Check my understanding (calling on folks who understand IR)

Internal resistance of a battery becomes easier to comprehend when you look at the processes in the battery in mechanical terms. This is a valid way to look at it because in the charging and discharging process, you are physically relocating ions from one location in the battery to another location. The ions have to make their way out of a graphite solid, travel though a separator material, and then travel into another solid made out of iron and phosphate atoms. As stuff moves past other stuff, it bumps around and loses some energy as heat, so the materials of construction for the battery are selected to make the pathways as open as possible.

However, if manufacturing quality is poor, these pathways might not be optimally placed. Or if a battery is abused, the pathways might get damaged, or plugged by molecules that were created in side reactions from over charging. The materials the pathways are made out of can get physically cracked and degraded through normal use. This is why measuring internal resistance can tell you if one battery cell is in similar condition to another battery cell. But you have to do the measurement under the same conditions because the state of charge (how many lithium ions are already parked in the pathways), temperature, and rate of charge or discharge (how many lithium ions you are trying to move at once) all affect the reading.

I found this animation to show lithium ions moving in the anode material. Before you run it, grab the image and rotate it around, and from one direction you will see that it has large openings. These are the pathways the lithium uses. Thought this might help some people who are visual learners.

Thank you, this is a good and intuitive explanation of internal resistance on a chemical / physical level!

Connecting your comment with Cinergi's comment on the greater practical significance of IR in parallel, I now see why many/most of the videos on youtube about Internal resistance center on testing used RC batteries, or testing small cylindrical cells.

It's entirely possible that some of these large format prismatics actually undergo a capacity change during their initial few uses (and during their long passive period sitting on the pallets and boat). It stands to reason, if there are mechanical changes (electrolyte expansion, case-widening, bloating, etc.), then that physical change might alter ion mobility in some way.

A controlled test would be hard to conduct, but some of the folks seeing slightly lower capacities might do well to commit the time to put one cell through a number of full cycles. It would be pretty interesting to see how repeatable the results are.

nebster said:

It's entirely possible that some of these large format prismatics actually undergo a capacity change during their initial few uses (and during their long passive period sitting on the pallets and boat). It stands to reason, if there are mechanical changes (electrolyte expansion, case-widening, bloating, etc.), then that physical change might alter ion mobility in some way.

A controlled test would be hard to conduct, but some of the folks seeing slightly lower capacities might do well to commit the time to put one cell through a number of full cycles. It would be pretty interesting to see how repeatable the results are.

Click to expand...

This speaks directly to one of the points I was trying to make.

As Dzl pointed out it is stupid simple to measure capacity with great specificity. That is the alluring part of doing this.

However there are a few pitfalls here.

If these specific measurements are exactly repeatable then I would agree it is a simple thing we can all do. That is not what I am seeing however.

With a lack of repeatability it raises a few general concerns:

Is the constructing and testing method sound? For many folks I believe it is.

Are the measurements accurate? For folks using multiple measurement tools or calibrated instruments this can be controlled also.

Is the test subject stable? This may be where majority of the problem lies.

Specifically there may be significant variability within one test subject sample. In that case you will need a series of tests to see what is a significant difference not just a measurable difference.

To figure out a statistically significant difference you will need a number of tests to speak with accuracy. I know just enough about statistics to be mildly dangerous but in general you need more test data with a variable test subject to tease out real differences. That means a series of more tests done in a fairly repeatable fashion.

I have no idea if the number of tests should be n=5, n=50 or even n=500 but I firmly believe that one test of a cell may not be adequate. Add 16 cells in a complex interaction to the test and n definitely is not 1.

So I guess I am advocating for either repeatable tests or a larger series of tests before folks can accurately claim that capacity is X.

The other solution is to conclude that the results are "close enough" to determine the value and accept there is a certain percentage of unknowable stuff (inaccuracy) that cannot be eliminated without a very time consuming and expensive testing process.

Can you accept 5% or 10% variability? That's probably close enough for me. YMMV.

Dzl said:

technically this should be minimum capacity or rated capacity but usually they are one and the same).

Click to expand...

My point precisely regarding lifepo4 batteries specifically. Until vendors and buyers demand specificity in actual rated capacity (the minimum a battery can provide) then it will continue to be unnecessarily ambiguous.

Dzl said:

'Nominal Capacity' is all one term.

Click to expand...

Nominal capacity is only one term because it's allowed to be. Using it actually compounds the problem. Don't we have the capability to accurately measure minimum capacity and therefore give a battery that rating? I think we do, but I could certainly be wrong because I'm new to this.

Knowing that a battery or solar panel is meant for 12V, 24, or another voltage application is completely necessary but completely different from knowing what AH capacity a battery has.

Can I use a specifically rated (i.e., 280AH) battery in any voltage application so long as they are connected in such a way to provide the required output? I think I can, but I'm new with this.

