one 120Ah and one 100Ah manufactured 2 years apart, at least.

What happens in practise is during a heavy load they both supply a slightly difference amount of current but the cell voltage stays close. Once the load is taken off and solar start charging again they will sort themselves out. The only thing to remember is the SOC display from the BMS will be different due to the difference in capacity. However, since this is only a rough guide to be glanced at, I don't worry about it. Lead acid batteries probably also did the same, we just never measured it to hundreds of a volt or at cell level.

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.

However, LFP is just a store of power, it does generate it, so if you use it, well it has to be replaced. If you are not replacing it every couple of days when will you? Also what are you running that needs such a large bank each day? We know from our stats that the yacht plus electric cooking uses an average of 1.7kW each day. We have a total of 3.8kW in the bank. Might be more useful to increase the electrical generating systems, rather than a larger bank.

Some stats attached for info.