I ran across a very interesting video courtesy of viewer Brandon Hollinger http://ampREVOLT.com/. Brandon has done a lead acid conversion of a 1970 Saab 96 that looks quite well done for a Pb car. And he’s started a LiFePo4 Miata project for someone else.
Jay Whitacre received his Phd in Materials Science in 1999 at University of Michigan Ann Arbor – so he’s a lot smarter than I am. He worked for Jet Propulsion Laboratory 2000-2007 mostly on lithium battery projects. Today, he is an assistant professor in Materials Science at Carnegie Mellon university. He’s also head of Whitacre Research Group which studies functional materials for electrochemical storage.
The group received a $5 million DOE grant for the development of a Sodium battery using a Sodium Metal Oxide cathode (Na4Nm9O18) with an aqueous electrolyte. These cells are 1.6 volts or thereabouts and a scant 45 Wh per kilogram. But they are totally benign, have a cycle life well over 5000 cycles, and cost about as much as lead acid cells to produce. Because of the low energy density, they are probably not suitable for electric vehicles. But for grid storaage, they are ideal. Inexpensive, rugged, very long lifetimes, environmentally totally benign, easily scalable, and did I mention inexpensive. Like buying lead acid cells that last 20 years and have no lead. The spinoff company is Aquion Energy.
They apparently do not have a web site up but HERE is a bit more about their product development.
In this video, Professor Whitacre addresses a group of EV guys to explain the chemical processes of Thundersky LiFePo4 cells.
The presentation is very interesting and answers a number of temperature questions quite well I thought. Amazingly, at about 1:05 into the presentation he’s taking questions regarding Battery Management Systems. And he basically tells this group, who are all totally convinced you have to have a BMS, that essentially NO manufacturer, including Tesla, GM, and Toyota, use a cell level BMS and in fact are moving away from the module management they have now as just too expensive and apparently unnecessary. He stresses EXACTLY what I’ve been saying, closely matched capacities and string level management.
I found it fascinating. So I thought we’d store the video on our Amazon server and create a blog entry for reference.
Enjoy
Jack Rickard
Jack,
Good to hear, you’re not alone in your opinion regarding BMS. A man needs that so now and then. You’ve got me convinced without the right Phd.
But a totally different question: What have you done with your broken MES-DEA TIM600? Do you want to get rid of it?
Regards, Jan
IT is being repaired by one of our viewers. It had a blown IGBT.
Jack RIckard
Hi Jack,
Thanks again for your latest video and this one too. Have you heard of CALIBpower based in California? They seem to be the official distributor of CALB here in the US. Do you order your CALB batteries form them?
Just one more note: I really don’t want to point out mistakes but(maybe its due the wine :))… anyway, to refresh on AC theory: you were explaining to Brian about AC 120V being 60V on the positive half cycle and 60V on the negative half cycle. The 120VAC from the outlet is 120Vrms, so the peak is 168V peak positive half cycle and -168V peak negative half cycle, etc. Also, pulsating DC is still DC and not AC. AC is zero crossing and have opposite & identical half cycles.
JohnM
A really interesting video – thank you for posting the link. Interestingly he didn’t touch on self-discharge rates which seem to be a potential source of drift in cell equalisation
Self discharge doesn’t seem to be much of an issue with these cells.
JRP3
Self discharge has nothing to do with the entire mythical concept of “drift” in cell equalization. This is a hallucination of amateur BMS designers.
I can’t find evidence of self discharge itself.
Jack Rickard
At the very end of this video(~74:20) he says that a differece in about 1-2% will magnify over the cycles. That would suggest that the cells WILL GROW APART. Or did I interpret that wrong ?
I find it very comforting to hear that he likes to do a bottom balance and have well matched cells as the most cost effective approach for getting a long lived and trouble free pack.
As from getting well matched cell that will be the challange. It would be good if some third party would actually match cells from these big manufacturers, like TS and CALB. Or if you could pay a little extra to get 1-2% variance only.
Best Regards
/Per
You did Per. These beliefs, once you hold them, are hard to let go of. And everything tends to confirm them.
It is possible that the capacity differences grow very slowly. The rich get richer and the poor get poorer. That does not mean that they “grow apart”. The best treatment, as he very clearly said with his entire audience virtually begging him over and over “but don’t we need a BMS” the answer was clear – no treat them as a string. A cell level BMS is an absurdly complicated and expensive proposition.
There is an ONGOING confusion between capacity and state of charge. It just won’t let go. There is NOTHING you can do about capacity variations except more efficient manufacture or selection. And no matter HOW much better you do that, there will REMAIN capacity differences.
But you CAN keep the cells at essentially the same state of charge.
Now if YOU redefine SOC as having the same voltage at the end of the charge process. and simply refuse for it to be anything else, you are never going to get there.
If you imagined, for the sake of discussion, that ALL your cells were at exactly 50% SOC. And you simply ignored the end voltages at the end of charge, and at the end of discharge. ANd you did NOT do this goofy top balance thing ever, the cells would operate just fine essentially forever.
Over time, the difference in voltage at the end of charge would grow – quite slowly, but it very possibly would. And the differences in voltage at 100% discharge, would ALSO grow, very slowly over many cycles, but it did. And you neither discharged 100%, nor charged 100%, why would you care? You’re never going to get, and never WERE going to get, any more capacity out of the pack than the LEAST capacity cell in any event ANYWAY.
And over 2000 cycles, that total capacity is going to decrease to 80% of the original capacity ANYWAY.
Now what is “growing apart”????? The capacity difference from one cell to the next?
