Friday, July 16, 2010

4C LiFePo4 Power Tests

I was able to post this week's video ON Friday. We shot a good bit of it last week. So this week we do a 4C load test of four CALB 100AH cells.

The test actually consists of 40 pulses of 400AMPS lasting 2.5AH or about 30 seconds. The CALB cells list 4C for 30 seconds on their spec sheet. We do a full 100 AH with every single tron coming out at a 400A rate (well, until the end when it more or less collapses).









Here is the Microsoft Xcel spreadsheet chart of voltages and temperatures experienced across the test.



The good news is that the cells will indeed do 400 amps quite handily. I think the bad news is that at that rate, the Peukert effect does come into play and you will get a little less than 100 AH from the cell.

Overdischarging the cell of course is damaging. We had venting after 95AH and indeed lost a cell. But this was not entirely accidental. One of the things we wanted to demonstrate was failure under a substantial load from overdischarge. No fire. No explosion. No thermal runaway. In fact, when we did discontinue after 99.9AH, all four cells cooled fairly rapidly.

Interesting test. You'll want to watch this one.

Jack Rickard

27 comments:

  1. Any thoughts about which loads will be usefull for estimating the Peukert constant ?
    .3C, 1C and 2C ?

    Right now I have a lot of .3C curves and the setup to do up to 2C. Higher loads will require a new setup or just a suitable piece of copper ;-)

    Martin.

    ReplyDelete
  2. Obviously the higher the load, the more pronounced the Peukert effect. I've suspected it for some time, but even at 100 Amps, it is so minor you don't really know if its there. At 400 amps it's there, but it looks like 5-7%.

    So like most of the things that apply to Pb or NiMH cells, they ALSO apply to LiFePo4 cells. But they are an order of magnitude or more less pronounced, and so you don't really have to deal with them. In general, the Peukert constant, the concept of diverging balance, charge efficiency, and even to some degree temperature compensation are just not things you need to address with LiFePo4 cells. Not because they don't exist. But because they are so minute they don't matter.

    Jack Rickard

    ReplyDelete
  3. Yet another excellent show! Lithium ion sure is turning out to be a bulletproof design. Seems Toyota is willing to bet that the 18650 cells will be their battery of choice in the new RAV4EV, so that's encouraging as well. I can't wait to hear what you have in store for the 400 amp cells....

    ReplyDelete
  4. "Obviously the higher the load, the more pronounced the Peukert effect. I've suspected it for some time, but even at 100 Amps, it is so minor you don't really know if its there. At 400 amps it's there, but it looks like 5-7%."

    Ok - With my paired 90Ah cells and a maximum current draw from the controller of 270A, I will reach 1.5C and I was planning to do some test at that point.

    Martin.

    ReplyDelete
  5. Mmm, I can almost sense it in the air, a quad 9" or 11" DC 4-wheel independent drive, high-performance monster, which you hinted at during the tests. Hope it happens, this seems to be the way taken by Mercedes with their e-cell and audi with their e-tron concepts (except they are AC, but still)

    ReplyDelete
  6. Great testing-video, Jack.
    I got 120AH from SkyEnergy (now CALB), which seem to be 130AH. My Curtis 1221C will do up to 400 Amps. I'll tell you my experiences, if the car runs (hopefully soon).

    ReplyDelete
  7. Jack; You mentioned ultra capacitors... Could an array be used to store electricity for (later use) from regenerative braking even as you are showing that their seems to be no benefit in real world driving, from regenerative braking as this pertains to getting maximum mileage/ performance from the vehicle. Could this be from (parasitic losses) in the motors trying to overcome driving while braking. Doesn't sound like I'm making my point but if anyone can understand what I'm trying to say, it's you.
    I've mentioned that my 2010 Insight charges going uphill and its' battery, such as it is is full going downhill. Perhaps it's just dyslectic...
    Honda missed the boat on this little bastard, big time but I still am glad I bought it and enjoy driving it back and for to the truck stop then driving the diesel all week. Aways trying to get better mileage.

