Matched IR, in series or parallel
1 - 20 of 39 Posts
Joined
·
7,791 Posts
Hi RIP,
I assume you are talking about how to sort and group cells by the spec sheet internal resistance. My opinion would be to keep the parallel groups as close as possible in the cells' resistance. That way they will current share more equally.
Series connected cells (or groups of paralleled cells) would not be sensitive to differences in resistance since the current will be equal by virtue of the series connection. So they should charge and discharge equally.
If you intend to use the battery hard, you might try to keep cells having higher resistance on the outer edge of the pack as they would heat more.
Even though the resistance values look small, there is actually a plus/minus 14% spread.
Did you purchase individual cells or multicell packs?
Regards,
major
Joined
·
3,218 Posts
I think, EVFun, he is asking how the cells should be grouped according to internal resistance.
Ideally, you want the equivalent resistance of each "cell" in series to be identical. So you parallel cells with various internal resistance figures to result in the same overall resistance as every other bank of cells that are paralleled together to make up a series string (ie - a battery).
The spread of 9 milliohms to 12 milliohms is pretty wide - but very typical of these Chinese cells, regardless of chemistry or manufacturer. Parallel cells so that each group has the same mean internal resistance as all of the cells considered together. Ie - if you have 10 cells at 0.009 ohms, 20 at 0.01 ohms and 10 at 0.012 ohms then the average (mean) resistance of all 40 cells is 0.01025 ohms. Aim to make each paralleled cell group have that resistance and you'll have the best possible balance w/r/t voltage drop under load for each "cell".
An additional consideration is to group by capacity, but that's sorta making things even more complicated. Might want to use a computer to sort that mess out. Not even gonna try it here.
Ideally, you want the equivalent resistance of each "cell" in series to be identical. So you parallel cells with various internal resistance figures to result in the same overall resistance as every other bank of cells that are paralleled together to make up a series string (ie - a battery).
The spread of 9 milliohms to 12 milliohms is pretty wide - but very typical of these Chinese cells, regardless of chemistry or manufacturer. Parallel cells so that each group has the same mean internal resistance as all of the cells considered together. Ie - if you have 10 cells at 0.009 ohms, 20 at 0.01 ohms and 10 at 0.012 ohms then the average (mean) resistance of all 40 cells is 0.01025 ohms. Aim to make each paralleled cell group have that resistance and you'll have the best possible balance w/r/t voltage drop under load for each "cell".
An additional consideration is to group by capacity, but that's sorta making things even more complicated. Might want to use a computer to sort that mess out. Not even gonna try it here.
Joined
·
7,791 Posts
Hi Tess,
You know I hate to disagree with you
And Woodsmith says we should not resort to insults, namecalling and threats. So let's look at a parallel set of cells. Take 4 for example. Use a simple model for the cell of an ideal capacitor and series resistance. Draw the circuit of the 4 cells connected. Use the same C value for each, but different resistance values.When you put a load on the parallel group, each capacitor will see a different current depending on the individual resistor. The voltage on each capacitor will decrease as charge is removed. Once the load is removed from the parallel group, the voltage on the 4 capacitors are different. So, current will continue to flow within the group of 4 parallel cells until all 4 reach equilibrium.
This is called charge redistribution. It is undesirable because it is current flow within the system which does no useful work for you. Worst case; it can cause excessive heating. Lithium cells have high Columbic efficiency, so there will be little loss of capacity, just the energy lost in resistive heating.
And I guess I don't see a good reason why series connected cells need to be R matched. There you should match capacity.
Regards,
maj
Joined
·
3,218 Posts
Who are trying to kid, you crusty curmudgeon! You love to disagree with me!?You know I hate to disagree with you
Yep, I do agree, but I saw this as the lesser of two evils at really high currents (relative the the cell's capacity) because if a paralleled group consisted of all high internal resistance cells it might trip LVC prematurely. I was keeping in mind what RIPPERTON here would likely be using the cells for - namely, in racing motorcycles. Of course, you have some practical experience in this matter whereas mine is entirely fabricated, so...
