This week we take a break from the Escalade and the VW Thing and go back to a little bit of battery theory.
I’m learning to use stolen graphics with the green chromakeyer so that I can walk around in front of a green wall waving at things I can’t see while babbling things that could very well be true – or not.
This is a skill I’ve always wanted to develop.
I have learned the hard way not to read these open online forums like DIYelectricjunk and EndlessFear. They are not good for my head. But I did get drawn into a mailing list I kind of monitor when a gentleman whom should clearly know better started making some preposterous claims agains my dearly beloved LiFePo4 cells. Anyone reading this and acting on it could clearly lose significant ducats on their battery pack and so I spewed forth my usual blunt correction.
It led to some further comments from this “technical” group. The most striking was from the far left field who offered his “opinion” on battery swelling and including a specious appeal to authority via an unknown “university research paper” he was simply unable to locate at the moment.
I was struck by a recent television commercial for an insurance company where a young girl related what she had heard on the internet with the assertion that they would not put that on the Internet if it wasn’t true. And then she introduced her French Model boyfriend she met on the Internet who’s “Bonjour” somehow had faint traces of an Arkansas accent.
Unfortunately, the same world of misinformation that permeates popular EV discussion forums also extends to “university research papers”.
Which leads me to a preposterous personal claim. I’m really good at wading through piles of horseshit in my never ending quest for a pony. I’m not sure why or how. But I have spent years reading and writing technical information, some of it so esoteric it reduces itself both by its volume and obtuse verbiage to an audience of nearly dozens. To quote my hero Sir Winston Churchill, “Sir, your document by its very weight and volume defends itself ably from the dangers of ever being read.”
The form of journaled, peer reviewed scientific papers purports to weed out the noise and leave only gems of true scientific information. It fails somewhat grandly but perhaps DOES arm the reader with more clues. The unfortunate biproduct is it leads to a stilted, “scholarly” writing style that is essentially unreadable. This has become so idiotically ingrained in our University structure that it becomes an end unto itself. My wife’s advisor during her doctorate chided her basically for communicating well and urged her to alter the tone to more scholarly format. In a mirth spasm, I took the offending paragraphs and rewrote them deliberately to produce the most obtuse, schalarly garbage I could concoct, and deliberately altered the nut of information buried within to nonsense and the exact OPPOSITE of what she was saying and handed it to her. I dared her to submit it as a correction. She did and her advisor LOVED it despite the fact that it made no sense at all, and the information within it was totally untrue.
Rigourously structured and referenced documents do not BY LAW have to be irretrievably BAD at communication. They customarily are as an exhibit of poor communication skills on the part of the authors. It’s not technically REQUIRED.
In any event, it can lead to some handy shortcuts. If you include some of their more pedantic terms in your google search, often you will eliminate the total nonsense of the wider Internet and reduce your gleanings to the more narrowly defined nonsense of technical journals. For example, instead of OVERCHARGE use ABUSE TOLERANCE or instead of SWELLING use GASSIFICATION. Etc.
But even once you are down to the elite and magical world of “university research papers” which we all know are the LAST WORD on all things technical, you still have to use a bit of critical thinking to separate the wheat from the chaff. On any particular topic, I wind up with three piles.
NICE. This is quite good. He has a theory. He has observations. THey seem to match. It could work that way. Keeper pile.
SAVE FOR LATER. I don’t know. Save this and see what related papers bubble up over the next few months. This guy has some ideas but they are not very well supported by observation and the observations he does link don’t fit very well. THey COULD mean that. They could mean something else. If we collect more papers on the same topic, this will become clear.
NONSENSE, This paper should be downworded and posted on EndlessFear. He’s striving for a strained theory to explain unrelated observations randomly.
This all derives from the problem that our reach on many things rather extends our grasp. By this, I mean the following:
1. We can BUILD perfectly operational lithium ion batteries. They store energy. Deliver current. Can be recharged. And this can be repeated a lot.
2. We would like to build better batteries.
3. To do that, we need to experiment.
4. To experiment efficiently, we need to understand how the battery works.
5. We don’t REALLY know how the battery works.
Unfortunately, item five extends in all directions to everything. The quantuum mechanics of the atom itself, as described circa 2013 is very very different from that taught in the 1970s. Is it more accurate? Hopefully. It is certainly more complicated.
And we are further reduced to the fact that it doesn’t MATTER a whole lot. If our conceptualization through analogy is sufficiently accurate, we can make changes and the outcomes will be consistent with the changes made. And that’s good enough. Indeed some times a simpler analogy serves better to envision changes and their outcomes than a more accurate, but mystifyingly detailed one. And so we deal with the unviverse by analogy – essentially theories. Complex theories or simple theories. But they should be theories consistent with our observations and the observations should be reproducible.
