Relationship between internal resistance and capacity

[PV1]

So, I've steadily been gaining capacity in my traction battery (as tracked in the thread EVBatMon Readings), but voltage rise/sag seems to be getting worse. During the last quick charge, input current seemed low for the SoC. Coming home from 7 Springs, cell voltage was dipping below 4 volts above 90% charge, which I don't remember it doing before.

Am I going crazy, or can internal resistance increase without affecting capacity?

 

[mdbuilder]

While you were driving? I'd be surprised if it didn't dip below 4.0 under load.

 

[zzcoopej]

Am I going crazy, or can internal resistance increase without affecting capacity?

I can't help you with the answer to the first question you pose as I sometimes think the same thing (lol), however I believe somewhere in this video Prof Jeff Dahn hints at your latter question being true (its a fascinating video on battery tech anyway if you havn't seen it) https://www.youtube.com/watch?v=pxP0Cu00sZs

 

[Phximiev]

See this post for the same video also and the Battery Degradation topic in general for the notions:

http://myimiev.com/forum/viewtopic.php?f=23&t=2325&p=20877&hilit=Check+out+this+video+on+YouTube#p20877

The posts in that topic also link to the article by Dahn:

http://myimiev.com/forum/viewtopic.php?f=23&t=2325&p=20979&hilit=Dahn#p20979

 

[mdbuilder]

So, I've steadily been gaining capacity in my traction battery (as tracked in the thread EVBatMon Readings), but voltage rise/sag seems to be getting worse. During the last quick charge, input current seemed low for the SoC. Coming home from 7 Springs, cell voltage was dipping below 4 volts above 90% charge, which I don't remember it doing before.

Am I going crazy, or can internal resistance increase without affecting capacity?

A few thoughts based on years of experience with model airplane lipo's.

Other than the first few cycles when the pack is 'breaking in', IR does nothing but get worse over time when measured under the same condition of pack voltage / SOC, temperature, and how recently was it charged.

IR is lower at higher cell temps within the normal operating range. If you had a pack you charged to 100% and its temp was 90F it will show a lower IR than the same pack at 100% that is 60F or 70F etc.

I wonder how the car is measuring AH which EvBatmon reports? I suppose the two choices are the BMS or the charger. In either case, is it removing the count of capacity that is shunted off to balance the cells at the top? I would guess not. And that implies if there is a lot of balancing going on at the end of charge it may report higher Ah counts which aren't really actual capacity. On the other hand, your capacity growth is quite a lot - hard to attribute to balancing activity eh?

 

[PV1]

So, starting at 36 minutes, he explains what happens when we see a failed cell in an i-MiEV battery pack.

My pack is balanced, and has been since I got it last summer, so balancing is likely not a factor in the growing capacity. I would likely attribute that to sustained higher temperatures and taking the car on increasingly longer trips, which allows the BMU to align itself to the battery pack better, sort of like doing full battery cycles on a cellphone so that it shuts off at 1% instead of 10%.

I will be pulling my CaniOn logs and investigating, but the pack was a tad cooler during the last quick charge (about 80-85 F IIRC), and was considerably cooler during last night's drive home (pack average of 70 F, which climbed to 76 F as I drove). The pack also seemed more in balance at 33% SoC than it did before, but I'll be able to tell for sure once I pull my logs.

The video did an excellent job of demonstrating what goes on inside the cell.

 

[PV1]

Guess I'm going crazy . Just pulled my logs and analyzed some data points.

At 41,910 km, with 96% charge at a pack average of 19.2 C, a 100 amp draw nets a pack voltage of 346 volts (3.931 V/cell).

At 57,398 km, with 92% charge at a pack average of 22.5 C, a 150 amp draw nets a pack voltage of 344 volts (3.909 V/cell).

Just today, at 66% charge, a 110 amp draw pulls pack voltage to 338.2 volts vs. 345 volts at near 0 amps (3.843 V/cell vs. 3.92 V/cell).

At 41,910 km (my first data point), 0 amps pack voltage is 352.9 (4.010 V/cell vs. 3.931 V/cell under load).

So, in summary (and in US units):
100 amp voltage sag at 26,041 miles and 66.5 F was .079 volts per cell.
110 amp voltage sag at 35,740 miles and 76.1 F was .077 volts per cell.

