LiFe cell internal resistance

BB.

A good first pass approximation.

-Bill

Suntrekker

OK and I’ll keep you posted.  Thnx again.... 

Suntrekker

Discharger, work in progress.

Suntrekker

 

BB.

I would suggest bolt up connections (to battery/etc.)... Alligator clips are not very reliable for high(er) current applications.

Bolts up also limits the chances of a connection being knocked off/shorting--But probably not a big issue here.

-Bill

Suntrekker

Makes sense, thanks!

Suntrekker

And the power resistors/heat sink are Hot to the touch. 

BB.

Sounds good... If they start "smelling hot/smoking", more heat sinking/air flow (set plate sideways to increase air flow).

But the resistors should be relatively cheap and easy to replace if ever needed.

-Bill

Suntrekker

the 10 awg wire is cool ; )

Suntrekker

Bill, I’m bench testing and see there’s a discrepancy in charge times between cells by approx. 3 minutes.  Is it a big deal in your opinion, maybe place the slow charging cell upfront (in series) to set the pace?  

BB.

I am not quite sure I understand your question... If you have a single series string of cells, all cells "see" the same amount of current--So position does not matter.

There have been members here that for their larger lead acid battery bank swear they have seen battery/cell position sensitivity to physical position. One end of the string position seemed to have the "lower SG" readings at the end of a year... And they would "rotate batteries" to even out the "wear".

You can try swapping the cell in question to the other end of string and see if you find a difference. My guess you would not unless one end of the string is hotter (poor ventilation, in sun, etc.) and the other end of the battery bank is well cooled.

The difference in charging time... Also depends on the charging current. Some examples:

• 1C charging rate (1 hour rate): 3 minutes / 60 minutes = 0.05 = 5% "difference"
• C/10 charging rate (10 hour rate): 3 minutes / 600 minutes = 0.005 = 0.5% "difference"

5% difference could be worrisome with new cells... 0.5% difference, not at all.

One of the reasons we top (typical for solar systems) or bottom balance... You want the cells to reach 100% SoC at the "same time"... If you have 5% difference, that means you can "safely" only recharge to ~95% State of charge (if one cell is 5% less capacity, it reaches 100% while rest of unbalanced cells are still at 95% SoC).

Same issue with "bottom balance"--All cells reach 0% at the same time... Rather than one cell reaching 0% first and the rest at 5% (or whatever)... And the rest of the cells "reverse charge" the one cell to 0% to negative SoC--A cell killer for most charging chemistries.

-Bill

Suntrekker

I guess what i’m trying to say is the internal resistance between my LiFePO4 cells is notable due to faster and slower charge times.  It’s an observation, not sure if it’s an issue

Suntrekker

What I mean for example, there are two individual cells at 3.302 volts.  Then I discharge each cell to 3.270 volts.  Next I charge each back to 3.302 volts and note the charge times.  I find one cell reaches 3.302 in 1.5 minutes whereas the other takes 3 minutes.  I use a CC of 3.50 amps. Does the different charge times matter?  And should I consider this when placing cells in series (24V)?

And Thnx for the detailed explanation, good to know!

mike95490

Are you sure it's internal resistance and not different cell capacity ?  Constant Current avoids the IR problem - so I'm leaning to different capacity.

BB.

Generally, for series cells, you do want to "match performance" in each string. If you have 3 cells with 10 AH capacity and one cell with 8 AH capacity, then the string has 8 AH capacity--The capacity of the weakest cell.

For "high performance"  battery packs (such as drones), the owners do go through a qualification process to group cells by capacity for each string/pack.

Just using time itself--I am not sure I quite trust that. I would by using an Amp*Hour meter of some sort to log each cell. Otherwise using the automatic charger cutoff time may be a bit less accurate (a little temperature/voltage variation, wiring resistance, etc.) can affect the time, but the cells could still have "identical" AH capacity.

https://www.amazon.com/s?k=dc+ah+wh+energy+meter

-Bill

Suntrekker

Yeah my crude test isn’t accurate and I haven’t done an actual capacity test.  All cells are identified as 280AH.  Anyway I’ll bottom balance, recharge to 90% (monitor) and call it good.  Thnx 

Suntrekker

Bill, I get different voltage readings from the battery posts as I discharge.  For example at the threaded screw 3.241V and at the flat section 3.275V.  Which is more accurate in your opinion?  Thnx 

BB.

The measurement to the "flats" is probably more accurate. The threaded studs could be something like stainless steel or other material and just not as good as a conductor.