And it is ENTIRELY possible, that at the end of 2000 cycles, at 50% DOD, that ALL the cells are more or less 50% DOD…….
There are two things I have no doubt of. After all that, they will NOT be all at exactly 50% DOD. And second, the diffferences will be such that it won’t matter at all.
There is a huge difference between “doesn’t happen” and “doesnt matter”. ANd it mostly lies in degree….
Jack Rickard
Cycle life seems very uninteresting, to me at least.
Take a pack at 25 kWh, roughly 160 km worth of energy. The 2000 cycles means 320000 km driven – few people can do that distance in just one car’s life.
In the Netherlands that 25 kWh pack would break even with gasoline at 575 cycles or roughly 57200 miles.
I’m using 150 Wh/kilometer as a “mileage” here.
The figures I hear mentioned on EVTV’s convertables indicate a better mileage then that.
Could you provide us with some mileage figures on your EVs, Mr Rickard?
@Jack Rickard
What do you think of active battery management systems that try and keep either state of charge or voltage level across all cells during use (both charge and discharge)?
Alex,
Are you serious? Do you watch his videos and read his posts?
Jack,
When he was talking about the positive effects of pulsing I thought that might be another plus in favor of regen, a short reversal of the discharge process. Thoughts?
JRP3
Jack thanks for posting the video! He seemed very knowledgeable on the subject. I would love to hear his opinions on your bottom balancing video as he stated it didn’t make a difference where you balanced at. I realize you have come off of the bottom balance idea somewhat, since they come from the factory closely matched as of SOC. I would love to see him as a guest on your show and have a battery discussion.
Ditto JP on comment about Alex LOL!
W5LEE:
I haven’t come off of bottom balance AT ALL. In fact, I’ve never advocated it. I’ve been enormously clear on this all along. NO BALANCING. You use them out of the box. We do bottom balancing as a remedial measure when cells have to be replaced, or when configuring a pack like the Vantage van where some cells were used, and others not.
IT is BETTER than top balancing, simply because that is where a 500 amp cell reversal event would happen. Slight overcharging at 15 amps is a NON event. But we have never said you should bottom balance your pack. Not ever.
If you HAVE top balanced. Bottom balancing is about all that is left as a solution. Actually, the REAL solution is to take each cell, charge it fully to spec, discharge it fully to spec, and then charge to precisely half AH – 50% SOC. As a practical matter, it is nearly impossible to perform in a month of all day effort. So bottom balancing is all that’s left.
Jack Rickard
Khalen:
We think cycle life is actually the central issue with the cells. IF they last the life of the car, the expense becomes part of the capital cost of the vehicle, rather than a recurring expense. At the price of these cells, it is important that they last for many years.
In practice, we’re removing the sells from the original speedster largely because they are obsolete. We can do better with more current cells, and ifact I’m kind of waiting on the newer cells about to be announced.
As to mileage, it sounds like we are more or less on the same page. 150wH per kilometer is 240 wH per mile. That’s right in the ball park with the two Speedster’s and the Spyder = maybe 225 wH per mile driven carefully and 250 at reckless abandon.
The Mini Cooper is a much heavier car at 3500 lbs and we see 350-375 wH per mile. What this means is broadly speaking you can calculate wH per mile as a function of weight – 0.11 wH per mile per pound. So a 2000 lb car runs 220 wH per mile while a 3500 lb car would be 385 wH per mile. Ergo a 7200 lb vehilcle should come in at about 792 wH per mile. And so on. It’s a rough rule or thumb, but it seems to work more or less within the variance that occurs in driving style anyway.
1.609 km per mile x 150 = 240 wH per mile.
Jack Rickard
Well Jack. I think you have rightly put to bed this business of BMS.
Again.
A brilliant film. Information packed in every moment.
If you see that Jay Whitacre; give him a tea bag and tell him the drinks on me. 🙂
Pulse charging.. It *could* be used to speed up charging speed in theory because RC across the battery poles make a high pass filter. Pulsing a charge also makes the charge travel outside the skin better. So *guessing* push/pulling the ions across the poles more efficiently. There is no way I would exceed the battery’s electrolyte limits at any moment in pulse time. I’d play chicken until we know better and stick with say, 3.7V max per cell.
Your “wH per mile as a function of weight – 0.11 wH per mile per pound.”. Is what I want to hear. Loving it! Numbers like this give important baselines for any future project. I’m looking to an XKR Jag but still attracted to a “mommy wagon” for good battery storage and practicality.
Jack, I think your new AH meter covers most of what anybody needs instead of an OBD-ev-car-pc. Do you know of LinuxICE? Its pure multi “everything you want” bling which has all the toys for ev’s.
Great video, Jack. Thanks for posting it. I could have missed it but I didn’t hear him specifically say anything about charging a cell when its temperature is below 0C. It was clear there wasn’t any problem discharging with it in the -20C to -10C range.
If you go to 1:09:22 and listen to what he says he clearly said that the engineering way to do it is to put in cell level management. He further says that for what the group is doing they probably should monitor each cell. I think the biggest issue right now is that so many have had issues with other battery chemistries when there was no BMS that the assumption is that every chemistry has to have it. As Jack has said many times, poorly designed BMS systems are worse than no BMS at all. Even if you believe you have to top balance it is silly to go with some of the circuits which bypass 2A of current. That is just a heat induced disaster waiting to happen.
