Manufacturing factors affecting the internal resistance of lithium battery pack

[zee123king]

What are all the manufacturing considerations that affect the performance of the battery packnand how do they affect the performance. For instance, nickel strip purity, number of welds, position of welds, strip thickness etc.

Also, what is a good sorting and grading criteria for the lithium cells (IR difference, voltage and capacity)?

Can I also get links to published research if any related to this.

 

[Korishan]

what is a good sorting and grading criteria for the lithium cells (IR difference, voltage and capacity)?

Hahah, definitely a @Wolf question there
He also has detailed threads on the topic

What are all the manufacturing considerations that affect the performance of the battery pack

There's a lot, probably too many to list.
However, the biggest is that every connection there's resistance. So every nickel strip welding connection is a bit of resistance. So to minimize that, you do multiple spots. The general practice is 4 spots, which translates to 2 tacks as there's 2 probes. Some ppl will do 3 tacks, or 6 welds. More than that on small cells, such as 18650's or smaller, is probably going to increase the chance of destroying the cell casing when done on the Neg end, and potentially break the Pos end cap.
- Side note, when spot welding, good idea to wait a few seconds between tacks so the heat can dissipate a bit. If multiple cells are lined up, you can help this out by doing multiple cells in a line, then going back and doing the 2nd run. Heat is the #1 killer of batteries

The other thing is the type of nickle. Obviously nickel plated is worse than pure nickel. But this can also depend on how many amps you are planning on running through those strips. If the amps will be low, say <1A, then nickel coated steel probably would be fine. But if you plan on pumping 5+A through that strip, then pure nickel is probably the way to go. And possibly multiple layers when getting higher.

Strip thickness would depend again on how many amps

 

[100kwh-hunter]

Hahah, definitely a @Wolf question there
He also has detailed threads on the topic

Not only Wolf, but the whole forum is also based on this.
Indeed Wolf made a very good guide's/user manual, should be a sticky.

Also, what is a good sorting and grading criteria for the lithium cells (IR difference, voltage and capacity)?

It is not so simple to answer this one, you need to read a lot and start to ask more specific questions.

 

[zee123king]

Can someone please share the link of the user manual or the thread shared by @Wolf .
P.S. Thanks Korishan for your answer.

 

[100kwh-hunter]

The Passion of the Li-Ion 18650

As we all are aware there are hundreds of millions if not billions of 18650 cells in all kinds of devices and there will continue to be a demand and supply of this form factor. Too many designs for this form factor would have to be changed, molds redesigned, BMS redeveloped, loads recalculated...

secondlifestorage.com

next time dont be lazy

 

[zee123king]

The Passion of the Li-Ion 18650

As we all are aware there are hundreds of millions if not billions of 18650 cells in all kinds of devices and there will continue to be a demand and supply of this form factor. Too many designs for this form factor would have to be changed, molds redesigned, BMS redeveloped, loads recalculated...

secondlifestorage.com

next time dont be lazy

Thanks man Wont be lazy

 

[Wolf]

Also, what is a good sorting and grading criteria for the lithium cells (IR difference, voltage and capacity)?

Short answer, If you want the best possible batterybuild, the cells will be all same manufacturer, same part number, same date code, at least 95% SOH, and an IR difference between all cells no > 5mΩ.
Now since we don't live in a perfect world my next scenario would be same manufacturer, same part number or at least same capacity, same chemistry ICR or INR but not mixed. (in any battery build I would not mix ICR with INR ever) SOH at 90+% and an IR delta of no more than 10mΩ.
Note to clarify this statement:. (in any battery build I would not mix ICR with INR ever)
I would build a 12/24/48 volt battery with exclusive INR cell and another 12/24/48 volt battery with exclusive ICR cells and parallel those.