    ReplyDelete
  8. According to Jack's measurement it will not matter what you use to store the energy in (batteries, supercaps, buckets,...) as long as there is a less than 100% efficiency in getting it in or out of the storage.
    The Supercacitors might be more efficient than the lithium batteries, but you will need a dedicated subsystem in order to utilize them efficiently - and that will cost $ and weigth. When you use the batteries the normal inverter is used and it will not cost you anything in terms of $ or weight - but Jack's point is that it will dictate a less than ideal driving pattern and thereby increase the overall losses. The last point would still exist with the supercap solution.

    Martin.

    ReplyDelete
  9. Hi Jack,
    I enjoyed your show again this week. It was good to see that the CALBs can take a beating like that at 4C, although not indefinitely! I did notice a detail in the video that I had to rewind and watch several times to make sure that I saw it correctly. I noticed in the last couple of 400A cycles (starting with the cycle that went from 92.5Ah to 95Ah) that the current load indicated on your test equipment was no longer at 200A each. The lower current load "box" starting dropping off suddenly right at the end of that cycle, approximately 94.1Ah. You were busy gathering data as the sole test operator, so I wasn't sure if you caught that change. For the 95Ah to 97.5Ah cycle, the pack starts off at 400A, but quickly drops to less than 300A on the current load readout when you reached the end of the cycle. In the final cycle, the pack immediately starts dropping from 400A and ends at barely 225A by the time it gets to 99.9Ah.
    I'm not a EE expert, real or imagined, so I don't know if this information matters. I do understand from your test that by the time we're in this part of the discharge curve, we're likely doing damage to the cells, so we just shouldn't go there in practice. Do you have any additional thoughts about the importance of this data, or is it simply more evidence that the cells are self-destructing in their attempt to provide such a high current output?

    One other minor detail: Careful with your Celsius to Fahrenheit conversion. 75 degrees C is actually 167 degrees F, so 185.5 is quite a bit above the CALBs specified temperature. Of course, that was well into the dangerous part of the discharge curve, and the pack did hit 167 degrees F at the about the same point that it reached the knee of the curve, and it was still able to put out 400A with no problem at that point.

    Thanks for answering a newbie's questions. I know there are many others asking far more intelligent questions than I here on your blog. Thanks again for posting all of your hard work for us to nitpick apart! :)

    Best,
    Alex Vieira

    ReplyDelete
  10. I shed a tear for those poor cells. Such needless abuse, you couldn't have stopped at 95ah? They could have gone to good homes. Have a heart man!

    JRP3

    ReplyDelete
  11. "Great testing-video, Jack.
    I got 120AH from SkyEnergy (now CALB), which seem to be 130AH. My Curtis 1221C will do up to 400 Amps. I'll tell you my experiences, if the car runs (hopefully soon). "

    You should be within spec and with a little headroom to spare. I think you'll find they will do it quite ably.

    Jack Rickard

    ReplyDelete
  12. "Jack; You mentioned ultra capacitors... Could an array be used to store electricity for (later use) from regenerative braking even as you are showing that their seems to be no benefit in real world driving, from regenerative braking as this pertains to getting maximum mileage/ performance from the vehicle. Could this be from (parasitic losses) in the motors trying to overcome driving while braking. Doesn't sound like I'm making my point but if anyone can understand what I'm trying to say, it's you.
    I've mentioned that my 2010 Insight charges going uphill and its' battery, such as it is is full going downhill. Perhaps it's just dyslectic...
    Honda missed the boat on this little bastard, big time but I still am glad I bought it and enjoy driving it back and for to the truck stop then driving the diesel all week. Aways trying to get better mileage. "

    As you all well know, I am NOT a hybrid fan. But I've actually been looking at this fascinating little car quite a bit - with an eye to deep sixing the hybrid part and converting it to electric drive. It has some very interesting characteristics and I think you are quite correct, Honda didn't realize what they had here.