On my pack I have what I believe (hope) are the more safe BMS boards on the market. They can top balance but I don’t use that function by not charging up to 4.00Vpc and if they do get to 4.00vpc they will only shunt 500mA. Assuming the boards are “perfect” and never fail (my BMS boards have no adjustable items and is fully conformal coated), was it a good investment? My particular system cost me $40/board with brass bolts which have a tapped hole in the center for mounting the boards independent of the connecting straps. (Incidentally the brass bolts haven’t loosened in 5000 miles and I only have a bronze split washer on top of a brass flat washer for holding them.) Not counting sales tax, my 40 TS-LFP100AHA cells cost me $4840 or $121/cell. I spent $800 for the BMS. I could have bought 6 more cells for less than $800. Since I’m running buddy pairs that is three cell pairs worth. This means that in my 2p20s pack I could handle killing three sets of cells and still have spent less money than I did with the BMS system. Even with Jay Whitacre’s comment at 56:00 I don’t recall Jack having a burned down GEM from reversing a few cells so the probability of burning a car down from over discharge appears to be minimal. It is clear that the cost benefit of not using a cell level benefit for the OEMs is worth it. It appears that unless one is going to be pushing the limits of the charge/discharge curve of your LiFePO4 batteries that the cost benefit to the DIY is to go the same route.
Jack, you mentioned the “mythical concept of “drift” in cell equalization.” Is there evidence of this either way? I don’t remember seeing much. There are your hand full of EVs and my 4 months of non-balancing charging on my cells to support this but is there any other data you know of to support this. Maybe it is my skepticism of the belief that the cells will behave similarly to each other during their entire lift span that I am asking. What is the worst that can happen if, for example, a fully charged cell gets an internal short? Will it catch fire and is the risk of that happening lower than the risk of a properly designed cell monitoring causing an issue?
David
I don’t know if everybody sees the confusion about the M of BMS. The professor in this video reads the M as it is intended: Management. While a lot of us, read the word (subconsciously) as Monitoring.
A real obtainable BMS can do both or just one of them or half or whatever. It has to monitor the voltage, sometimes it gives that information back to the user, and sometimes it uses this information directly without feedback to act and starts resistors or shunts or whatever.
The professor clearly states that an acting BMS is not an option. He talks about taking batteries out of the pack as soon they reach a certain voltage. That is not smart, according to him. Too expensive and in the long run completely useless.
But on the other hand he also says that monitoring is a good thing, at least not a bad thing. But with the nuance he tells earlier in the video about what can go wrong with adding wires and sensors to a pack.
That is pretty much in accordance with what I think Jack’s opinion is.
It’s just like in the real world with managers. Sometimes they do nothing, just monitor and report to their boss. Some can ruin a good company by just ‘monitoring’. While others do that by acting without reporting it. Anyway the company will go down whatever management you choose, if the workers don’t corporate.
Jan
I don’t know if everybody sees the confusion about the M of BMS. The professor in this video reads the M as it is intended: Management. While a lot of us, read the word (subconsciously) as Monitoring.
A real obtainable BMS can do both or just one of them or half or whatever. It has to monitor the voltage, sometimes it gives that information back to the user, and sometimes it uses this information directly without feedback to act and starts resistors or shunts or whatever.
The professor clearly states that an acting BMS is not an option. He talks about taking batteries out of the pack as soon they reach a certain voltage. That is not smart, according to him. Too expensive and in the long run completely useless.
But on the other hand he also says that monitoring is a good thing. But with the nuance he tells earlier in the video about what can go wrong with adding wires and sensors to a pack.
That is pretty much in accordance with what I think Jack’s opinion is.
It’s just like in the real world with managers. Sometimes they do nothing, just monitor and report to their boss. Some can ruin a good company by just monitoring. While others do that by acting without reporting it. Anyway the company will go down whatever management you choose, if the workers don’t corporate.
Jan
Jack, you’re famous! (Well, even more famous!) You and bottom balancing are mentioned, and critiqued starting on this page:
http://books.google.com/books?id=o-QpFOR0PTcC&pg=PA74
Battery Management Systems for Large Lithium Battery Packs
By Davide Andrea
Jack I discharged some cells in series right out of the box. They started within +- 0.01 V of each other. On the first discharge they were about 1/2 volt out of balance at the end! I was using a load tester. Anyway, I then bottom balanced them, and they have stayed within +- 0.02 V of each other since then, despite high current discharges and charges. Anyway, this is a data point for needing to start with a bottom balance.
Haha, the link with the book on BMS was hilarious! 😀 I am in no way trying to defend jack here, he has proven that he is more than capable of dealing with fanatical idiots, but this guy just broke the moronity-meter (it’s a unit that was recently invented). I mean, how dumb must you be to consider the voltage of a fully charged cell as 4.2V and get the energy by doing something like 4.2V*7Ah??? This clearly shows that the guy in the book has never seen a battery and has no clue that the voltage stays 4.2V for just the first few percent of discharge, until it rapidly reaches 3.7V-3.8V and starts slowly decreasing from there. Not to mention his blabbering on what destroys a cell and what makes it explode (he even mentions it 3 times in a page). If he had any clue he would know this only applies to lithium-cobalt batteries, as pointed in the video, the electrolyte is the reason for explosions, and in LiFePO4 the voltage will never reach the critical 4.5V, when you are charging with the recommended 3.6V for example. A guy who has worked in JPL on batteries for the mars rovers, and has a PhD in material science, says that LiFePO4 batteries will bear overcharge and overdischarge like nothing, and some clueless “writer” says they will explode… Hm, i wonder who would I trust?