My least favorite scenario is the Frankenstein approach, and then only main stream cells. Sanyo, LG, Samsung, and some Panasonic.
Sony's older cells as the US18650GR i would not mix with any pack as their IR is usually in the stratosphere along with the LG S3 2200mAh, or more commonly known as the ICR18650S3 consistently >70mΩ.
For a frankenstein pack I would try to keep a mAh delta of 250mAh and an IR delta of no > 10mΩ.

For more reading material see my shameless promotion of my study
See https://secondlifestorage.com/index.php?threads/behaviour-of-cells-in-parallel-take-2.8857/

Wolf

 

[zee123king]

Short answer, If you want the best possible batterybuild, the cells will be all same manufacturer, same part number, same date code, at least 95% SOH, and an IR difference between all cells no > 5mΩ.
Now since we don't live in a perfect world my next scenario would be same manufacturer, same part number or at least same capacity, same chemistry ICR or INR but not mixed. (in any battery build I would not mix ICR with INR ever) SOH at 90+% and an IR delta of no more than 10mΩ.
Note to clarify this statement:. (in any battery build I would not mix ICR with INR ever)
I would build a 2/24/48 volt battery with exclusive INR cell and another 12/24/48 volt battery with exclusive ICR cells and parallel those.

My least favorite scenario is the Frankenstein approach, and then only main stream cells. Sanyo, LG, Samsung, and some Panasonic.
Sony's older cells as the US18650GR i would not mix with any pack as their IR is usually in the stratosphere along with the LG S3 2200mAh, or more commonly known as the ICR18650S3 consistently >70mΩ.
For a frankenstein pack I would try to keep a mAh delta of 250mAh and an IR delta of no > 10mΩ.

For more reading material see my shameless promotion of my study
See https://secondlifestorage.com/index.php?threads/behaviour-of-cells-in-parallel-take-2.8857/

Wolf

@Wolf Your study is extensive and astonishing. Thanks

 

[DG98]

Sony's older cells as the US18650GR i would not mix with any pack as their IR is usually in the stratosphere along with the LG S3 2200mAh, or more commonly known as the ICR18650S3 consistently >70mΩ.

Wolf

@Wolf - can you please clarify this statement for me? I have a bunch (~3000) of LGDAS31865 cells, with the vast majority of them capacity testing at very good SoH -- 95%+ is typical. I don't intend to mix them with other cell types when I build a large pack. I have smaller quantities (~400) of Samsung ICR18650-22F cells, which I may build a smaller single-brand pack from. Lastly I have few dozen each of Moli ICR18650H, Sanyo, Sony US18650GR and other cells that mostly have terrible capacity test results.

I read through the thread you linked and saw one of these LG cells tested toward the end. Is there an inherent issue or problem you have found with the LGDAS3 cells that I missed? Thanks much for any link or further reading, I appreciate it!

Cheers, John

 

[100kwh-hunter]

Short answer, If you want the best possible batterybuild, the cells will be all same manufacturer, same part number, same date code, at least 95% SOH, and an IR difference between all cells no > 5mΩ.

You forgot one major important thing.... temperature.
The difference between testing them at 5C or at 25C is 10% difference in SOH.
Sorry Wolf.
I am atm thinking about a formula for the calculation factor to implement in my excel sheet to get over the tempratuur.
But for the perfect world, if you have enough cells, you can create the perfect battery.


From my point of view, mixing different chemistries does not matter so much.
What matter is the deviation in the packs itself.

Example:
You need 280 cells for a 2.5kwh battery and you have 140 icr chem and 140 inr chem, with ir around 20 to 40 and ir 60 to 80.
Now ifyou give every pack the same amount of inr and icr, i see no problem....BUT
If you do the first 7 packs only icr and the next seven pack only inr.....that could give problems.
But how much you are going to take from each cell, in my case...IN MY CASE not even 25mah..
Charge and discharge.

Every chemistry has its own charge and discharge speed.
In Wolfs data sheet it is easy to retrieve.
Mine also, but mine is not online. incl in between what IR per cell number/manufactory/ect to keep one or at what IR to skip a cell.
There is a lot to figure out from IR measurement alone.
Testing in a follow up order without IR rules out a lot of junk cells to.
I use both, to be at the save side(full range)......my house, my time and my effort....my price for a lot of decades....i hope.