Joined
·
2,050 Posts
pack will be 32s 12p. 52ah 128v 6.6 kWh
4800 peak Amps which is irrelevant because that will never be used
controller can draw 500A cont which is 41Amps per cell
the pack will be ventilated with 1mm air gap between each cell
be used over 20 minute stint to complete 30km at race pace (90kmh av ?)
counting on burning 200Wh/km
theres going to be some very aggressive regen going on too
Just did a cell count out of the 450 cells I bought (need 384 for the pack)
57 cells are 0.009 ohms
234 cells are 0.010 (52%)
71 cells are 0.011
14 cells are 0.012
makes 376. theres a few 0.008's that I will throw in
there were a few freaky ones with 0.004 and up as high as 0.034 which
were sold to a friend for destruction testing.
I also thought to place matched cells in series to minimise disparity during charging and discharging. theoretically all cells in series would charge equally but balance out in parallel
4800 peak Amps which is irrelevant because that will never be used
controller can draw 500A cont which is 41Amps per cell
the pack will be ventilated with 1mm air gap between each cell
be used over 20 minute stint to complete 30km at race pace (90kmh av ?)
counting on burning 200Wh/km
theres going to be some very aggressive regen going on too
Just did a cell count out of the 450 cells I bought (need 384 for the pack)
57 cells are 0.009 ohms
234 cells are 0.010 (52%)
71 cells are 0.011
14 cells are 0.012
makes 376. theres a few 0.008's that I will throw in
there were a few freaky ones with 0.004 and up as high as 0.034 which
were sold to a friend for destruction testing.
I also thought to place matched cells in series to minimise disparity during charging and discharging. theoretically all cells in series would charge equally but balance out in parallel
If the cells are used for racing, so high discharge currents, I wonder how relevant the Rs measurements are, since the effective resistance at higher currents will be very different due to concentration polarization, as described by the Warberg factor, according to the CM video. In that case it seems what would matter is how much the W factors differ between cells, which you would have to determine yourself by discharging them at high rate. The Warberg factor appears to be the "effective" series resistance at high discharge rate, and is much larger than Rs, resulting in much more cell heating, and larger drop in cell voltage.
Joined
·
88 Posts
Daniel, I am surprised if your Enerland cells have an IR as high as 9-12 milliohms. For racing you should get considerably higher discharge power from your LiFeTech X1P cells since they all should be under 3 milliohms. After you ordered your LiFeTech X1P Power cells the factory started labelling the IR of the cells on the the top left corner of each cell tray as part of the QC cell grading process. This makes it much easier to match cells for packs used for racing applications. I would be suurprised if any of your cells have an IR of higher than 3 milliohms since all the the LiFeTech X1P cells are all typically in the range of 2.5-3.0 milliohms.
Attachments
Joined
·
1,378 Posts
Major, I don't have any empirical data for this, so not disagreeing, but I don't understand why that would be the case. It seems end batteries would lack warm neighbors on one side, and would have a longer battery cable to draw away heat. Why would end cells run hotter despite more ability to shed heat?
Joined
·
7,791 Posts
Hi David,
Maybe I didn't state that clearly, or you didn't read it clearly, or I don't read your post clearly
Anyway....One would assume all the cells in the battery discharge at the same current. Therefore cells which have a higher resistance would generate more internal heat than cells with lower R, right? So what I was saying, it is smart to put those high R cells which would get hotter on the edge of the battery pack. Cells on the edge would be able to shed heat better than cells in the middle, right?
I think it is common to see cells in the middle of densely packed batteries get hotter during heavy current discharges. That has been my experience.
And David,
Do you agree about matching internal resistance on the parallel sets?
Regards,
major
ps...
{On those Kokam cells, for some reason the most positive terminal on the battery gets the hottest. The rest of that most positive cell seemed about the same temperature as the other cells and the cells (edges which could be measured with IR gun) in the middle ran a few degrees hotter than the end cells. But for some reason, that positive terminal ran significantly hotter, on multiple tests, on multiple batteries. Anybody care to guess why?}
Joined
·
3,218 Posts
Yeah, I'd like to know this too, since I think major is full of crap...
