Monday, November 8, 2010

Itsy, Bitsy, Spyder 550 and Battery Balance Myths

In this week's episode, we assemble new mounts for the 2009 Mini Cooper Clubman drive train - somewhat colorfully.

We also wrap up our Vantage Green Van Battery Makeover. We have thrown out 775 lbs of AGM lead acid batteries and replaced them with 27 High-Power brand 200AH cells that I never liked well enough to put them in a car. The resulting pack weighs 350 lbs.

Because I had used some of the cells in some experiments in the lab, we were forced to bottom balance the pack, a three day adventure. Young Hauber subsequently drove the Vantage Green Van 103 miles on a single charge. This is a bit of an improvement over the 16 miles I had achieved during one test drive with the AGM's which only had 600 miles on them. Gentlemen. Lead is dead. Long live the LiFePo4. Get over it.








Also featured this week is Duane Ball's most excellent Spyder 550 build. And therein lies a tale.

Duane is about to take delivery of a Beck Porsche 904 from Special Editions Inc.



This is a fantastic new Beck reproduction. Just the roller will cost over $50,000. To raise funds for this project, Duane wanted to sell his beloved Spyder 550. So for a mere one Brazillion dollars, I bought it.

The car is a superb build with essentially duplicate components from Speedster Part Duh, but with a much better Brusa charger in it. Duane did use a PB-6 potentiometer box and a cable assembly to act as a throttle/accelerator, and everyone is pretty much unanimous that this was the one POS error in the car. It just feels awful. We'll be changing that.

In the meantime, the original Speedster is nearly 2 years old and has about 10,000 km on it. Duane's Spyder has 5811 km on it. So young Hauber and myself took advantage of a crisp, cool but beautiful October Sunday afternoon and drove on U.S. 61, a winding blacktop hilly highway to Perryville Missouri and back.

For some reason, this took more energy than it normally does. But to an interesting result. In the Speedster, we actually rolled to a stop just in front of our new shop at 601 Morgan Oak. We had to push the car inside by hand.

We had used a total of 195 AH out of the Speedsters two-year-old 180AH pack. And six cells were well under 0.5v STATIC once we pushed it inside and put a meter on it. This car has had NO BMS of any kind, other than a single CellLog8s misadventure for three days monitoring 8 cells and resulting in, you guessed it, a fire. It has been repeatedly run to full 100% DOD in testing. ANd we still managed 195AH from a 180AH pack (two strings of Thundersky TS-90AH in parallel).

So we've finally destroyed it. Well, at least we haven't done it any good. But as of last night, all cells were taking a charge and climbing in voltage. I'll check later today to see if any also LOSE their charge after the full charge. If they don't, this is incredible.

The Spyder 550 was also interesting. The cells are much easier to access, and we have young Hauber now to do the heavy lifting. It used 176 AH on the trip. It is much lighter (as is Hauber) accounting for the difference.

The Spyder is 1890 lbs with a 44% front and 56% rear weight distribution. As the cells were all still over 3.00v, I sent young Hauber for a couple of sleds of Stag Beer, his favorite housekeeping task here at EVTV. On return, the vehicle was showing 157 km total on the trip, 183.5 AH used, and a static voltage of just 95.65 volts or an average of 2.517 per cell.

After letting the cells rest and recover for a few minutes, I had Hauber take a reading of every cell in the car. Here's what we found....


Of course, the excellent news is that with NO BALANCE DONE EVER, and NO BMS EVER, a car with 5811 kilometers (3611 miles) on it was in excellent shape, excellent "balance" and all on the very vertical face of the discharge curve well below 3.00v and with 183 ah withdrawn from a 180 AH pack.

I have an odd ability at pattern recognition that is only occasionally useful for any thing. But the first 10 cells sort of jumped off the page at me. It raises a question.

The first 12 cells are arrayed across the front of the car, and so they may be getting dramatically more cooling than the cells in the rear of the car. But there are 12 of them. And cell 11 and 12 simply do not match the first 10.