You DO know who the author of “the book” is I assume. David Andrea is the designer of the ELITHION BMS system – the same BMS system used on LincVolt in the recent warehouse fire. NOT his first fire I’m afraid. Andrea has personally burned more elecric cars to the ground than anyone I know. Vice President at A123 personally told Brian that they would never sell individual cells to the EV hobbyist community other than in modules THEY designed. He cited David Andrea by name.
Andrea worked for a company in Boulder Colorado that did a battery pack for a car using A123 cells The car burst into flames while driving down the highway. The driver managed to pull the car over and escape the vehicle. It burned to the ground – literally burning the tires off the car.
The principal at A123 believes this is one of the reasons they lost the contract for the Chevy Volt.
Jack RIckard
Hi Jack,
Are you no longer allowing your videos to be downloaded??
Cheers
Of course they were “out of balance” 1/2 volt at the end. They are SUPPOSED to be.
I am simply failing to communicate this effectively. IF you received a pack of well selected and matched cells from the factory, they will ALL be at very nearly the same SOC – 50%. But they will still vary in capacity slightly.
This exhibits at both the TOP (last moments of a full charge) and the bottom (over 90% discharged). If you do NOTHING the differences in capacity are roughly split between top and bottom.
IF you bottom balance, you have brought all to the same voltage at the bottom, but EXACERBATED the capacity differences at the TOP. No real harm, but you would have been able to charge a tenth or so higher without it.
Of course, if you balance at the TOP of the charge, you simply move all the differences to the BOTTOM (end of discharge). As long as you never go past 80% DOD, no problem. But the first time you go over 90%, the least capacity cell will go over the discharge curve first and the others will drive it into cell reversal and destruction.
It is actually best to do NOTHING to new cells. But if you have to balance, for remedial reasons, do so at the bottom. TOP balancing routinely is simply the religion of those who absolutely know nothing about these cells and it is indeed a dangerous operation – particularly unattended.
Bottom balancing is not at all dangerous. For one thing, you’re there doing it manually. That is you are PRESENT to take action if there WERE some kind of a problem. There really can’t be. But if there were, you would be there.
These guys are top balancing not as a maintenance procedure, but as part of their normal unattended charging procedure. And they are causing fires – the latest Neil Young’s LincVolt, at 2:55 AM of course, using David Andrea’s ELITHION BMS and a Manzanita PFC charger. Those are just the facts.
As enough people start using these cells and learning for themselves, I’m going to win on this one guys. I’m not guessing. It’s not a theory. And it’s not an “opinion”.
Here, I offered a video of a guy saying explicitly the same thing, and I’ve actually got people HEARING from it something different than he said, because of what they WANT him to have said.
I give up.
Jack
OF course our videos can be downloaded. Click the title of the videos and they download in the background.
Jack Rickard
Regarding video downloads: On my Mac, I figured out that I must click the yellow video title while holding down the “alt/option” key.
YOU should write a book Jack. Then you can name and shame David Andrea and charge people 130 bucks to read it too!
Excellent article on the charger – I am sure that what happened to Tim could easily catch people out. I certainly never would have expected the voltage to rise having once set it. I hope this is made clear in the Manzanita manual?
I have recently come across a great candidate for a roller coversion from a company in australia who makes a car based on the 2005 shelby cobra concept. Would love to “11 inch” that beast… http://www.gt40australia.com.au
They have a old school cobra and a VERY nice GT40 repica too.
Re self discharge.
You guys may be right – maybe it isn’t a problem in practice. It wouldn’t be if all cells in a pack were close on self-discharge rate, but the CALB spec sheets show a self discharge of “=< 3%" per month, so I was curious. Jack – what is your experience of taking old cells out of the packaging after a couple of years?
You do not need to hold any keys on a mac. Just click on a title and it will start downloading.
We have recently pulled some Thundersky TS90AH out of crumbling cardboard boxes received in June of 2008. We cannot detect self discharge with test instruments. All cells were about 3.28v if I recall – about 50%SOC and all were identical. We didn’t really do any capacity tests on them, but I may do that just to see if there are any surprises.
Yes, the manufacturers claim =< 3% per month. I would say way <. Jack Rickard
Sorry for cross talk on your blog there Jack.
To Nick Smith!
Your link didn’t work. Don’t you mean these sites?
http://www.drbsportscars.com/
The new owners or
http://www.gt40australia.com/ — Same site.
Jack in my case my cells were not delivered in “middle balanced” condition. 1st discharge had them at 1/2 Volt out of balance. I then bottom balanced them. They stay within +-0.02V at the top and at the bottom, when fully charged and after delivering full capacity, outside the flat portion of the curve and at the knees at both ends. My experience proves my cells were not delivered middle balanced — I wouldn’t count on that.
For those people struggling to download Jacks Videos, you need to select “see and try our New HTML5 page” first, then select “archive of past weekly Friday shows”
Selecting the “archive of past weekly Friday shows” from the main page of EVTV.me will only take you through to streaming video’s…
Cheers
Simon
Tesla recently anounced they were able to reduce the size of there battery pack….coincidence?
Actually your cells WERE delivered in balanced condition. They were at identical 50% SOC. On first DISCHARGE you take them to the bottom, and it is not at all unusual for them to have a 1/2 volt difference. This is simply a voltage indication of their varying capacity. If you fully charged them, at the end of the charge while STILL CHARGING, you would notice the same thing, only the batteries that were LOW on the discharge side would now be the batteries that are HIGH on the charge side. They have less capacity.
By bottom balancing, yes, you make the VOLTAGE appear balanced at the bottom, but now you’ve moved the variations in capacity to the TOP. That’s much better than top balancing, but not as good as having left them alone.