    As to capacitors. The entire capacitor theory seems to be all bound up with the concept that you might somehow overcharge your batteries with regenerative braking. I guess I don't buy into that. This is where Dan's total confusion with wH DOES matter. You have to raise the voltage quite a bit to force current into the pack and the difference between the charging voltage and float voltage on these cells is pretty substantial.

    To "overcharge" them, you have to pretty much take them off the charger and jump off a six mile cliff with regen full on.

    Dr. Nourbaksh certainly feels capacitors could increase the efficiency but he's kind of jumping through buck/boost switching hoops to get there.

    I actually did quite a bit of testing with Maxwell 3000F Ultracaps two years ago. We could drive a GEM about a block and a half just on what was held in the caps with NO batteries connected.

    My view of the caps is that they can be used very simply in a parallel configuration with the battery pack. This has TWO main advantages:

    1. Instant power delivery. You can push huge currents from teh caps for really periods quite akin to a full acceleration. So the 4C limit goes away.

    2. Battery life. Batteries are pretty conservative beasts. They don't like change. Like sudden charges and sudden discharges. When you first put the throttle to the wall, you are drawing a huge amount of current of course. This is initially supplied instantly by the capacitors. As the seconds tick by, the caps drain down. As they are in parallel with the batteries, the cells start to provide current precisely commensurate with the falling current ability of the caps.

    Once the acceleration is over, if you go into regen, the caps can instantly absorb quite a bit of energy - shutting off the energy drain from the cells. And the voltage in the caps then begins to "charge" the cells at whatever rate they will take it.

    With no regen, the battery current peaks as the caps become depleted but when you take your foot OFF the accelerator, the caps begin to charge. So the cells CONTINUE to provide a gradually diminishing current as the caps charge.

    The result in a theoretical world is that the pulses from the pack are converted to much more gradual sawtooth waveforms. In actual practice, the cells and caps interact in a more closely coupled fashion than this. BUt the effect is that the caps BUFFER the cells dramatically.

    There was a guy in Australia years ago that got over 100,000 miles from a LEAD ACID battery pack this way. Talk about life extension.

    ReplyDelete
  13. The PROBLEM with Ultracaps. They take up space. Anmd they add weight. In fact, they take up about the same space, and weigh about the same, as cells of about that capacity.

    So, in any case where I could use Ultracaps, I can get more RANGE by adding batteries of the same size and weight. Unfortunately, while the Ultracaps give me INSTANT POWER and GREAT LIFE CYCLE EXTENSION, they do NOTHING to increase range.

    And they are quite expensive. I just bought ten 16.2v units made up of six 3000F caps - about $7500 delivered.

    So you wind up needing more space and more weight, at even more expense, and it goes no further than it did.

    That's the cap trade off, and why almost no one does it.

    The project we're going after is a bit larger physically than we've done, and the weight and space constraints not quite as severe. But moving all that will take a lot of power, and so on balance, to protect a very large and expensive cell pack, and be able to produce that level of power, I think the Ultracaps will finally make sense more or less and I've always wanted to work them into a project.

    Jack Rickard

    ReplyDelete
  14. "According to Jack's measurement it will not matter what you use to store the energy in (batteries, supercaps, buckets,...) as long as there is a less than 100% efficiency in getting it in or out of the storage.
    The Supercacitors might be more efficient than the lithium batteries, but you will need a dedicated subsystem in order to utilize them efficiently - and that will cost $ and weigth. When you use the batteries the normal inverter is used and it will not cost you anything in terms of $ or weight - but Jack's point is that it will dictate a less than ideal driving pattern and thereby increase the overall losses. The last point would still exist with the supercap solution."

    We would be using supercaps WITHOUT a convoluted boost/buck system, simply as a buffer. But we would do this for power delivery and the advantages of buffering the cell pack. While there might be some slight advantage to the regen, we'll never know. The project will not feature regenerative braking. We're going back to a series DC system for this one.