I went to look at the overall pack voltage, and of course Duane had used a Xantrex. I recall the same problem on the Mini Cooper. I had originally used the lower 10 cells to power the Xantrex at the 35 volts it likes. Over time, this depletes the cells actually. The Xantrex doesn't draw much, but it draws enough. Duane had apparently used the first 10 cells, just as I had, and later went to a voltage divider across the entire pack to power the Xantrex, just as I had. It is eerie to see your MISTAKES replicated in somebody elses' build.

We put a charger on the 10 cells at about 30 amps for probably 5 minutes. They all fell right in line. And while we were doing it, with hit Cell 22 with a very brief shot to bring it up as well. I don't have a cogent theory for that one.

Here's what I DO get from all this. The "they gradually go out of balance" theory used to support the increasingly dubious case for the necessity of the BMS is just pure D Grade Bullshit. That is, bullshit that is not even very high grade bullshit. Here we have two cars that have NEVER been balanced in any direction, with a couple of years on one and 3600 miles on the other, where NONE of that has been done, and they both operate so well, that I can take 195 AH out of a 180AH pack apparently without mortal damage.

The BMS adherents, who increasingly start to look also like the BMS designers, who also look a lot like the people selling the BMS's are simply pumping BS for cash. Their products will actually DAMAGE your car, and the purported gain from using them is total nonsense. At this point, since we've made this information public numerous times, I believe that they KNOW or SHOULD HAVE KNOWN it was bullshit, and are actually lying to you for cash. There is no innocent "difference of opinion" going on here.

I have actually set out to cripple a perfectly operational electric car that I have well in excess of $70,000 invested in and have apparently FAILED to KILL IT. As I'm normally very good at these things, we'll see with some folow up tests that almost have to show some DAMAGE at least. But I could not kill the car, and did not drive any cells apparently into reversal.

If you top balance your cells, you CAN. And we have in the past done that empirically and publicly where anyone can duplicate the experiment.

ANd here is the problem with the online millieu, people have extended the right to "have their own opinion" to somehow include the right to have their own facts. It does not.

That said, they remain YOUR cells, do with them as you may.



The Rinehart controller we are installing in the Mini requires a switched ground signal to indicate the brake light is on. I don't quite get this. Brake lights run on 12v and you usually see 12v coming on when you apply the brakes. But it's easy enough to invert. A relay would work quite well, but we put on brakes a lot and so we're going to wear out the mechanical relay. Another way is simply to switch a transistor, which inverts the signal. A MOSFET is a good candidate because of very low forward resistance. We used a 75 Amp 30v MOSFET that is gross overkill to switch milliamps. But this will work quickly and probably won't ever wear out. This is the little diagram to invert our brake signal to a ground instead of 12v when we press the brakes. The 12v brake signal switches on the MOSFET applying ground to the output.



Jack Rickard

52 comments:

  1. last week I had a german electronics journal (www.all-electronics.de for the german speaking fans) in my hands with several articles published concerning with BMS electronics. One article by Jack Marcinkowski of National Semiconductors. You may look for an english version for your enjoyment, but I don't have a link.
    The trouble seems to be that the automotive industry, ramping up slowy for ev-production, does not have any trust whatsoever in the batteries and would probbably like to complement them with a complex system to at least have the impression of having some kind of control. I really wonder what goes on in their laboratories where they undoubtedly also do a lot of battery testing to destruction.

    ReplyDelete
  2. Hi Jack et al.

    Regarding balancing of cells... :)

    I did a lot of work before I mounted the cells in my car - but it took almost a month to get through all 64 cells - so at the end of the initial charge/discharge/charge cycle the some of the cells had 1 month worth of self discharge - so they was not the same SOC when they was mounted in the car.

    I have measured the cells at different points and was a bit concerned about the difference at the end of charge. At first I could measure anything between 3.5 and 4.1 volts in the string. My charger is set to an average cell voltage of 3.68V (TS LFP90AHA's).

    My charger will charge CC at 0.1C until 118V is reached and the hold this voltage in 15 minutes.

    In order to (top) balancing my cells I purchased a 5V/120A power supply (HRP-600-5). This power supply can be adjusted to 4.1V and will work in constant current mode from 3V. It has a build in fan which is current controlled.