For example, of chemistries:
NiMh is a chapter by itself, they will never except a charge bigger than 200mah when completely empty.
Or give more than 160mah for a discharge....
Some li ion will take happy 3000mah and give happy just 500mah...or the other way around....some can not handle more than 500mah give or take and they can hold 3500mah or the other way around! (all above 90%soh btw)

 

[Wolf]

You forgot one major important thing.... temperature.
The difference between testing them at 5C or at 25C is 10% difference in SOH.
Sorry Wolf.

Don't be sorry, we all learn here.
However I may not be in Netherland but it does get pretty cold in Maine also.
Therefore I do all my testing in the comfort of my home which generally resides in a temp range of 18°C to 24°C.
So using your 10% SOH difference with a spread of 20° my 6° difference would amount to ≈ 3% .Sure I may have some cells that would be just under the cutoff but not that many.

From my point of view, mixing different chemistries does not matter so much.
What matter is the deviation in the packs itself.

This is where I disagree I think chemistries are so fundamentally different that the combination within single packs should be avoided no matter what your charging/discharging rate is.
You also state that

Every chemistry has its own charge and discharge speed.

Therefore logic dictates that you want to maintain the same "charge/discharge" speed in each battery/pack.

with ir around 20 to 40 and ir 60 to 80

IR from 20mΩ to 80mΩ is a delta of 60mΩ. There is a huge difference in how a 20mΩ cell next to a 80mΩ reacts when charging/discharging.
They will be constantly battling each other.
You can see all this in graphic format in my parallel cell study.

Wolf

 

[Wolf]

I read through the thread you linked and saw one of these LG cells tested toward the end. Is there an inherent issue or problem you have found with the LGDAS3 cells that I missed?

No you didn't miss anything that I am aware of.
The issue with the LG is very well described here https://secondlifestorage.com/index.php?threads/the-passion-of-the-li-ion-18650.11456/post-80489

It is the high IR of the cell that makes it difficult to incorporate it into a frankenstein pack. But as you mentioned you have 3000 of these cells and many of them test well so build away keep the capacity as close as possible and the IR the same.

Wolf

 

[100kwh-hunter]

@Wolf our forums IR master! (mind you, there are always two...a master and a apprentice)

Therefore logic dictates that you want to maintain the same "charge/discharge" speed in each battery/pack.

Agree 100% but if every pack has the same amount of high IR and low IR cells, than the balancing would be the same on pack and cell level.
Probably i got my things in wrong order.
Cell--->pack--->battery--->system

So if one would have a high drain battery the cells would have to work hard, i think in this case you are right.
So if one would have a low drain battery the cells would have a easy life.
Meaning low drain all the cells can give the asked mah. high drain, the cells with high ir can not give the mah that is demanded.
but the low ir cells can get the first punch and the high ir cells can fill up the low ir cells?

You have a very well put together battery and a big one.
Oke now there is a heatwave in Australia and here are the electric heaters burning, so high drain it is for both of us right now.
Despite the fact that the solar chargers power directly your airco.
But if you would have a decent size battery bank, why would you separate low from high?
Only reason i can think of is that the low drain cells would go bad a bit earlier?
I think if divided (high and low) evenly it does not matter?

I think that low drain are better suited for energy storage than high drain cells, why not benefit the best of both worlds in case of a small energy storage.
Let the high drain take the first punch?

Don't be sorry, we all learn here.

Or am i completely wrong here?

IR from 20mΩ to 80mΩ is a delta of 60mΩ. There is a huge difference in how a 20mΩ cell next to a 80mΩ reacts when charging/discharging.
They will be constantly battling each other.
You can see all this in graphic format in my parallel cell study.