    Jack Rickard

    ReplyDelete
  15. "We would be using supercaps WITHOUT a convoluted boost/buck system, simply as a buffer. But we would do this for power delivery and the advantages of buffering the cell pack. While there might be some slight advantage to the regen, we'll never know. The project will not feature regenerative braking. We're going back to a series DC system for this one."

    in that case you will not use the energy stored in the capacitors - you can more or less use ordinary capacitors for that task. You will not get the same amount of F, but the low ESR and current delivering capabilities will be there - imho.
    Another alternative could be to use Lithium Ion capacitors - a supercap with a base potential. This would reduce the number of capacitors you would need for the string voltage. The package is not the best for your purpose since the ones I have seen is in the pouch form and will need quite a bit of packaging.

    Martin.

    ReplyDelete
  16. Mr. Rickard,

    I keep my fingers crossed that your next show will talk about capacitors. I was reading somewhere where someone succesfully combined Lithium and lead/acid together in a pack. My first thought was NO WAY. You just can't combine the two because the internal resistance of the batteries are is so different. Then it occured to me that mixing capacitors with lithium is similar to mixing lithium with lead/acid, right? Maybe I'll put Prius cells along with some 18650 cells together with my lead/acid pack and see which one explodes first...:P

    ReplyDelete
  17. Yes, Martin, we'll get great energy out of the capacitors just as they are - whenever we call for more than the cells want to deliver. We've done this before. The performance increase is phenomenal.

    I have also fairly successfully run lead acid and lithium cell strings in parallel. There's not much use for such a beast, but I've done it with no harmful effects.

    Jack Rickard

    ReplyDelete
  18. "I shed a tear for those poor cells. Such needless abuse, you couldn't have stopped at 95ah? They could have gone to good homes. Have a heart man!"

    We're forever destined to different worlds JRP. I do understand you want to preserve, them, that's why I take such joy in destroying them. If you hook up a cell, say some magic words, incorporate some magic dust BMS and sacrifice a pig in full moon conditions, and they work, WHICH thing that you did was important?

    If you can get them to melt in front of you, you usually know why and where.

    BTW, I had a very interesting terminal failure in the Mini this week we'll talk about - a melted terminal bolt and a welded strap to the cell. Replaced it with one of the four you saw me stressing at 400C. Drove it 96 miles today requring 81AH from the 100AH pack, and it is back on the charger at the moment. But it looks like it is making good capacity.

    I think torturing cells is kind of a mission we've taken on - so you don't have to.
    Jack RIckard

    ReplyDelete
  19. "Yes, Martin, we'll get great energy out of the capacitors just as they are - whenever we call for more than the cells want to deliver. We've done this before. The performance increase is phenomenal."

    In which way do you see (measure) or feel this increase?

    The energy from a capacitor is
    0.5 * C * V^2 and delta energy is
    0.5 * C * (V1^2 - V2^2)
    A 144V string will require at least 54 Maxwell EDLC's and result in a C of 55.6F.
    If we assume a voltage drop of 10V it will result in 77kJ or 21Wh

    The resulting power might be a different point.

    What about cell balancing? Supercaps WILL be destroyed if you increase the voltage above the design limit (2.7V for the Maxwell K21 series). And after repeated charge/discharge cycles and even no load conditions you will get different voltages across the caps.

    Martin.

    ReplyDelete
  20. "If you can get them to melt in front of you, you usually know why and where."

    Jack and Brian,
    Love your show. Learned a lot about testing this way when I was a kid. :)

    Right now I have a mini that I bought with a blown engine. (garaged and very low miles,sweet) I'm thinking about the ac-5o. Still learning how to spec an electric motor. Think it'll be happy in the mini? Maybe a NetGain DC?

    Curt Nuzum

    ReplyDelete
  21. The key advantage of using capacitors in EV is their ability to source or absorb a huge amount of current in a small amount of time, which is a direct result of their very low internal resistance. Since acceleration (sourcing current) is really not much of an issue with a good bank of batteries in a normal commuting application, the capacitors biggest benefit is to absorb more of the regen energy than the battery alone can do (translated efficiency). Flywheels have a similar advantage. Capacitors have a higher internal discharge rate than batteries, so it is best to use this energy first or to transfer it into the batteries for later use. Using them as "battery helpers" would seem to have significant benefit potential.