    I measure the voltage of each cell pair after a full charge and top off the ones which is below average. With the power supply it is real easy, since the fan will start up when the current is 120A and stop when the 4.1V is reached.

    I understand your reason for bottom balancing your cells, but I think the assumption here is that the cells needs to be very close in capacity in order for it to work.

    My way is so much easier to do on a regular basis, since it will only require a full charge and a 10 - 15h for the batteries to rest.

    My approach has the disadvantages at the end of discharge, where my cells will not be at an equal SOC. But I very rarely (never) even approaches the end-of-discharge point, but I do on an almost daily basis reach the 100% charged mark (or at least 90 - 95% mark). And the inverter in my car will go into emergency mode (daughter mode) when the midpoint of the pack is not within +/-1V of 50% pack voltage. So this takes care of the problem of reversing a cell. If two cells drops symmetrical in the pack I have a problem, but now we are just making up unlikely scenarios.

    I am not trying to tell you or anybody else what is the best approach here, since it may change from setup to setup, but I am just telling my story for what it is worth :)

    If anybody wants to se numbers, feel free to do so at: http://www.citystromer.dk/lithium%20equliazation.xls

    In order to track changes I plot the normalized values, since the OCV is a function of to many things (temperature, time, SOC, ….).

    Regards,
    Martin.

    ReplyDelete
  3. 10'000km in one and 5'800km in another over two years.
    do you guys just not drive your cars?

    interesting the car with more km failed at 100% dod, when the other did not.

    perhaps its more a matter of how many cycles done rather than time that causes imbalance.

    ReplyDelete
  4. Thank you for the show, it is always worth waiting for!

    But please tell Matt to wear gloves when welding. UV radiation from the spark will give some nasty sunburns.

    ReplyDelete
  5. I wasn't going to say anything, we're all adults here, but it pains me to watch Matt hurting himself.

    The UV from welding is unlike any naturally occurring radiation on earth, because the atmosphere absorbs these wavelengths from the sunlight that hits the surface. The local welding "sunburn" we who weld have all gotten from time to time, is in fact a pretty nasty UV blast of unknown danger...

    Welding without gloves is literally turning yourself into an epidemiological laboratory animal in an experiment of exposure to this very exotic and obviously powerful radiation, which is likely a mutagen/carcinogen, but there's almost no actual data on it. Animal studies haven't been done, and they wouldn't apply very well anyway because human skin is uniquely sensitive to radiation and we live so much longer than rats and pigs. The anecdotal data that exists isn't encouraging.

    Get used to the gloves, Matt, and everybody else from American Chopper to Junkyard Wars, too. Its worth the trouble. You'll want your hands to still be there and working properly 40 years from now, yes?

    This practice is like painting without a respirator. Not very wise...

    TomA

    ReplyDelete
  6. Jack,

    Shakin the hornets nest again I see. Kudos to you. Love the knee jerk reactions. Love the Speedster too.

    Matt,

    I won't say one way or the other but you need to know that the UV from welding is worse than what you get from the Beach. It is quite strong. I'ts a safe bet to coverup when welding. Nice job on the cv mount, well engineered and constructed. Nice job on the van too.

    Pete :)

    ReplyDelete
  7. Anonymous:

    There is no mystery, and no failure. The Speedster weighs 2190 lbs. The Spyder weights 1890. The Speedster driver weighs 260. The Spyder driver I would guess at 160. The two cars drove the same distance, in fact on the same drive. They had identical pack sizes of 180 Ah.

    Spyder took 176.4 AH. Speedster took 195 AH.

    So what's the mystery?

    Very pleased to report that all cells on the Speedster came up. The car was fully charged with the addition of 191.5 AH, indicating a probably loss in capacity of 3.5Ah.

    As to how much we drive the cars, I don't quite know how to respond. Apologize for not driving enough to suit you? What??

    The ability of the online crowd to pick at nits has now reached "you don't drive your car enough."

    I suppose if I did THEN I would need a BMS? Enough being precisely enough to need the BMS of course.

    Jack

    ReplyDelete
  8. Martin:

    Your considerable efforts to conserve your battery pack are murdering them. Quit it.