If you had fully charged the cells initially. ANd removed the charger. After about 90 minutes, all cells would show about 3.35 or 3.38 volts. And they would stay very nicely together until they reached the bottom. The DIFFERENCES in capacity were arranged roughly 50% top and 50% bottom. Now the differences in capacity exhibit 100% at the top.
As I say, that’s better than top balancing. But entirely unnecessary.
Jack Rickard
Re self discharge
Thanks for the response Jack. An ounce of data is worth a tonne of urban myths.
“Yes, the manufacturers claim =< 3% per month. I would say way <" I wonder if it varies with SOC? If self discharge is higher at higher SOC then this would account for the very mild self-balancing effect that Jack has alluded to a couple of times in the videos.
Tesla reducing battery size. Probably coincidence. They have a very good relationship with Panasonic and I would guess they are getting higher capacity cells in the same physical size. So to reduce costs they can reduce cell count while maintaining the same range.
I would have rather seen them extend the range. But realistically, I guess 200 miles is good enough for most of their customers and compares well with Leaf’s 100.
Jack Rickard
Not to mention it takes weight out of the car which should improve performance and actually increase the range slightly.
JRP3
Jack, did you manage to make your window transparent around your clocks?
I might of given you the wrong info earlier.
If your clocks are a drawn background and not an object in code. Did you save them in .svg, .png or .gif formats. Which allow transparency?
On the Tesla. If it has a smaller pack with same rating. Its greatest gain will be in saving weight. Cornering is what sports cars are supposed to be best at. Top gear in the UK found weight to be its major drawback around the track.
Does anyone have experience with Thundersky cells that have yttrium instead of phosphate? They say its an improvement but they have been on the market a fair while now.
Andy, they still have phosphate, just added yttrium, LiFeYPo4.
JRP3
The “clocks” are transparent PNG files. The problem I had was making the NSView frame background color different and I still don’t know how to do that. However, I have made it an Apple HUD panel instead which gives it a dark background that is semi transparent. By using a black desktop, I can get a pretty good effect.
But I would like to know how to change the color of the view background. It is done in Interface Builder, so I really don’t even know how to address that particular view programmatically.
Jack Rickard
…just a quick jump back to complete the thought on skipping top or bottom balance and going with the 50% SOC out of the box.
my question is… ok, so how do I stop my charger when the cell with the least capacity is full and voltage starts spiking, and the rest of the string may not have started up enough for the charger to notice yet?
i.e. my cheapo charger sees only the end voltage of the whole pack. so without any BMS, and without top balancing, it seems to me the risk is that if any cell has a significantly lower capacity, it will rise first, and cook before the total pack voltage goes up enough to stop the charger?
Dan,
The point (at least my understanding of it), is that this simply won’t happen if you follow a few simple guidelines. Take the batteries, preferably from the same batch so they aren’t really different, out of the box and charge them as a string. They will only vary by a few percent. Don’t go to the absolute tippy top of the charge curve, nor to the absolute bitty bottom. Just don’t do it. There is no reason to, because there is so little energy storage gain by doing so anyway. Yes your range will suffer by 1-5%. Who cares. Stay a few tenths of a volt away from the top and bottom of the curve. Then, NONE OF THE BATTERIES WILL BE SIGNIFICANTLY DIFFERENT ENOUGH TO OVERCHARGE OR UNDERCHARGE enough to cause any problem. Ever. Period. Full Stop.
Sorry about the yelling, but it’s a little frustrating. I mean on side says: follow my advice and you won’t overcharge or undercharge. And people keep coming back with, but what happens when it overcharges or undercharges? Which misses the point. The point is that it doesn’t do that. Of course you need a proper high voltage cutoff on your charger and maybe an amp hours used cutoff on your controller or daughter mode switch. But none of that is a BMS. A BMS is unnecessary and dangerous.
Final, point: If one or two or three batteries do in fact undercharge or over charge because their capacity is so much drastically lower than all the other cells, then it is a bad cell to begin with, and nothing will ever fix it, so get rid of it.
Jack has kindly demonstrated time and time again how much effort it actually takes to purposely kill a single cell that hasn’t been top balanced or bottom balanced out of whack with its neighbors. It’s a ton of work, a cell that is just a few percentage points different in capacity is simply not going to blow up or be permanently damaged in any way by following Jack’s method.
I, anonymous, share Jack’s frustration with the BMS crowd. Sure in theory the BMS is great and all, but I personally know of 20 DIY’ers who take their new cells, hook up a crappy BMS, destroy their cells because the BMS didn’t work like they expected, and then use this as further evidence that a BMS is necessary. Alternatively, I see people routinely try to avoid a BMS and go about pain-stakingly top balancing their cells to within 0.001 (seriously 0.001) volts. Then, they destroy a few cells this way and use this as further evidence that a BMS is necessary. That’s just stupid. Stop it.
Thank you thank you thank you. Yes, what he said. What I said.
The BMS thing is just enormously circular, a self fulfilling idiocy. And it is very powerful. There are a dozen voices feeding it for sales dollars on a high margin easy to build device. And quite understandably a vulnerable audience with a very legitimate concern for a very large investment in batteries. I am enormously sympathetic.
But like anonymous, I have reached a level of frustration that is difficult to deal with. I was one of the top technical writers in the country. I have recast this at least four times trying to find a way to communicate it. And I think I’m losing here. If I cannot communicate it, it is MY failing, but doubly the more frustrating thereby.