Then again, perhaps all of my angst is due to the fact that lead acid battery manufacturers have been getting away with selling batteries that are pretty much only usable for about half of their rated capacity if any kind of acceptable lifespan is to be expected of them, and buyers like me went along with the ruse rather than insist on more specificity regarding the "actual" usable day-in-day-out AH capacity in all voltage ranges.

nebster said:

It's entirely possible that some of these large format prismatics actually undergo a capacity change during their initial few uses (and during their long passive period sitting on the pallets and boat). It stands to reason, if there are mechanical changes (electrolyte expansion, case-widening, bloating, etc.), then that physical change might alter ion mobility in some way.

A controlled test would be hard to conduct, but some of the folks seeing slightly lower capacities might do well to commit the time to put one cell through a number of full cycles. It would be pretty interesting to see how repeatable the results are.

Click to expand...

So far mine are repeatable, down to the Ah.

cinergi said:

So far mine are repeatable, down to the Ah.

Click to expand...

How many cycles have you applied?

nebster said:

How many cycles have you applied?

Click to expand...

2 on the tester and approximately 10 as part of a battery. I'll be performing more on the tester on the high and low cells after I've tested them all on the tester. I still have 28 more cells to test on the tester.

Sounds good -- I'll be intrigued to see if anything changes over time.

nebster said:

How many cycles have you applied?

Click to expand...

I haven't tested the individual cells. But I have run my pack through 5 or 6 cycles with low C rates and the end result is the same. I get 272ah's. I have one cell that is a bit weak.

cinergi said:

So far mine are repeatable, down to the Ah.

Click to expand...

And I believe you have some of the best tests detailed here. Please know I’m not criticizing folks, just trying to understand.

In your thread, posts 111 and 112 indicate different results from consecutive tests I thought.

Perhaps those were different test subjects?

JoeHam said:

If these specific measurements are exactly repeatable then I would agree it is a simple thing we can all do. That is not what I am seeing however.

Click to expand...

Can you point to any examples of this?

You raise a bunch of good general points about testing in general, specifically if we are trying to test to tenths of an amp-hour or a specific amp-hour maybe, or if we were talking about just one or two examples. But we aren't talking about that level of precision, or at least I'm not.

With the level of specificity we are looking for, I think you may be posing a bunch of problems that, in practice aren't nearly as much of a factor as you imagine they might be.

Off the top of my head, I can recall two people who conducted multiple back to back capacity tests (both have commented in this thread already), my recollection is that in both cases they observed the same general trend. Moreover, what they observed over multiple capacity tests is corroborated by more cases of people who have done single capacity tests.

While I don't doubt there is some imprecision and probably some issues with methodology here or there, and I would not trust test results down to the amp-hour, its hard to dismiss the general trend as an outlier or a collective inaccuracy. The fact that people with different tools, in different locations, with somewhat different methodologies and levels of experience are observing the same general trend, goes a long way towards neutralizing the impact user error, or poorly calibrated / inaccurate tools might have. Particularly if you acknowledge, that lay people have been capacity testing lithium batteries for a long time, and to my knowledge, there has been no widespread problem semi-accurately validating the capacity of good cells, or no widespread observed phenomenon of lay peoples tests skewing low.

All that said, there is one variable that is still nagging at my level of confidence, temperature. I don't know how sensitive LFP is to temperature off the top of my head. From what I've seen many of the folks doing the capacity testing have met or exceeded every other part of the standard test conditions. Temperature is the one variable I'm not sure of, the manufacturer tests in the mid to high 70's, most people don't state what temperature they test at. I wouldn't think 10 degrees one way or another would have a significant impact, but honestly I don't know.

I imagine that temperature differences is the least variable factor at work but that is just my guess.

@Dzl I posted one example of what you asked for just above your request for an example ?

It was posts 111 and 112 from this thread which happens to be one of the most thorough:


What is not clear to me is if those are two different sample cells or assembled batteries hence my asking for clarification. One was 262 and one was 270.

Sillyputty said:

My point precisely regarding lifepo4 batteries specifically. Until vendors and buyers demand specificity in actual rated capacity (the minimum a battery can provide) then it will continue to be unnecessarily ambiguous.

Nominal capacity is only one term because it's allowed to be. Using it actually compounds the problem. Don't we have the capability to accurately measure minimum capacity and therefore give a battery that rating? I think we do, but I could certainly be wrong because I'm new to this.

Click to expand...

I agree with you that sellers should be held to a minimum standard. And agree that there is a problem. I respect your skepticism and desire to hold this market to a higher standard (as it should be).

But I think your mistargeting the term 'nominal capacity,' and misunderstanding the term / purpose. Nominal Capacity is a useful term if used properly and properly understood. Its not precise, but its not meant to be.