    Jack Rickard

    ReplyDelete
  9. On lead vs lithium , I had a bit of a revelation a few weeks ago. Here's a copy of a post i made on the diyev forum:

    Ok can you guys check my maths here cos something can't be right. So i'm messing about with a few headway cells for the bike project and got to thinking....16ah cells can easily sustain a 150amp discharge so a 4p pack could easily support a 500amp acceleration for say 10 seconds.

    So a 45s4p pack of 16ah cells would yield 144v and 64ah. Little over 9kwh. Cells weigh about 400g each so pack weighs 72kg. Throw in the cell blocks and links call it 80kg.

    Right now i'm running 120v of 10 odyssey pc2150 agms. At 80% dod thats about 9.6kwh. This weighs a whopping 340kg!

    So i can dump a massive 260kg and achieve the same (if not better due to the weight loss) performance from the car. Allowing for the loss of the 10 gallon petrol tank then the car would come in at stock weight if not a little less. Picking up an extra 24v would also help top end rpm.

    At 23usd per cell i'm looking at $4140. At current exchange rate thats about 3k euros. In or about the cost of the odysseys.


    Most importantly that pack can be easilly expanded as funds allow. so...........what am i missing??

    Yesterday I paid for the cells. I was particularly interested in the bottom balancing as my experiments to date would seem to indicate the headways are very imbalanced out of the box but repeated charge / discharge cycles actually "pull" them closer together.

    I know the above might have been glaringly obvious to some but not to me until recently.

    then again i'm not that smart ........

    Damien

    ReplyDelete
  10. Jack: Your considerable efforts to conserve your battery pack are murdering them. Quit it.

    Could you elaborate a bit. Is it the 3.68 V average voltage of the Thundersky's you think is murdering them or...

    Martin

    ReplyDelete
  11. Damien:

    You're making a lot of connections there with 45x4 cells, but no, I think you're getting the picture.

    I cannot stress strongly enough how different the vehicle becomes in ALL ways with the weight loss. It accelerates better. In the case of the Vantage Van the severely sprung front suspension suddenly stood tall, and it of course takes MUCH less energy to move the car, while we have MORE energy to do so. The advantages are coming at you from all directions at once.

    Lead is dead. It is SOoooooo dead....

    Jack

    ReplyDelete
  12. I run a bottom balanced pack with no issues so far. Probably saved some cells during a very low SOC run. Instead of a BMS I would recommend the best thing you can do is try and specify a closely matched pack if you can. My 36 SE/CALB 100 cells came in at 110-114 ah. If you could order all cells at say 110, or even 110-111 ah, where most of mine are, even if you had to pay a little extra, you'd essentially be running top and bottom balanced. See if you can get internal resistance closely grouped as well. Once the cells are close enough in capacity and resistance I would think the whole top, bottom, and BMS debate becomes moot.

    JRP3

    ReplyDelete
  13. Jack, would you like to elaborate on why you think I am killing my battery?

    Martin.

    ReplyDelete
  14. Hi Jack,
    You might laugh about this, but
    I had a completely weird idea these days!
    ....has anybody already tried to (re-)balance
    his cells by switching an extra cell
    to the pack and toggling it between the
    highest and the lowest cells in order
    to equalize them ????? or just one by one
    without further controlling or measuring.
    Ok, switching is a little bit tricky,
    but with four mosfets per cell it should
    work more silent than with relais.
    if a little difference in voltage is enough
    for transporting some energy between the cells
    it might work....on top as well as on bottom.
    I have not yet tested this, but some day i will. ;-)
    greetings from Germany!
    Franz

    ReplyDelete
  15. Tesla, TSLA near 30.00/share. Up 16% today. Go Tesla!

    JMS

    ReplyDelete
  16. Heh! Hey Jack, At what price does Tesla pay for your roadster? I only grabbed a few at the IPO but a 55% bump from then is only the beginning.

    I'm sure EVTV could find some interesting tests for one. You can always make a small concrete bunker to isolate it from your belongings if the bms burns it to the ground. :)

    ReplyDelete
  17. Actually, I have a little side bet going with Dan Frederiksen. He keeps telling me to sell, and every time he does, I buy another 5000 shares.