The terminal probably fans out internally and so you are seeing increased heating because of current crowding... that's my guess, anyway... Keep in mind I've never even seen one of these cells, so...
I guess that things that are electrically isolated from each other are also thermally isolated. That means that the higher temperature at the positive terminal indicates that whatever process is responsible for the internal resistance of the battery is active at the positive plate. For the manufacturer, it means if you want to reduce resistance, look at the positive plate structure and reactions.
Gerhard
Joined
·
7,791 Posts
It is just the positive terminal on cell #14 which is at the positive end of the 14 cell pack. All cells are identical having a positive and negative foil type terminal. Except for the cells on the ends, the foil tabs are compressed together with the adjacent cell tabs with aluminum bars. The same aluminum bars are used on both ends, just compressing a single cell tab with a cable lug screwed to it. And from memory, it is like 20 or 30 degrees hotter than anywhere else on the battery. I have some notes about it in the lab I'll look for. Here is a photo of the battery.
Joined
·
1,378 Posts
Got it, no confusion now. Mr. Literal here read "keep cells having higher resistance on the outer edge of the pack as they would heat more" as cells on the outer edge would heat more than in the middle, and that was desirable for some reason. Here's a clarifying restatement for the Literal Readers like me in the crowd: "keep cells having higher resistance on the outer edge of the pack since they heat more."major said:Hi David,
Maybe I didn't state that clearly, or you didn't read it clearly, or I don't read your post clearly
I have not tested this myself, but my intuition is it is better to match resistances for the paralleled cells. Here's a roundabout way to show it: A local EVer mixed cells of different capacities, I forget the exact capacities but it was in a 2:1 ratio (like 50 and 100 Ahr). He figured the 50 Ahr cells would have twice the resistance, and so would supply 1/2 the current of the 100 Ahr cells. He figured the cells would reach empty at the same rate. He was right, it worked for him. His daily commute almost completely emptied the cells each day, so it wasn't a case of shallow discharges. BTW his drive home was a long uphill with a big elevation gain.major said:
Running with this, suppose he had cells of 2:1 internal resistance but the same capacity. It seems to me one cell would reach empty twice as fast as the other cell, greatly stressing it.
I'd guess it's a thermocouple effect. You have current going from metal A to metal B for the + terminal, but from metal B to metal A for the - terminal -- for different metals generally one way is hotter than the other.major said:
Joined
·
7,791 Posts
Interesting theory David. The Li cells have one aluminum and one copper terminal. Can't remember which is which, positive negative. But between the end cell terminals and the copper cable lugs I have an aluminum clamp bar. Hopefully I'll get back into testing this month. I think I will make a copper bar clamp to replace the aluminum one on the positive end and see if it makes a difference.
Is there such a thing as a copper aluminum thermocouple
as? since? I know what I mean when I type it. Sorry, but all you have to do is ask. If I confuse you, chances are others are confused also. I'd rather be questioned than leave something misinterpreted
major
Joined
·
308 Posts
That's an interesting theory, I could be wrong but I think the negative terminal is a nickel coated copper foil and the positive is aluminum foil so wouldn't thermocouple effect be most pronounced at the negative terminal? It is easy to test the thermocouple theory though, does the effect reverse when charging?
Other theory, the aluminum to aluminum connection at the positive terminal is higher resistance then copper+nickel to aluminum connection at every other connection point?
Joined
·
2,050 Posts
Maybe so Armin but power isnt what I want
I want Power to weight ratio.
I can go 12 parallel with the LiPo's but only 4 parallel with the LifeTechs
and still end up with a 20kg lighter pack.
Plus I wouldnt contemplate for a second to not have air cooled LiPo's
so higher running temp isnt a prob.
The LifeTechs will make a good commuter battery....one day
1 - 20 of 39 Posts
Join the discussion