This is the central difference between theories and opinions. Having an opinion about lithium batteries is not a very useful pastime. First, the batteries don’t know your opinions and are curiously disinterested in them. Your opinion will not alter their operation. In fact, unbeknownst to virtually the entire online body politic apparently, the cell won’t be terribly impressed even if you can get LOTS OF OTHER IDIOTS TO AGREE WITH YOUR OPINION. It’s not really a very democratic process. Voting on it doesn’t do any good at all – which at a higher level is the archilles heel of the peer-reviewed journal as well.
But it helps to have a process of observation, recorded data, and then to propose theories to explain those observations. But observations are tricky.
For example, if you find your battery mysteriously swollen, you might suspect something is wrong. If you test the battery by hooking it up, and find that it produces current, from that limited observation you might conclude that battery swelling is harmless and that the battery is operational The problem with that it is its kind of a half a question. You don’t have a theory as to why the battery swelled in the first place. And batteries do more than produce current. They also produce it over a set period of time, storing a set amount of power, and at a potential rate. If the rate of power produced decreases, then the fact that it still produces current is not really a win. If the current production is diminished to milliamps, then the fact that it makes current is almost moot.
If it makes full current, but doesn’t do it very long would indicate a strongly diminished capacity. And if it does it at full current, and at full capacity, but that capacity is now fading quickly with each subsequent charge, the life cycle ability of the cell is reduced.
And so you need to observe ALL the aspects of battery performance by observation to determine that the swelling of the battery was irrelevant. But worse, you really need a theory as to why the battery swelled in the first place.
Prior to measuring ALL the important parameters of the cell, and having a working theory as to how it came to swell to begin with, having an opinion on the matter is TOTALLY inane. It’s idiotic even to claim one.
My favorite papers work like this.
1. We made an experimental change to a cell.
2. It had the following measureable and reproducible result – in most cases a good one.
3. Here’s our theory of how the cell works at all – referenced from these other papers.
4. Here’s our theory of why this change might work.
5. Here’s our theory of how it did work. It improved such and such because of the interaction of a and b. Or it failed to improve such and such because A and B don’t apparently interrelate the way we thought they would in 4.
6. Which kind of calls into question all we know in 3. Or which tends to be consistent with what we know in 3.
The ongoing problem with these studies is scope. If you try to do too much with a test setup, you wind up with too much data that could mean too many different things. A single observation could support three different theories as to why that was observed. If you observe three things we are up to nine theories and we can’t even keep track of them.
If you do too little, you miss interactions with other aspects of the cell operation. Yes, it still makes current. Does it get hot?
I tend to prefer the narrow scope, but it leads to the problem that all answers lead to further questions.
All of this is an attempt to get you in the chair of how hard won the knowledge of lithium cell operation really is. Starting with a basic theory, that polyanions might make more stable cathode materials than simple layered oxides, at the University of Austin they built a cell that WAS more stable and had good battery characteristics as well. They tried several materials and one of them produced a good result. THere were of course 300 other materials they could have tried.
From that point, it is really cool to come up with a more detailed theory of how it all works so you could FURTHER IMPROVE IT. But it is a THEORY. And it is refined and improved continuously.
And you wind up with an entire shopping list of properties of lithium batteries. How fast they can be charged. How fast they can be discharged. How much charge they can hold. How either of those is affected by cold temperatures. Or hot temperatures. This extends endlessly with each question partially answered and leading to further questions. This week we actually look at the difference between cycle life gains from gains in temperature BETWEEN simple cycles of direct discharge from full to empty and the effects of temperature on cycle life on cells that are only PARTLY charged and discharged. Turns out there is a HUGE difference.
And that leads us to serendipity. While opinions don’t matter, luck can. And the advantage of the WIDER scope is that once in a great while even a blind hog gets an acorn now and again.
I meant to do a bit of a chalk talk on the CURRENT THEORY of what happens when you overcharge a cell and why that act can cause a very hot very dangerous fire. We know by observation that using what is PROBABLY the SAFEST of these cells, LiFePo4, we can actually burn a warehouse or a ferry to the ground with a single car. These fires are horrendous and my observation is that it is almost always an overcharging event, usually caused by some well intentioned battery management system, that causes this. But what is the theory of operation of HOW it does this.
But understand this is the CURRENT THEORY I can come up with by reading a LOT of “university research papers” and attempting to translate that into something humanoids COULD be expected to get their head around in one brief session. Note that there is no universal agreement on this. There are a LOT of theories. But if you take the more prevalent and cited ones, and kind of artfully put them together into a vague and broad consensus, that’s more or less what it looks like currently. It may be my presentation, but it isn’t really my opinion. It is relating mostly the work of others in as cogent a terms as I can muster.
