Lithium-ion multiple batteries on parallel - same challenge as Lead Acid or not?

[Roil]

I am considering connecting multiple (20-30 or even 40) 24V Lithium-Ion batteries together in a large 24V batteribank. As each module is 24V (58Ah) and as the batteribank is also 24V they all need to be connected in parallel.

Is the challenges of connecting such a large number of batteries in parallel the same as for lead acid?

i.e to keep the batteribank healthy they all need to have exactly the same resistance between the collection point for load and charging?

Achieved this this with his many modules is a nightmare - is it really necessary?

batt_new_2.gif

[SunEagle]

Any Lithium-ion battery system requires a BMS to properly charge and balance each cell. You really can't wire them like you would an FLA type battery. They need to be wired so that each "string" or cell will get the required voltage and you can't do that by just wiring them in parallel.

[Roil]

Each of these modules are already built as 7s3p - not much I can do about that (except tearing them down and rebuilding them in a different configuration)
(Yes I know they should have been 8s3p - but that's another topic)

So I have all of these 24V modules and I want (have to) connect them in parallel if I am going to use them - is it doable or will they "self-destruct" without a rigorous cabling scheme?

Screenshot 2018-11-17 at 18.51.10.png

[nebster]

You can do it, sure. But it might not be cost effective to make it safe.

Each of the 24V strings (one of your batteries) needs its own protection and monitoring mechanism. You need to fuse each one of them and also have a contactor to isolate them, unless each string already has that built inside. You need a battery monitor that can open that contactor in a high or low cell voltage situation.

If your strings already have those features built in, there is not much issue with paralleling. If they do not have those features, it will be very expensive to acquire all of that hardware for 20, 30, or 40 small strings.

The wiring issues are mostly about extracting the most capacity from your parallel strings. Equal resistances will do that better, so simply wire them all up to a pair of bus bars with equal length cabling going to each string. Since you will have unavoidable, very slight differences, some of the strings will get used slightly more than the others. This will result in very slightly uneven wear, but it is unlikely to be significant in my experience.

Those pictures of lead acid parallel wiring are mostly a red herring for lithium packs. I think they have issues because of voltage drop and challenges getting back to absorption and 100% SOC. None of those are problems with lithium chemistries.

[Sunking]

You can do it certainly as long as you do not mind destroying the batteries. No problem.

Each battery takes its own BMC and charger. So here is the problem. Say you have 8S4P. As individual strings each would require its own bms and charger which only a DIY or Fool would do. The proper way is to wire than 8S4P at the cell level so you only need one BMS and charger like you see below.

[Roil]

Each of the 24V strings (one of your batteries) needs its own protection and monitoring mechanism. You need to fuse each one of them and also have a contactor to isolate them, unless each string already has that built inside.

Why do I have to be able to disconnect every single module just because they are connected in parallel?

A Tesla 5Kwh module have many cells in parallel that are not equipped with this type of hardware.

[Roil]

You can do it certainly as long as you do not mind destroying the batteries. No problem.

Each battery takes its own BMC and charger. So here is the problem. Say you have 8S4P. As individual strings each would require its own bms and charger which only a DIY or Fool would do. The proper way is to wire than 8S4P at the cell level so you only need one BMS and charger like you see below.

I understand the need for a BMS per cell. That would mean many individuell cell monitors but they seem to be available for around 12$ a pcs with an 8W capacity. www.batrium.com have a solution for this and they do support cells wired in a string/parallel configuration all the way up to almost 200 cells

Why would I need individuell charger per module?

[nebster]

Why do I have to be able to disconnect every single module just because they are connected in parallel?

A Tesla 5Kwh module have many cells in parallel that are not equipped with this type of hardware.

First of all, at issue is what degree of protection you want, and you could, of course, choose to have less protection. It would work fine until it maybe didn't, or until it turned into a blowtorch. Blowtorches are more likely with some lithium chemistries and almost impossible to achieve with others.

Any time you put two or more lithium cells in series, you need to be able to see the intermediate voltages on each cell (or block of paralleled cells) in order to ensure that each is not getting too high or too low. When one of those cells/blocks gets too high or low, it is "imbalanced," and if you let it go too far, you can damage a cell and/or start a fire.

The Tesla modules are 74p, and then 6s. Then 16 of those modules are placed in series to make 74p6s16s. Tesla uses a monitor on each of those 96 74p blocks. If you just buy a Tesla module, you would want to implement a similar monitor on your 6 blocks.

Within one of the 74p blocks, Tesla implements a fuse on each of the 74 cells. There is no monitor on each cell, because that would be too expensive relative to the value of that little cell. But there is safety: if one of those cells goes out of whack, hopefully the fuse melts and takes it out of the circuit. Then that block loses 1.3% of its capacity, but you probably still drive home okay. You might not even notice it, ever, since a small imbalance is not the end of the world on a large pack that stays in the middle of its charge envelope most of the time.