Manufacturers do specify a minimum rated capacity and they (and especially the grey market resellers) should be held to that standard if they represent the product as grade A or meeting capacity specifications. To my knowledge, all manufacturers do hold themselves to that standard, cells that don't meet the minimum capacity standard are sold off as Grade B cells. And the nominal capacity is usually the minimum capacity. Many reputable manufacturers use a 'nominal capacity' that is substantially below the typical tested capacity (under promise over deliver).

The big problem with accuracy/accountability is with grey market resellers, most of whom claim every single cell they sell is grade A or "A-". They are salespeople in a very competitive low margin market, and one with different norms than we are used to. The only time I put much faith in what a reseller says about grade is if they state a cell is Grade B, or in the past have acknowledged another cell is Grade B. Unfortunately the few sellers that are honest about this don't get a lot of our business, so they have a disincentive to do so.

Sillyputty said:

Knowing that a battery or solar panel is meant for 12V, 24, or another voltage application is completely necessary but completely different from knowing what AH capacity a battery has.

Click to expand...

I agree, there is a difference.

Sillyputty said:

Can I use a specifically rated (i.e., 280AH) battery in any voltage application so long as they are connected in such a way to provide the required output? I think I can, but I'm new with this.

Click to expand...

Depending on how you configure them, yes. 4 cells in series = 12v, 8 cells in series = 24v, 16 cells in series = 48v

JoeHam said:

And I believe you have some of the best tests detailed here. Please know I’m not criticizing folks, just trying to understand.

In your thread, posts 111 and 112 indicate different results from consecutive tests I thought.

Perhaps those were different test subjects?

Click to expand...

Yes different subjects.
And no worries - there's a LOT of information to consume coming from many different people. Keeping it straight is only possible for people with eidetic memories lol

JoeHam said:

I imagine that temperature differences is the least variable factor at work but that is just my guess.

Click to expand...

This would be my guess too, but apart from broad user error or broad measurement error in the same direction, its the only factor I could see that would lead to tests skewing low.

JoeHam said:

@Dzl
It was posts 111 and 112 from this thread which happens to be one of the most thorough:
What is not clear to me is if those are two different sample cells or assembled batteries hence my asking for clarification. One was 262 and one was 270.

Click to expand...

I'm not 100% clear on that either, so I will let @cinergi clarify, so I won't speculate.

...me not speculate? did you honestly believe that? ? I have to indulge myself with at least a bit of educated guessing here...

I believe what you are looking at is two parallel tracks of testing:
In post #111 (Capacity test #4) is a full pack capacity test (limited by the lowest cell to 262Ah). What you are seeing in the bottom right of the computer screen is the Orion-BMS measuring each cells individual voltage as the pack discharges
In post #112 (Cell test #3) may be a single cell capacity test (though I'm less clear on this, hopefully @cinergi can clarify)

Going back through the thread I think these are the results of the pack level capacity tests:
Capacity test #1: 252
Capacity test #2: 261
Capacity test #3: 261
Capacity test #4: 262

Does that look correct @cinergi

Dzl said:

All that said, there is one variable that is still nagging at my level of confidence, temperature. I don't know how sensitive LFP is to temperature off the top of my head. From what I've seen many of the folks doing the capacity testing have met or exceeded every other part of the standard test conditions. Temperature is the one variable I'm not sure of, the manufacturer tests in the mid to high 70's, most people don't state what temperature they test at. I wouldn't think 10 degrees one way or another would have a significant impact, but honestly I don't know.

Click to expand...

Temperature definitely affects capacity. Here's one study of small form-factor LFP:

nebster said:

Temperature definitely affects capacity. Here's one study of small form-factor LFP:

View attachment 28973

Click to expand...

Totally agree but most folks here tend to do all their testing in the same room or garage so that shouldn’t affect one person’s results.

It does make it impossible to compare results from a forum member in Antartica to one in Ecuador though on the same 280Ah cells ?

nebster said:

Temperature definitely affects capacity. Here's one study of small form-factor LFP:

View attachment 28973

Click to expand...

I know it is a factor, especially at very low temperatures. What I'm not clear on is how sensitive it is. In cell test #2, the Chargery is reporting 75*F, in cell test #4 the Orion shows 24*C both are just about exactly standard test temperature. I believe @cinergi mentioned his garage stays in the range of 60-70*F, and Gazoo's testing has been in his apartment so I would guess close to ~70*F as well.

I wouldn't expect a large difference at those temperature differences but I really don't know. Like the graph you posted, all the data I have seen on temperature usually only shows big gaps in temperature (extreme low, freezing, room temp, hot). No indication of what ~5-10*C difference would make.

@Dzl here are the results of a search:


I only went through the first few pages but there are a few people reporting discrepancies.

Most if not all are at the battery, not the cell level which reinforces the need for n>1 in a complex arrangement of cells