    Right now I have 10,000 shares of Tesla and I'm up about $60,000 - I've been in and out several times. I figure another $15,000 and I'll sell and order a Model S.

    Jack Rickard

    ReplyDelete
  18. One good thing to do with those horrid shunt balancers........lower the set point, put some crocodile clips on it and use it as a bottom balancing tool!

    ReplyDelete
  19. Jack: I fail to understand why you ignore my question about why you think I am destroying my battery. If you think I am just wasting your time and is an idiot for not seeing why, can you at least tell me so.

    I fail to see why my way of top balancing my battery will destroy it. I also fail to see why an average maximum voltage of 3.68V will destroy the battery. My normal charge cycle will stop at the 90% charge mark and the balancing charge is less than 2Ah (about 1%) so it of great consequence for the cells involved.

    If I am unable to understand this should be a problem, I will bet there's also a majority of your readers and viewers that is unable to do so - so please tell us why you think it might be a problem.

    Your MOSFET inverter will work just as well with a NPN transistor and your statement of the forward (ON) resistance is irrelevant is type of application.
    I would at least double the 100 ohms and 470 ohms resistors (and switch them if using a NPN transistor). You are burning almost a quater of a watt in those two resistors to no use at all.

    Martin.

    ReplyDelete
  20. Martin:

    I'm ignoring you because it is just too painful for words. First, I'm not burning a 1/4 watt through the inverter. 12/570 =0.021. That's 21 milliwatts and as usual you are off by a factor of 10x.

    The rest of it is more than adequately covered in our videos. We've demonstrated precisely how to murder a cell using Top Ballancing, and you've apparently taken that as an instruction to do so. They're your cells. Murder them if you like.

    Repeating the same information to you over and over is just not a good use of my time, or anyone else's here that has to read all this.

    So I will ignore your question if I feel like ignoring it.

    Jack

    ReplyDelete
  21. Oh, and it only uses 21 mw when the BRAKE is on.

    Jack

    ReplyDelete
  22. Ok got your point, but I do think your are wrong or at least misunderstanding me. I am not charging my cells different from what you are doing - as far as I has been understanding it at least. You are, as Matt showed us in the last video, equalizing your cells at the bottom at the charge curve. I am just doing it at the other end.

    What I am trying to do is to insure that the cells will reach the same voltage at the maximum SOC point (in my case approx. 90% SOC). I do this because I saw a rather large difference in the voltage over the cell at the end of charge point (average 3.68V in my case).

    I will illustrate my point with an extreme example: Take a 80AHA cell and a 90AHA cell and connect them in series.

    If you equalize them at the bottom and charge them to an average 3.65V - what will the SOC be of each of the two cells at the end of charge ?

    If you equalize them at the top and discharge them until you have a difference of 1V in the two halves of the string and then charge them until the average is 3.65V - what SOC will the two cells be at?

    I know the example above might seem extreme but I did measure all my 66 cells (90AHA) and they had a capacity from 90 - 100 AHA.

    Since I normally use my pack from 90% to 40% SOC before I recharge them it seems more correct to equalize them at the end-of-charge point (90%).

    I do not see that I go against what you have said in any of your videos.

    And by the way:
    P = U * I = U * U / R = 12 * 12 / 570 = 253 mW
    An I do know it's only during braking and that an normal brake light is about some 20W. But why waste energy needlessly?

    I am sorry if I have been offending you with my ignorance.

    It is your blog and you may behave as you feel like on it, but I do think it is kind of rude to ignore people just because they might not be able to catch your steep learning curve.

    I hope you like the Nord Lock, which I kind of pointed you toward. You are making a mistake with the additional washer - remove it as the good people of Nord-Lock states in the material on their web site.

    Martin.

    ReplyDelete
  23. Martin,

    The link to the video seems to be missing during the revamping of the site. (By the way Jack, I like the direction your headed with the new site).

