Making use of low C rate LiFePo4 cells

[amberwolf]

It is due partly to that, and also to the low resistance of whatever connection you have between the cells.

But there is always *some* resistance between cells in parallel, and so they are not always exactly the same voltage. As long as the cells that are supplying the most current are not sagging in voltage, then those plus the other cells in parallel will be at the same voltage.

As soon as the cells supplying most of the current begin to sag (internally), current will begin to flow thru the parallel connections from the other cells that aren't sagging, and there will be a minute voltage drop across the cell interconnects during that time--for that time, the paralleled cells will indeed be at different voltages.


Really, that's irrelevant, but just thought I'd point that out for those that dont' see how they can be at different voltages in parallel.

 

[dogman dan]

Some kind of device that limits amps of the low c rate packs would be the way to go I think, even if it wasn't particularly needed.

Not a cut off or full bms, but just an amp limiter that would keep the power flowing from the low rate cells to the high rate ones. So the effect would be a 1c flow from the weaker cells that was pretty continous. That would be the way to go, for a larger pack on a motorcycle or something. The effect would be like a charger running while you rode.

But why bother, with only 10 ah of cells at risk for a bike.

 

[joe tomten]

Not the same situation but:

When I parallel my 48v15ahr ping to same sized keling (not a the cell level) and I watch the charging and discharging wattage (and usage) with 2 watt meters -

I see that the keling (5ahr cylind. cell format) takes up most of the load earlier on in the charge cycle, as if it sucks up most of its charge at 3.2v/cell, compared to the ping preferably likeing 3.3v/cell.

When I unplug the charger at full battery - The ping feeding the keling with say 10 or 15 watts cause of the pack voltage differential, and this evens out after some riding.

I wonder if the flat part of the volt/capacity curve differs between the a123 and the lifepo4 cheapies that you have, and how much of an effect this would have.

 

[John in CR]

Some kind of device that limits amps of the low c rate packs would be the way to go I think, even if it wasn't particularly needed.

I think the different IR resulting in a drastically different sag will accomplish that result.

But why bother, with only 10 ah of cells at risk for a bike.

True, however, low power or not they're part of my lithium stockpile, so they're my babies. I'll leave the battery killing to rich guys who have fun doing it like LFP, and those who don't understand that BMS stands for Battery Murdering System. I love that acronym, because it challenges the e-guru's to come with a truly bulletproof BMS. Even the manufacturers haven't figured that out, because otherwise they would be directly attached to each cell. Then we'll have it made, because cells will self limit their current, and instead of a whole pack shutting down only the low cells will turn off.

John

 

[dogman dan]

Yeah, don't we all wish we had Lukes kill stuff for the fun of it budget.

I guess what I was trying to say was relying on the resistance to do it on a bike size pack is one thing, but if you did it on a large EV pack you'd want to be more sure the low discharge rate cells never saw a high rate, like at the low voltage end of the discharge for instance. On the bike, you can just be the bms and not use all the throttle for the last half.

 

[John in CR]

I did some testing today down to over 75% DOD, and it worked exactly as predicted...actually even better for my type of short blast stop-n-go riding. I say better because I was discounting the low power cells contribution under load and how much they recharge the A123's when off the throttle.

Here's what I did (sorry but video watching multimeters and the iCharger would have been silly):
I paralleled my 3 worst A123 cells, since I didn't feel like soldering on good ones for just a test. I also couldn't discharge more than 2C, so fewer and worse A123's should give comparable results to more and better A123's at higher power. Each A123 has only 1.9ah of capacity. I put that group in parallel with a 10ah LiFePo4 cell that I previously tested to have a bit over 9ah of capacity at 1C. I put my multimeter on the 20A limit current setting as one leg of the connection, but its resistance was giving me far to different results depending on which end I put my discharge mains, so I paralleled it with a copper wire that reduced the resistance 4X, which put the overall resistance of that leg far below the internal resistance of the low power cells themselves, so I'd get valid results. I varied the current as well as stopped and started it, and watched what happened with the current flow between the batteries.

Results:
The low power cell was only an identical voltage as the A123 group to start the test. Once I started discharge, the A123 voltage was always lower, so current was always flowing from the low power cell to replenish the A123's even after stopping for as long as 10 minutes. I was also never able to get the current higher than .5C out of the low power cell, though that would have to change in a continuous high power situation continuing well past the capacity of the A123's. In that case the increase in voltage sag would be quite obvious, but in my type of use the change in voltage sag will be much more gradual because the A123's are being continuously recharged.

Unless someone requires continuous high power and deep discharges I don't see any downside to this approach, and I'm sure my plan to team 23s A123's with 20s Konions or lipo will work equally as well. For lipo that 20s/23s combo looks like great protection from dangerous over-discharge. Bulkier and heavier, yes, but high risk and can't utilize full capacity offset much of the lipo benefit.

John

ps- The good DrBass should give me a commission for helping sell those DeWalt packs he put up in the for sale section.

 

[John in CR]

I did another simulated ride test with longer periods of battery drain including fewer several minute stops, and this time I took it down to 90% DOD. The current coming from the low power pack still never got to half of the current draw, so the continuous recharging of the A123's from the lower power pack. This time the current flow during stops did stay higher than before, still 2-3A even after a few minutes and as high as 4A.

For my use it appears that this strategy for a pack will work perfectly for my kind of riding. I'll use 4 or 5p of A123 M1's depending on which fits together best for a physical arrangement, and parallel one 10ah nominal lower power cell with each A123 group. That will form one high power 18-20ah 23s pack, and I'll cannibalize some pings I have (23s15ah of good cells from my 2 version 1 36V packs and a broken 48V20ah pack I bought from someone plus spare cells). Even after I inevitably mess up some cells during cannibalization, I should be able to get 25ah23s of Pings, which will team up nicely with the 8p23s M1 pack I finished up last week. 43ah23s nominal of lifepo4 in a high power pack sounds perfect for the 2wd street beast I have in the works.

John