HOWEVER, a DMM draws very little current when making readings... So I would be surprised that a "poor" connection to the studs should not make any difference (the meter input resistance is probably something like 1-10 million Ohms--So stud resistance is not the issue).

With low voltages and measuring into 1/100's to 1/1,000's of a volt, you can get into issues with voltage from dissimilar metals (steel, plating, meter leads, etc.) and have a thermometric offset:

https://meettechniek.info/measurement/faults.html

You could test the resistance of the studs by placing  a load on them (electric lights, etc.) and see the no load/loaded voltage difference between the stud and the lead(?) flat of the cell.

However, if you are actively charging/discharging the cell...

If the cell is being charged (current flowing--alligators clips to charger?) while making these measurements? If so, I would not make "accurate" or "precise" voltage measurements on the studs. But in this case, then the voltage on the studs should be slightly higher than the voltage on the "flats".

If the alligator clips are to your resistor load bank and under load, then you would see a voltage drop at the studs vs the flats of the cells (the flats are where the wire connections and current flow take place--The studs are just for mechanical holding forces).

So under load/charging, yes the flats are more accurate. Using a nut and washer to "bolt" (even finger tight) the meter leads to the battery should give you best results (more and consistent pressure from flats to lead tips).

-Bill

Suntrekker

Got it.  I may replace the alligator clips though I like the quick disconnect feature (and the DC charger has alligator clips).  ...I’ll measure the flats.

I read reduce all cells to 2.5V.  I guess that means 2.500V (not 2.510, etc.).  I don’t know how critical this is or maybe just splitting hairs.  Your thoughts?

Hopefully not a pita to match 8 cells.

BB.

I am not the Li Ion expert here... So I cannot say much other than when measuring down to 0.00x volts--You are in the area of needed calibration/checks and lab grade equipment. So don't drive yourself nuts.

The simplest method to balance is to wire all the cells in parallel then charge/let sit for a few days. Of course, the cells need to be relatively close first. Don't put a 100% charged with a 0% charged cell in parallel--That will be a lot of current.

-Bill

Suntrekker

Great.  Once all cells are around 2.5 volts, I read “bottom balance” by connecting them in series and charge to desired SOC. I like 26.6V (90%).  ...will keep you posted.

Suntrekker

Wow, about 6 hours to discharge one cell from 3.300 to 2.505V. And that’s only the first pass lol.  I think I’ll get the cells near 2.500, passive balance to achieve same voltage then connect in series and call it a day or week 

Suntrekker

I wish there was a faster way to discharge, maybe double the power resistor size.  But I read not to discharge quickly.  ...definitely a learning curve. Pic shows cell #1 requires more passes to get near 2.500V.  I’ll leave it for now and jump to cell #2.

Suntrekker

The first attempt to discharge a cell did not go well, I left it unattended (fell asleep, dumb).  The good news is I revived the cell for another project or spare and ordered a replacement from the same manufacturer.  Live and learn....

Suntrekker

This is interesting...tried discharging four cells to 2.500V but individual voltages varied; 2.537, 2.546, 2.560, 2.568. After many passes I decided to passive balance a bank of four.  At first cells read 2.558V but within 1.5 hrs, 2.563V.  My guess is they’ll be in sync soon.  I’m using 300A rated copper busbars for the battery (may help equalize quickly).

Suntrekker

Bottom balance cells for a 24V battery bank takes time (8 x 6 hrs). But the rewards are worth it! Reminds me of an old snickers commercial lol
https://youtu.be/WPOZbG7ibEE

Suntrekker

After 4 cells are passive (parallel) balanced, the voltage is now 2.6.  I’d like to discharge them ‘in parallel’ rather than individually to around 2.5V.  Can anyone comment on this practice, pitfalls?  Thnx!

BB.

Not anything major with matched cells in good shape.

Ideally, discharge connections should be kittycorner to reach other for even discharge currents between cells.

You could always increase your load bank current if you want to shorten the timeframe.

Bill

Suntrekker

I forgot about kitty-corner, I’ll give it a try!   ...no luck discharging individual cells to about the same millivolt. 

Suntrekker

Things are looking sunny ; )  

By discharging (kitty-corner) and passive balancing two (4 cell) banks, the difference between the banks is 1 mV, 2.570-2.571 (after numerous passes and discharge estimates).  Next passive balance 8 cells overnight then discharge near 2.5V - bottom balance complete for my application...what a trip 😬

Thnx Bill and All for guidance!