Some of you guys are just wearing the spots off these cells. I mean just wearing them OUT screwing around with them. That’s my job.
1. Take em out of the box.
2. Put em in the car.
3. Get a GOOD charger – I know, there aren’t any.
4. Focus on instrumentation to get the maximum range while avoiding overdischarge.
5. Undercharge slightly. This isn’t to avoid a BMS. It just extends the life of the cells in and of itself.
6. Recharge at every opportunity.
7. Be aware of parasitic loads – any little draw from instrumentation on SOME cells but not others causes problems. 20 ma is way too much because it is cummulative over time.
8. Be aware of ground leaks. ANY connection, including through a voltmeter or ammeter or anything else that causes a leak from your pack to the vehicle frame ground is quite dangerous.
9. Terminal connections. Clean. And tight. And clean.
10. If you just HAVE to do SOMETHING to protect your cells, go back to item 9.
Jack Rickard
Regarding your advice:
Point 9. Terminal connections. A couple of shows ago you showed us how to use the Nord-Lock washers. Unfortunately YOU did not read the manual on those washers. Please correct that misstake before a number of viewers actually follow your advice and installs them the way you did and atually are worse off then with normal spring washers. Do a video errata on that topic. Show why you were wrong and what is the right installation procedure.
Being one of the top technical writers in the US does not give you immunity on this kind of things.
This issue has been pointed out by me and others already but there has been no real comment from you on the subject. I really hope you did not istall Nord-Locks in the Mini Cooper that way you showed in the video, because then you will have to do it all over again. But hey, maybe that is the kind of iterations of perfection you like….. 🙂
Best Regards
/Per Eklund
We used all Nord Lock washers on the Mini Cooper now with the EVWorks braided tinned straps. We like them very well. They are extraordinarily secure. In fact, it takes quite a bit of pressure to get them off.
We highly recommend them.
Jack Rickard
Ok thats good that you installed them on the Mini. BUT as I said, I really hope you did not put that flat washer in there also(as you showed in the Video earlier).
Best Regards
/Per Eklund
I managed until now to see the video and had no time to read all your comments (which I’ll do), but this video is realy helpful to understand the things going on in my trunk 🙂
The information Jack presented was correct in regards to the Nord Lock Washers. I think you need to go back and watch the video and watch and listen a bit closer. Video: news91010-1280.mov
Listen and watch this time for yourself.
Pete 🙂
Hi Jack,
Thanks for all the work on the shows and your continued effort to educate about not needing BMS. Don’t give up! Every time you say it, it’s in a slightly different way which is exactly the way another couple of people need to hear it, and you’re slowly gaining traction. Just think back to the days when you first introduced the concept here and in the forums.
I have a request for a video topic, and it probably won’t take a whole episode. Could you do a segment on what testing and monitoring equipment an EV converter ought to have at their available when they begin a project? I have a little electronics experience, and I own a little hand-held digital volt meter, but I don’t know if I should have anything fancier than that. Since we’re dealing with pretty high voltage and current, I don’t know if my little DVM will be survive the EV conversion process.
Thanks!
Alex Vieira
Colorado Springs, CO
I had viewed the “No BMS” thing with some skeptiscism, but I have to say after bottom balancing Jack is right. My cells stay very balanced in my testing, despite some pretty extreme discharge and charging currents.
Jack, please look at my data below. I don’t see how these cells could have been delivered in middle balanced condition, but please correct me if I’m missing something. My cells were all delivered in the at the same time, in the same box. Note on the first discharge one battery goes from 3.30V to 3.29V, but another goes from 3.27V to 2.93V (0.34 V out of balance from the top cell). BTW this was withdrawing only 40% of the Amp Hours. Note that after a bottom balance the cells stay within +- 0.02 at the top, and at the bottom, after delivering 93% of their Ahr at 10C, and after charging to 97% full at 4C. Note all the measurements are at 0 Amps (well, just the tiny current used to measure voltage).
3.30V to 3.27V: Cells as delivered
3.29V to 2.93V: Cells at rest right after first discharge of 40% of Ahr
2.99V to 2.96V: Cells after more rest and bottom balance
3.33V to 3.33V: 12 hours after charging
3.22V to 3.22V: After 8C discharge for 4+ minutes
3.33V to 3.33V: 24 hours after charging
3.20V to 3.16V: After 10C 5.5 minutes discharge (93% of capacity!)
3.33V to 3.29V: After charging
etc. — they stay within +- 0.02V of average after big discharges and charges
Thanks Alex. Yeah, let me look at that. The basic tool kit.
Your DVM will be fine. You need a little load resistor and a bench charger or a little 10A 12v supply.
You need a 1/2 wrench with heat sink on it to insulate it.
And that’s most of it.
But we should probably do something.
Jack Rickard
Anonymous:
Did you get these cells directly from the factory? What brand? What size?
3.29 to 2.93 does seem extreme at 90-% discharge.You shouldn’t have anything over about 3.12 or 3.15 at that point. Did you record all the cell voltages? I would kind of have to see the mix.
In a mixed bag, bottom balancing is really the only answer. But I’ve never received a shipment that varied. And I’ve bought from a number of different places.
Jack RIckard
Dan, If I interpreted your question currectly, you did not intend in supporting the BMS theory but wanted to understand your “cheapo” charger?
To avoid lectures from anonymous’ (and jack) you should state what brand and model charger you are using.
What do you mean that your charger only reads the end voltage? Does it shut off when it reaches a set voltage? or go into a constant voltage mode once it hits that “end voltage”? (as prescribed by this army)
Or are you just curious how to pin point your lowest capacity cell?