And this i really dont get, sorry.
Oke i am mixing in my packs low and high, to a average! not delta.
To my understanding of "delta":

1. Delta means change, in case of ir, there is only change when in use.
2. Delta means when in use the v will drop or rise after a set point.
3. Delta means when i am charging my cells and the v is topping the ah is dropping, or the v is dropping the ah will be topping.
4. Delta means in case of NiMh a change in temperature when the v is done and keep on charging, what is needed with NiMh
5. Delta on Sanyo cells...after a x year of self-life they can handle less and less in charge/discharge mah due to ir, there soh is also going lower(all cells btw) but some in heat, especially Sanyo(for 4 years now Panasonic?)

To my understanding a cells ir only goes up when in use or in self-life (or abuse goes rapidly)
Some us18650 (green sony) are climbing in IR with 15mohm in just one year....self life, without using them and at a perfect 3.330v storage v.
So at this point you are perfectly right in your statement of your "perfect cell" "world".
Same type, same manufactory ect.

But again, i hope you can tell me the pinpoint why not mix low and high ir cells in a modest to big battery banks.
Up to a certain point it does not make sense to me.

I think in a real good modest perfect battery bank at pack level the cells would benefit from some high drain cells.
To clarify what i think is a high drain bank--->300mah per cel charge or discharge and higher.
The more cells in your battery or pack or more banks in your battery would give you lower drain and lower charge.

The more batteries in your bank/more cells in your packs would give you lower strain and drain on cell level.
And now the lady is going to cook with the ac unit on and you want to take a shower, all electric...oh yes you just park your ev at home.... high drain for a hour.
after dinner you are fresh, well fed in a cold house, only the ev is a bit charging, low drain.

With kind and best regards Igor

 

[100kwh-hunter]

But as you mentioned you have 3000 of these cells and many of them test well so build away keep the capacity as close as possible and the IR the same.

I have a bunch (~3000) of LGDAS31865 cells, with the vast majority of them capacity testing at very good SoH -- 95%+ is typical.

Indeed keep the capacity as close as possible in all your packs, so all your packs will have a smaler deviation of 5Ah and keep the SoH also to a minimum. (70-80 or 80 to 90 or 90 to 100)
Cell saviors can do a lot but SoH you must do yourself.
Blue bar on top of your screen, there is a drop/pull down menu <battery tools>, click on it.
Then click on Cell saviors pack builder.
If having all in excel, transport (COPY) all data in to a other excel and only keep the data that is needed, divided by ---> : <---
If you have troubles with this i will or someone will guide, you with examples, it so simple

 

[Wolf]

And this i really dont get, sorry.
Oke i am mixing in my packs low and high, to a average! not delta.
To my understanding of "delta":

Delta's most common meaning is that of difference or change in something. In this case the delta (difference) between 20mΩ and 80mΩ is 60mΩ

But again, i hope you can tell me the pinpoint why not mix low and high ir cells in a modest to big battery banks.
Up to a certain point it does not make sense to me.

I think in a real good modest perfect battery bank at pack level the cells would benefit from some high drain cells.
To clarify what i think is a high drain bank--->300mah per cel charge or discharge and higher.
The more cells in your battery or pack or more banks in your battery would give you lower drain and lower charge.

The chemistry of a "low drain" ICR cell is Lithium Cobalt Oxide which gives it a higher IR right off the bat.
The chemistry of a "high drain" INR cell is Lithium Manganese Nickel. The Nickel and Manganese give the cell an inherently lower IR
which leads to an obvious conclusion, which is, the biggest difference between INR and ICR is Internal Resistance (IR).

INR cells tend to have much lower internal resistance and can deliver more current than ICR.
In a pack together, the INR cells will discharge their capacity faster than the ICR under heavy loads.
However all the cells are in parallel so voltage must remain the same creating stress on the INR
For demonstration purposes let's assume we have 2 cells in parallel one with an IR of 20mΩ and one with an IR of 80mΩ
We also have a 1Ω load and the batteries are at 4V. Ohms law tells us that the 20mΩ cell will discharge at 3.921A and the 80mΩ cell will discharge at 3.703A that is a delta oh sorry difference of 0.218A or 218mA
So its not a lot especially if you have packs of 80p or greater.
However it is a consideration.