You described a 3p7s battery that you then want to parallel at 20-40. Each of those 3p7s units can flow a lot more current and is presumably worth a lot more. For safety, you need to fuse each of those units. For safety, you need to monitor each 3p block, and you need a way for your monitor to take a bad block out of circuit. Most people with paralleled strings do that by taking the whole string out, because power relays are expensive.

Does that help? You're asking good questions, but definitely make sure you fully understand this stuff before you start experimenting.

[Sunking]

I understand the need for a BMS per cell.

You are missing the point. You are correct that a BMS is at the cell level, but not understanding how to do it. First thing is you cannot define a BMS or what it is. It can be a voltmeter or a Vampire Bleeder Board. A BMS needs to be integrated into the charger.

OK say we have 4S2P as pictured above. Conventional method would require 8 Cell Boards connected with 4-cells per charger. Each string will not charge at the same rate. One would be charged before the other thus requiring a second charger or control which is expensive and a poor solution that does not work well. The Alternate method is what Lithium batteries use paralleled at the cell level, not string level. Regardless if it is 4S2P as shown, or 4S99P only requires 4 cell boards and one charger. Otherwise you would need 396 cell boards and 99 chargers if you used Conventional. Good luck with that DYI.

[nebster]

I understand the need for a BMS per cell. That would mean many individuell cell monitors but they seem to be available for around 12$ a pcs with an 8W capacity. www.batrium.com have a solution for this and they do support cells wired in a string/parallel configuration all the way up to almost 200 cells

Why would I need individuell charger per module?

No, you can't have a "BMS per cell" in your scenario. Once three cells are paralleled, they are electrically one. You can put a monitor on each 3p.

There is no need for a charger per module. That would be an incredibly expensive and even more unusual way to do it.

Finally, there are good reasons to parallel multiple strings of cells in the real world. The classic motivation is to have some degree of sub-pack redundancy: if one cell in one string goes bad, the computer can isolate that string, but the rest of the pack can continue to support critical loads. There are other reasons to go that route as well. But, certainly, it is less common.

Doing that requires more hardware, more expense, and more design consideration. It becomes unrealistic when the parallelism is very high (your example of 20 or 30 or 40 strings). My own pack has 6 parallel strings, and I wouldn't want many more.

[nebster]

One would be charged before the other thus requiring a second charger or control which is expensive and a poor solution that does not work well.

This is incorrect. The paralleled strings will charge at different rates, proportional to their capacities, but in so doing they will all reach the same SOC at the same time. It will not be the case that one string is "charged before the other."

[SunEagle]

This is incorrect. The paralleled strings will charge at different rates, proportional to their capacities, but in so doing they will all reach the same SOC at the same time. It will not be the case that one string is "charged before the other."

Without a BMS, parallel lithium cells will charge at different rates and some may exceed their maximum voltage range without a limiter to bring them back. So they may not all reach the same SOC at the same time and some may actually get too hot to stay stable.

[nebster]

Without a BMS, parallel lithium cells will charge at different rates and some may exceed their maximum voltage range without a limiter to bring them back. So they may not all reach the same SOC at the same time and some may actually get too hot to stay stable.

That's true: if a cell or a string in parallel goes nuclear, all bets are off. In that case, the notion of SOC goes out the window, too, though.

But in normal operation -- which a BMS must be in place to ensure at all times -- paralleled strings will charge and discharge just fine off a single charge bus. Even if the underlying capacities of the strings vary considerably.

[SunEagle]

That's true: if a cell or a string in parallel goes nuclear, all bets are off. In that case, the notion of SOC goes out the window, too, though.

But in normal operation -- which a BMS must be in place to ensure at all times -- paralleled strings will charge and discharge just fine off a single charge bus. Even if the underlying capacities of the strings vary considerably.

I agree if the BMS does it's thing. But even then I have seen LiPo's get very hot and ignite while they are being charged. That is why I use a thermal bag for my RC batteries. You just never can tell when one cell decides to commit suicide.

[Sunking]

This is incorrect. The paralleled strings will charge at different rates, proportional to their capacities,

Wrong, dead wrong. Batteries charge rate is determined by the Internal Resistance and SOC. A BMS is the root cause of Lithium battery failures. No two strings charge or discharge at the same rate. You are on the wrong track semi-right idea. A 100 AH Cell will have roughly half the Resistance of a 50 AH cell. No 2-cells of the same capacity have equal resistance.

Unequal charging is not a problem for any battery type except lithium. Top Balance is a killer of lithium cells and the root cause of failures and fires. All courtesy of the BMS. That is why EV manufactures DO NOT TOP BALANCE or ever allow the customer to ever fully charge a lithium battery. They would be sued into bankruptcy if they did on warranty claims and loss of property and life.