 

[mdbuilder]

so, 7 vs. 7.9 milliohms. Seems pretty good for such large cells. I think they are not directly comparable though if I followed correctly one is at 96% soc and one later at 66%?

 

[PV1]

Here's a fairer comparison between the 96% and 92% comparisons:

100 amp Voltage sag at 26,041 miles and 66.5 F was .079 Volts per cell.
100 amp Voltage sag at 35,665 miles and 72.5 F was .157 Volts per cell.

(the difference between 100 and 150 amps was very small (less than 1 volt for the pack).

 

[mdbuilder]

Lol! I don't think that helps <g>. That comparison says the IR has doubled which sounds really bad. Maybe it is but maybe it is other factors.

I presume you are extracting these from a data stream during a drive? It would be better if you want comparable measurements over time to set up a repeatable test. Which might be something like let the car sit and charge over night to 100%. In the morning very gently drive a short distance to a straight piece of road then do your short acceleration spike to capture the 100+ amp voltage dip. That way you are starting from the same near 100% state of charge and the temp measurement is the actual pack temp not a surface temp. In other words, if you are driving normally (current up and down regularly) for a period of time the temp inside the cell is going to be warmer than the measured surface temp as you are working it on the interior.

 

[PV1]

These are all points from driving.

I may have to go back and compare my quick charges. Driving is too variable.

 

[Lic]

last time on qick charger I was looking for cells which is rich 4.105v first. It was cell number 61. When I was driving back from charger I checked if I will see same cell with lowest voltage under load and it is it - number 61. So It's look like this cell has worst internal resistance. It's interesting if somebody else have similar result.

 

[JoeS]

last time on qick charger I was looking for cells which is rich 4.105v first. It was cell number 61. When I was driving back from charger I checked if I will see same cell with lowest voltage under load and it is it - number 61. So It's look like this cell has worst internal resistance. It's interesting if somebody else have similar result.

Hmm, that's disconcerting. I haven't seen such a variation, but frankly haven't been looking for it. When's the last time you did a full slow balance charge starting from below two bars?

 

[Lic]

May be month ago. I don't see how it could be connected to cell with worst internal resistance.

 

[JoeS]

May be month ago. I don't see how it could be connected to cell with worst internal resistance.

Valid point, as I'm not aware of any 'healing' that may take place while balancing. All of my quick charging stops at 80%SoC so I never see that high a cell voltage when DCQC.

Just out of curiosity since I am at this moment completing a full full charge (as I'm about to take a long trip in the i-MiEV in a few minutes) I checked the voltages but unfortunately all the cells are happily sitting at 4.105v at 99.0%SoC and still balancing. Next time I go to fully charge I'll turn on CaniOn and see if any individual cell goes up first; for that matter, I'll try to remember to have a look at the cells when I come back today to see if there is an outlier, as I expect to be down to two bars by the time I get home... that reminds me, have the 12v electric fleece blanket handy. :roll:

With your one weak (high resistance) cell I'd be paying attention to it as that alone will probably be limiting your range and if it gets worse will make your pack a candidate for replacement (if still under warranty).

 

[Lic]

It's already limiting speed of quick charge. last time when I was on quick charger temperature outside was about 4Cel. On battery I saw was something about 13-19Cel and still CC phase finished at 67%. I continued to about 75% until power went down to 3kw and I decided to leave quick charger because it's no sense to do slow charging on quick charger. So it was 2 cells which reached 4.105v first. I forgot other one, because I didn't see it after on my way home. Cell 61 could have worst internal resistance because of lowest it's own temperature. We have temperature sensors but still we don't know real temperature of each cell. But definitely I will look for this cell and I will try to look for "winners" in charging and discharging. It would be good to have ability in canion to see internal resistance of each cell. It should be not so difficult to do this calculation.

 

[PV1]

On one hand, 3 kW at 75% is very low (quick charging my cars are faster than level 2 until about 90-95%), but still getting full charger output at 67% is pretty good (assuming it's charging at more than 30 kW).

Do you know how to browse the CaniOn databases? If not, I can PM you instructions on how to do so.