But how much you are going to take from each cell, in my case...IN MY CASE not even 25mah..
Charge and discharge.

In this scenario you probably could get away with it.
But why would you even think about mixing cell chemistry then? As there is no need unless that is all the cells you have.

However in the scenario you describe below not so much. This is where I would use exclusively INR cells. They can provide the busts of power you need and when done just idle at 25mA and no harm is done.

The more batteries in your bank/more cells in your packs would give you lower strain and drain on cell level.
And now the lady is going to cook with the ac unit on and you want to take a shower, all electric...oh yes you just park your ev at home.... high drain for a hour.
after dinner you are fresh, well fed in a cold house, only the ev is a bit charging, low drain.

When cooking, AC and shower are done, the ICR will return some of its charge to the INR cell.
This causes extra charge cycles on the INR cell, which will age it faster.
With low current loads, this effect is minimal, but with frequent cooking AC and showering, it can drastically reduce the life span of the pack.
Again I am somewhat befuddled, if you have enough of the ICR cells to build a battery 100p14s or whatever you want in parallel why not just use them? Like wise if you have enough INR cells to build such a battery just do it.
If manufacturers thought it was a good idea to incorporate ICR and INR cells in packs they would have done it a long time ago.
Just saying
¯\_(ツ)_/¯
Wolf

 

[100kwh-hunter]

@Wolf Thank you very much for your extend and perfect explanation.
Now it makes a lot of sense.

No but and no however from my side!

I just looked at my testing results from the last 1000 and there is indeed a deviation.
There are two main groups, one with IR between ~20 and 30 and a group with between ~70 to 100.
All cells having a SOH leftover of more then 80%

My intension is to build packs of 20p14s, so 280 cells in a battery.
With a harvest of ~50 cells a week and 1500 already tested and a 1000 more to be tested, this must be a no brainer.
My wife will thank you.......sigh....it is a bit to cold in the shed right now to sort them out again, so on the kitchen table it is.
After testing my cells i segregate my cells from 50 to 60, 60 to 70, 70 to 80, 80 to 90 and 90 to 100 SOH.
So i will split them up on high and low ir's
Also i think i will put the Sony (green ones) on a separate pile, cause they tend to "age" faster.

The biggest benefit i will have (at least i think so) will be when the cells come to the point of low SOH, i dont have to desolder and retest the whole pack.
Just done is done.

Oke, no but or however, just a question
When building the batteries with matched IR they will end up on the same busbar.
Throw in the mix my lifepo cells and my NiMh, now what?
Those blue lifepo4 prismatic cells have a IR of ~20.
NiMh cells have a IR (good ones) from 4 to 10 Mohm.
Li ion's will have ~20 to 30 and 70 to 100Mohm.
All batteries will be on the same bus.
What is your thought about this?
I think there will be no problem, except that the inr would have some cycles more.

So when building a pack with li ions we have to take in consideration the SOH, IR(low and high) temperature, aging.
Exclude heaters, low SOH, out of average IR, self-dischargers, cells that take to long to charge.
"wake up" cells that have been to long on there storage v.


With best regards Igor

 

[DG98]

No you didn't miss anything that I am aware of.
The issue with the LG is very well described here https://secondlifestorage.com/index.php?threads/the-passion-of-the-li-ion-18650.11456/post-80489

It is the high IR of the cell that makes it difficult to incorporate it into a frankenstein pack. But as you mentioned you have 3000 of these cells and many of them test well so build away keep the capacity as close as possible and the IR the same.

Wolf

That's the link I had missed. Thanks very much for taking time to respond and pointing me in the right direction -- much appreciated!

Cheers, John