    Dan

    ReplyDelete
  22. We actually sag quite a bit in voltage when producing a large amount of current suddenly. I experimented with this several years ago with the little GEM vehicles - a 72 volt system that we ran in parallel with a bank of 36 Maxwell Ultracaps. In fact, we had a 72volt system of Seiden LiFePo4 cells, a full set of Trojan 135AH cells, and the Maxwells and I could run the vehicle on any, all, or any combination of these three by turning some mechanical heavy duty switches.

    The results were very interesting. For example, I could parallel the batteries and the caps and charge them up, and then DISCONNECT the batteries. The vehicle would run about a block and a half just on what was in the caps.

    But the one that haunts me is paralleling them. The throttle response was instant and would tear the tires off the buggy.

    With no buck/boost circuitry, the things still work very well with the batteries. The caps can put out instant current and the batteries can ramp up to whatever is necessary. The result is that rather than sagging when you first step on the accelerator, you accelerate. Over time, the sag eventually comes back into play - long periods of high current. But it really boosts acceleration.

    The GEM's did feature regen and yes, it did have some salutory effect on recovery, not as much as you would think, but some.


    A guy in Australia some time back racked up over 100,000 miles on a set of lead acid cells, or at least claimed to, using this technique. That's actually pretty incredible. The only explanation would be that the buffering action of the caps extends the life of the cells. And that's what I would expect here.

    Instant acceleration from the caps, with a kind of sawtooth waveform out of the battery. The caps never do quite discharge at all. They just supply the initial current and the cells can replace that current at the rate of cap discharge. Similarly, as current needs reduce, energy will flow into the caps at a reducing rate as they fill. The batteries see more of a sloped demand and release rather than the sudden pulses. The voltage more or less stays the same.

    Jack Rickard

    ReplyDelete
  23. Jack,

    Excellent video! It appears that cell failure could have been averted by having a low voltage alarm on each cell. Based on your graph 2.0 volts looks like a good number for a controller with a 400 amp limit.

    I am guessing that the 4 cells are not perfectly identical and I wouldn't be surprised if one of your cells actually reached zero volts (or reversed). This ultimately this is what damaged the cell. It is a shame you didn't measure each cell voltage as you brought your pack to its knees.

    Sure, setting a low battery alarm at 2.0 volts would cause an alarm at about 85% discharge at max throttle, but maybe it is a good idea to have a little reserve before you completely drain the pack! You could then think about using the remaining energy more conservatively as you drive to someplace to charge your battery. Similar to having a low fuel light in a gas powered car.

    Thanks for your sacrifice in time, effort and battery.

    Respectfully,

    Joe

    ReplyDelete
  24. Jack,

    I thought the quote below to be relevant to our little discussion on regen. If nothing else, its an interesting statement!

    We will be at KOSH Monday and Tuesday. Hope to meet you.

    Dan

    "I want to establish in your mind very clearly that you must not think I deny all that I do not admit. On the contrary, I think there are many things which may be true, and which I shall receive as such hereafter, though I do not as yet receive them; but that is not because there is any proof to the contrary, but that the proof in the affirmative is not yet sufficient for me"
    - Michael Faraday

    ReplyDelete
  25. Joe,
    Just because the cell sags to 2.0 or lower under load doesn't mean it's empty, or that it will be damaged. That's why a 2.0 alarm isn't that useful. I've pulled my pack to 1.77 average cell voltage under load with no noticeable problems. I wouldn't do it on a regular basis but it's certainly not the end of the world. Bottom balance your pack and you won't be pulling any one cell significantly lower than the rest.

    JRP3

    ReplyDelete
  26. Jack,
    Any chance you will be running some tests on Headway cylindrical cells? They have a new 16AH cell, which does 5C continuous and 10C for 30 seconds. I know you love the prismatics, but for us motorcycle guys, they are tough to fit.
    -Zoom

    ReplyDelete
  27. They are a little off our track, but yeah, I ordered six of them today.

    Jack Rickard

    ReplyDelete