Everything about connecting cells in parallel

[tomv]

Guys,

I'm trying to find more info on exactly what happens when cells are shorted together in parallel. General wisdom on the net says:
-It's ok to discharge cells shorted together in parallel, all chemistries
-It's ok to charge Lithium cells shorted together in parallel
- It's not ok to charge Nimh in parallel.

But that leaves out a lot. What happens if cells are mismatched different capacity? Different internal resistance? How much variance there is in capacity and resistance in different cells?

If I parallel two mismatched cells (say 3Ah and 2.7Ah) what happens to:
- Combined capacity? I suspect I'll get minimum cell capacity (so 2.7 * 2 = 5.4 instead of 5.7Ah) as stronger cell will be dumping its charge to the weaker cell.
- Effect on charge termination/balancing. The weaker pair will be reaching full charge sooner (after 5.4 Ah instead of 6Ah for regular cells). Good balancing circuit should allow those cells to dump charging current while other cells catch up.

More interesting question is what if resistance mismatches? Say my cells are 10 mOhm and 20mOhm. One cell would be getting twice the current of the other cell. Which also means twice the heat. So what are effects on
- Max discharge rating. I guess 2P cells have same C rating as single cell.
- Cycle life. One cell doing way more work than the other cell can't be good
- Cell matching. Harder woking cell would become weaker and weaker, complicating balancing
- Capacity - cell that discharges faster would then be charged by fuller cell, but this is inefficient
- In Nimh this is a runnaway effect as the harder working cell then drops its resistance even more. Do Lithium resistances increase with age? Stay stable?

Any published tests or manufacturer information on this?

Thanks a lot!

 

[Toorbough ULL-Zeveigh]

You got a lot on the plate there, so I'll start with a small bite & take the easy one.

If I parallel two mismatched cells (say 3Ah and 2.7Ah) what happens to:
- Combined capacity?

In the situation regarding discharge capacity, you're combined capacity, either real or rated as long as they're both apples, will be 5.7Ah.
You need a voltage differential depending on the chemistry from about a quarter to half a volt per cell for current to flow from the fuller cell to the (nearly) empty one, i.e. current has to flow downhill.
As the smaller cell nears empty (but not quite) it's voltage will stay at the same level as the larger one, just not contribute any current.
No voltage gradient means no discharge into the smaller cell.
If there's some exception out there I haven't considered I'm sure someone will point it out.

Next up.

 

[tomv]

Great info! In short then mismatching capacities of same chemistry cells = no problem!

What about mismatching internal resistance?

 

[Doctorbass]

Great info! In short then mismatching capacities of same chemistry cells = no problem!

What about mismatching internal resistance?

mismatching intermal resistance of the same chemistry will results in a situation where the lower int res cell will give more current and will heat less than the one that have higher int res.

Normally, internal resistance is significant index about the age of a cell. as the cell is aged, as the int res increase.

IT IS VERY IMPORTANT THAT TEMPERATURE MANAGEMENT OF LARGE PACK ARE RESPECTED

heat affect alot int res!!! so cells surrounding the pack will be cooler than the one in the middle.

if you want cell to age equally, their heat need to be the same all the time during cahrge or discharge.

A good thermal management can increase pack performance by 20% !!

read article on the battery university website!

I have been working on a large 2kW battery pack since 5 month.. it is very long to test each individual 2p 1s cells!!
(they are from my Makita tool service center (i got around 65 used or deffective 18V 3Ah pack.))
you can see that i'm testing cells capacity and int restsiance one by one!! each take 40min at 5A disch
but the 4 wire current sensing that i developped is very accurate for int res measurement!!

In the graph i have below, you can have a great exemple of around 280 cells that aged in 4s2p pack from different age and use.

at around the 2p1s cell group no 80, i upgraded my connection method with 4 wire current sensing... As you can see on the graph, the int res measured is closer to the official actual value and we see that before that method, the wire between my charger/tester and the cells influenced and introduced an error of around 7miliohm dur to the connection and wire resistance. now, i have no prob !!

Combining A123 dewalt and konion makita cells, and other AGM SLA batteri, i will have tested around 500 cells from now!

Doc

I will match them in a perfect 18p 24s (when i will have finish to measure all them! 360...) 88V at 25Ah !

Doc

 

[Toorbough ULL-Zeveigh]

...In short then mismatching capacities of same chemistry cells = no problem!

Welll, not exactly.
That was mathematically speaking how to go about calculating capacity.

For practical purposes you do want to match the capacities with the same size cells to maximize the cycle life & performance.
The bigger battery will be carrying the brunt of the full load once the smaller one stops contributing current & will tend to wear out faster.
I wouldn't characterize it in terms of safe or unsafe.
You can combine mismatched cells in a pinch as long as you're aware of the toll it takes & it's simply not good practice.

 

[tomv]

Very intersting data Doc. So just to try some numbers, let's assume I get real unlucky and pick two cells out of your batch that are:
C1 2.7Ah 30 mOhm
C2 3.0Ah 50 mOhm

Lets say I run them in parallel. I plug fully charged pack into my motor and draw 30 amps out of them (5C).
Current flowing/power disippation would be:
C1 19 amps, 11 Watts
C2 11 amps, 6 Watts

At the end of discharge, let's say after C1 is fully discharged and not contributing

C1: 0 amps, 0 watts
C2: 30 amps, 45 watts (!)

So it looks like running more cells in parallel gives you more capacity but not more power.