What you CAN”T have is nonsense with no theory. You can’t have a cell with a voltage of x, apply a higher voltage of y, and then say that you are NOT charging the cell because it is “not higher by very much”. To do that, you have to show a theory of where the excess voltage went. It could go off in heat. It could go off in light. It could go off in furlongs per fortnight. But you can’t simply leave the voltage or the power applied on the table to the whims of the gods. This is the heart of the Boeing problem They have a 3.72 volt battery held at 4.00v and a resulting fire. I think it is from overcharging. That’s a pretty good theory since we demonstrably have by DESIGN a higher voltage than the cell applied for an indeterminate amount of time. I’m quite ok with that NOT being the cause but it can’t be because we didn’t do it enough over voltage. Enough to do it right now?
But I was a little diverted by a fascinating paper I came across actually by one of the students at a university. State of Health Estimation of Li-ion Batteries – Cycle Life Test Methods.
This paper, by Jens Groot is of course entirely directed at something else. He’s developing cycle testing methodologies to aid in developing some sort of long term state of health estimation algorithm. But a couple of things jumped out. The first was the dramatic difference between the cycle life of cells that are charged to 100% SOC and subsequently discharged to 11% SOC, and the alternate case of cells charged to 50% SOC and discharged to 24% SOC. The former was depleted to 80% of its initial capacity after 2000 cycles. The later was at 9000 cycles and still enjoying 85% of original capacity.
We had long talked of this, as a theory, and here was a fairly rigorously crafted test that showed it in glaring form. It’s not a few percent. It’s four and a half times greater cycle life.
And it implies that our 2000 cycle cells could easily go 10,000 or 15,000 cycles driven conservatively. It might also imply a DIFFERENT operating procedure to extend battery cycle life – don’t ever fully charge it.
Secondly in that comparison was the fact that the 2000 cycle version was done at 3.76C, while the 9000 cycle version went as high as 23C current rates. I had heard many times that high power output of a cell reduced cell life. I have never been able to muster any observation to support that, but was vaguely ok with it as it “sounded like it made sense.” The implication in this test, not totally proved but certainly there in the data, no problem with high current levels.
The other thing that jumped out was that on the SAME driving cycle A, there was a hugish difference between 23 degrees centigrade and 35 degrees centigrade. And the direction CONTRAVENED everything I’ve seen written about the effects of temperature on cycle life. All my OWN observations in testing show a battery working bettter in all respects at higher temperatures of 35 or 40 degrees centigrade, so my question was, “why then would cycle life be shorted from higher temperatures.” Something short term ought to be “worse” at 35 or 40 degrees if those temperatures are bad long term.
Well apparently not. Same for same, he got LONGER cycle life from 35 centigrade than from 23.
Why does all this matter?
Operating procedures. How should you charge your cells. How far should you discharge them. Should you cool them if they get hot? Should you heat them if they get cold? And the answers matter. If you are using cell capacity, to cool cells so they will last longer, when they will actually last longer hot, you not only lose cycle life from the lower temperature, but also from the use of energy from the cell to cool it. You get a double debit in EXACTLY the opposite result your are attempting to achieve, with additional weight and complexity in your car to do a dumb thing.
At the same time, it might make sense that the reverse situation is also true, when in fact it can be totally false. We DO know that cells too cold when charged are damaged. So the yin and yang and logical ones and zeros that are attempted in online discussions because they “make sense” just are not the basis for cell care. You can’t really thought experiment your way into “logical” results. Every operation must be related at least to a working THEORY and the results should be both observable and reproducible. The cells are what they are.
Why am I excited? Because every time we touch one of these cells we learn something new. In fact, in every case where someone ELSE touches one of these cells, we learn something new – if they follow some basic methodologies and accurately report their observations.
I was kind of eggregiously reminded this past week that some of the online poseurs unfortunately will get so wed to a position that they will actually and quite deliberately present “observations” never made nor ever possible to make to support their theory. It is true I suffer fools with exceptionally poor grace. Even quite publicly.
But beyond that, we live in an exciting time. These energy storage devices are modern day miracles and indeed appear miraculous to me BECAUSE we do not fully understand them. The process of gleaning little bits of information on them a tiny bit at a time is actually very intensely pleasurable to me personally. ANd it implies a future of vastly improved energy storage – totally silent, without emissions, with no moving parts. Not just cars, but flight, levitation, all things become plausible as we learn to store and retrieve usable energy in useful amounts at the atomic level. It is the true promise of “atomic energy.” And I get kind of jazzed up at that. When I consider what an unfathomable amount of data can now be stored on a tiny thunb drive so small I lose it IN my pocket, at a price of less than $20 the concept of energy storage on the same scale is just explosive. Literally. Cranial detonations. I can’t get my head around it.
And so just a brief peek into my ongoing and very passionate love affair with batteries. For me, they are a peephole glimpse into the very land,,, the shadow of the figertips…. of God.
Below are the rest of the slides used in today’s show which you may examine at your leisure….