Need to be more specific to get an answer instead of a post response reiterating what has been said every other video.
JP. Sorry, yes, You are correct. Yttrium doping to keep the crystals smaller and a more regular size. My silly error.
==================================
Once upon a time I perchanced upon a video of a guy who serviced GM EV1’s.
He pumped up tyres and topped up the liquids. He reached between the seats and plugged in a PDA to check the batteries. For balance? It certainly showed rows of numbers.
He said nothing ever needed be done to the batteries and it was the easiest servicing of any car, bar none. My kind of car!
There was no BMS in the GM EV1. Nor in the Toyota EV Rav4, (One EV Rav4 on youtube claims it has clocked 375,000 miles).
No mainstream production electric vehicle using lithium batteries have BMS. Who can argue with that? Can’t imagine motor man’frs messing around with batteries and charging either.
NSVIEW.. I should Cocoa. 🙁
Really am out of my depth here Jack.
Had a rummage on the ‘net. Hope these make your life feel more complete and save you some time. Lots of explanation & samples abound:
Seems the standard colour command is like:
[[NSColor colorWithDeviceWhite:white alpha:1.0f] setFill];
Alpha being the opacity.
http://www.amateurinmotion.com/articles.html
Search: “NScolor”.
Added three spaces in this very importantly helpful url below in case it messed up the page.
http://developer.apple.com/library/mac/#docu
mentation/cocoa/Reference/ApplicationKit/Cla
sses/NSColor_Class/Reference/Reference.ht
ml#//apple_ref/occ/cl/NSColor
Remember, plaguerism is the best way to invent the wheel.
==================================
Jumping needles:
Been there! Old days on noisy A/D conversion etc.
Cheapo bar graphs update position in single steps against time to smooth them out.
Not as fancy as approximating a cosine over 180° across readings. (Cosine approximated from data to save processing cycles) Would give a softer start and ending and should keep the needles professional while cruising.
Hope it’s a help. All the best.
I’ve just started mooning at GTK+,Cairo & Glade.
Errata:
(Cosine approximated INTO data to save processing cycles)
Thanks Jack, I agree they shouldn’t have swung so much in Voltage right out of the box (3.29V to 2.93V, vs. 3.30 to 3.27V for another cell), but the cells did work great after a bottom balance. They were new China Hi Power cells. I didn’t buy them direct from the factory, but there was at most a week of delay going through the reseller, so it wasn’t like they were sitting around for years in a dusty back room. When I finish testing them I’ll write it all up with individual cell Voltages and temperatures and send you a link.
Oh yeah, forgot to say, these are 10 Ahr cylindrical LiFe cells.
Can anyone please tell me if not using a BMS would apply to these cells as well? Could I simply connect a 96v charger to a 32cell pack and walk away? The chemistry is slightly different.
– Nominal Voltage: 12.8V (4X 3.2 V)
– Nominal Capacity: 60 Ah
– LiFeMnPO4 chemistry
– Operation Voltage Range: 11.2 to 14.4V
– Weight: 9.2 kg or 20.3 lbs
– Dimension: 125X280X180 mm or 4.9X11.0X7.1 in
– Max Charging Current: 3C
– Max Discharge Current: 3C (continuous) / 10C (pulsed)
– Cycle Life : >1500 (80%DOD)
– Operating Temperature: -20 to 65 C or -4 to 149 F
– Self Discharge Rate: <3% monthly
I would greatly appreciate any input on this!
lespaul109_39 at hotmail.com
The only experience we have had with China Hi Power cells are the white cells we installed in the Vantage Greenvan. We had them for over a year. There was a reason we had them over a year without putting them into a car, and then put them in a $4000 car. They just don’t make the grade here.
The 10 AH cyclindrical cells I’ve never even seen. I cannot comment on their chemistry or performance, and have no position on what you should do about a BMS. That you got them from a third party instead of directly from the factory pretty much tells the story.
Nothing in any of our videos applies to your China High Power Cells. The little bit of testing we did on them did not even point to their USE here anyway. And we did not further testing. They were 200AH cells and prismatics. You have something else entirely, and I know NOTHING about them. I cannot advise you at all.
Jack Rickard
Jake:
LiFeMnPo4 cells are an entirely different chemistry and have some interesting safety issues. But we have ZERO experience with them and cannot advise you about them AT ALL. Not good. Not bad. Not at all.
One thing I HAVE learned is that lessons from Pb or NiMH cells for example, become total nonsense with LiFePo4. So I’m not an advocate of battery cross dressing at this point.
Jack
I have 64 TS160s that were all withing .01v of each other out of the box. They sat for 6 months before I installed them. I wired them in series, charged them with my Elcon, but kept an eye on cell voltages. One cell hit 4.0v before the others… I bypassed it and continued charging. This happened two more times to two other cells before I disconnected everything, wired in parallel, and charged to 3.8v.
In my case, there were differences in SOC. I balanced at the top, but designed the system to draw down to 50% most of the time… and charge to 3.8v normally.
Perhaps I was merely seeing the difference in total capacity between individual cells. But had I just wired in series and charged to 3.8v per cell (243.2v) I would have overcharged cells.
I merely present these as data points. I don’t have a BMS. I probably will not run one believing that staying well within the “middle” of the pack’s SOC is the safest method… be that directly in the middle or biased one side or the other.
To clarify further what Jay, the battery PhD, said was that the ideal solution *was* per-cell management… but from a cost standpoint it didn’t make sense. That’s not quite, there’s absolutely no merit in having a management system.
I am ASTOUNDED at the reinterpretation of what Dr. Whitacre has said here in this video. There is actually a discussion on DIYelectric where a noted BMS vendor explains not only that I don’t know what I’m talking about, but that what Dr. Whitacre REALLY said was that you DO need a BMS.
Several of you HERE have explained to me patiently that I have misinterpreted very clearly presented English language presentation by a noted battery expert. I simply have not. I read, write and comprehend English at a reasonably expert level. And for those of you who understandably don’t (45% of our viewers are non-U.S.) what he said is NO BMS and he said it in to me a very satisfying level of detail.
This video has caused such ire, that one BMS proponent had notified Dr. Whitacre that I was “using his video for profit” and he indeed contacted me. I provided him a link to the blog and explained what we were doing and he noted that he was perfectly happy for me to continue the post. We then discussed BMS and batteries briefly by e-mail. His final:
“Jack – we’ve been driving a large thundersky pack sans BMS for a while here at CMU, no issues. If you understand the batteries, you just don’t need it and you are safer without it.
See:
http://chargecar.org/conversions
BMS is just a dumb idea and you can quote me on it!”
So there. I’ve quoted him on it. There IS no confusion nor reinterpretation necessary. BMS is just a dumb idea. You can quote him on it. You can quote me on it.
What he said. What I said.
Jack Rickard
=-=-=-=-=-=-=-=-=-=-=-=-==-=-=-=-=-=-=-=-=-=-=-=
http://neon.mems.cmu.edu/whitacre/whitacre.html
Jay Whitacre
Assistant Professor
Materials Science& Engineering
Engineering and Public Policy
Carnegie Mellon University
4323 Wean Hall, (412)268-5548
Pittsburgh PA 15213
whitacre@andrew.cmu.edu
Jay Goldman:
What you were seeing is a perfectly normal exhibition of CAPACITY variance, not SOC variance. Had you observed precisely the same cells that were hitting 4.0 v at the end of charge at the end of discharge, you would have found that they exhibited LOWER voltages than the other cells. They are simply smaller batteries.
By top balancing, you have moved ALL the difference in capacity to the bottom of the charge curve. That means there is a greater distance between discharge curve knees cell to cell. And that poses a strong danger to your cells at 90% discharge and beyond. The gig is that during DISCHARGE, you might easily see 400 or 500 ampere draws, and if your least capacity cell goes over that knee at that rate before the others (and it will) it will be driven to zero volts and into cell reversal in SECONDS. Not minutes. Seconds. Like 5 or 6 seconds of acceleration.
SINCE you have top balanced, your only practical remedy is to bottom balance, a tedious process.
I have experimented with these cells extensively and can tell you you will simlpy get better results charging your string to 3.65v X n cells rather than 3.80. You gain almost NOTHING in range at the higher voltage level. And indeed, the very fixed charger TS sells for 72 volt packs of their cells charges to 87v – 3.625 volts per cell.
Climbing that vertical wall of death with your charger adds almost NOTHING to range. It really just causes heat at the gain of a miserable 1.5 AH. Your cells will last much longer at the lower charge level.
Jack Rickard
Jay Goldman:
Actually, that IS what he said. And I’ve just quoted him saying it again by e-mail. “BMS is a dumb idea. You can quote me on it.” There, I’ve done it again.
At the end of the video, he was posed a number of hypotheticals by the audience and responded hypothetically. It is all correct and I agree with all of it. In practice, it’s nonsense and poses dangers and expenses that the benefits simply do not warrant – even approximately. But if you want to “theoretically” discuss whether I would LIKE to know the exact voltage of every cell in my pack, at all times, the theoretical answer is of course YES. Now does that mean I’ve reversed my position and advocate use of a BMS? It simply does not.
The expenses, the dangers, and the complication weighed against the knowledge gained by knowing each cell voltage at all times fails in the analysis. It simply doesn’t make any sense. You are not PROTECTING anything, you are ENDANGERING it with negative temperature coefficient semiconductors and spaghetti wiring interconnecting VERY high power battery cells, and you can easily spend a couple of thousand dollars doing this to protect a pack of $6000 total expense. I can buy EIGHT 200AH cells for the cost of the Elithion for example.
If you want to experiment with the cells and learn about these voltages and currents, set up a BENCH like I have and test them there. I’m all about learning and testing and measuring. But don’t do it in your car. And for the love of life, don’t plug a car into the wall with one of these fiendishly clumsy amateur devices hooked up to it and go to bed.
BMS is a dumb idea. What he said. What I said.
Jack Rickard
Here is a bit more about these packs.
• Advantage of GBS 2nd Generation Li-ion Batteries Compared To Thunder Sky Batteries:
– 10% higher energy density by weight and by volume;
– Superior safety performance due to patented new safety valve and pressure cap designs;
– Batteries do not explode or catch fire when batteries are shorted or punctuated resulting in internal shorts;
– More robust connection due to new electrode terminal design using four rivets or four screws per terminal; Prevent loose connections caused by vibration;
– Reduced impedance due to improved electrode terminal design;
– Improved cycle life;
– New cell structure better facilitates BMS integration and battery pack formation, and improves air circulation.
Here is another fire in an EV with an Elithion BMS:
http://www.pressdemocrat.com/article/20110319/NEWS/110319406/-1/archive?p=1&tc=pg
Hi Jack,
I loved this video the first few times I watched it months ago–but now, it seems not to be working! I’m getting “Video not found or access denied”. This was an amazing resource for those of us just getting started in the EV world–can you try to bring it back?
Thanks,
Scott