    Martin, this blog post about the video in question goes over many of the same points in detail. Please read this.

    http://jackrickard.blogspot.com/2009/11/get-rid-of-those-shunt-balancing.html

    Michael

    ReplyDelete
  24. Martin,
    I think more accurate would be to say you have a higher risk of killing your cells by top balancing than bottom balancing. With top balancing if you ever get caught out on the road and take your pack lower than expected you could end up pulling high current through your lowest cells, probably damaging them, especially if they are driven to 0 volts. Bottom balancing avoids this. However, if you never over discharge a single cell then you won't damage anything with a top balanced pack, and top balancing does make charging a bit easier. Personally I prefer the comfort that bottom balancing gives me when I'm out on the road, and I deal with charging by setting the charger to slightly undercharge the pack.

    JRP3

    ReplyDelete
  25. Martin,
    Sorry, I have to agree with Jacks calculation on current across those two resistors.

    However, Nordlock washers being a hard Mat'l suitably splined in the flat faces will not be so "grippy" against the hard stainless bolt either. I'd suggest a soft Alu washer under the bolt head. After one is certain they have settled into place, bend the washer up against the flat of the head.

    I'm only saying this because nobody is keen on wirelocking across 350 volts :)))

    ReplyDelete
  26. Michael: I have read it and my opinion still stays the same. I would not feel any more secure by a bottom balancing of my pack and with my driving pattern.

    Andyj: Jack calculated the current as you also noted, but I am talking about the power.

    JRP3: I am nowwhere near the bottom of the cells. But I daily are around the top. If you have a faulty cell with less capacity, then that one risc a overcharge when you have bottom equalized the cells (imho).

    I'll leave it here, but if any of you want to take it up off-line feel free to contact me.

    Martin.

    ReplyDelete
  27. Martin:

    This is precisely what the latest blog post is about. You're entitled to your opinion. You're NOT entitled to your own set of facts. This concept that everyone gets to vote on reality is just absurd.

    We demonstrated in an early video how specifically to kill a cell, and indeed killed 3 in front of you, doing precisely what you're doing. This week, we rolled a Speedster to a STOP without losing a single cell. It has NEVER been bottom balanced at all. We just failed to top balance it. Had we top balanced it, there is no possibility we would not have lost a cell.

    So have a triple serving of opinion....it is TOTALLY irrelevant.

    Jack RIckard

    ReplyDelete
  28. YAY!!! some people are getting it....when you bottom balance, if you do drive the car further than you should your less likly to root cells!! Don't say that will never happen, it will!! when its pissing down with rain, your late for one of the most import dates of you life, or a job interview........you switch your rational brain off and keep driving!!

    If Top balancing was as safe as Bottom Balancing why would you bother spending $2-3K more to do it....it's just plain dumb!

    ReplyDelete
  29. Sometimes the lazy brain overpowers the rational brain

    ReplyDelete
  30. I noticed Duane's Spyder 550 on eBay the other night...

    If you top-balance, set up your Ah meter to shut down your ride when the Ah capacity of your weakest cell is reached. One just has to go through and figure out the Ah capacity of all the cells every so often.

    A battery pack conditioner that can bottom-balance a pack would be a handy device. A tap off every cell and have it do automatically what Jack's apprentice does manually. A BPC instead of BMS. Used only occasionally and not built into the vehicle like the charger might be. -Klaus

    ReplyDelete
  31. Again, the assumption that because I am strongly advising AGAINST top balancing, I must be a bottom balancer.

    Not precisely accurate. In the case of both the Spyder 550 and the Speedster, we didn't really bottom balance.

    Occasionally, it is necessary for a variety of reasons to do a "corrective balance" because of an unusual sitution. In the case of the Vantage Van, it was because I had used 10 of the cells elsewhere out of the pack and they simply did not match in SOC with the other 17 on installation.

    In the case of the Spyder, we simply brought up the 10 cells that had been dislocated by the Xantrex power requirement.

    Generally we do NO balancing. But as I've said numerous times, if you add a new cell, replace a cell with a terminal problem, or in situations just like these two, we try to get them "married in" by bottom balancing.

    Top balancing, THE most common BMS technique, will absolutely cause cell failure on even fairly slight overdischarges. Not damage. Failure. Damage is where you don't get to drive as far. Failure is where you get to buy another cell.

    Jack Rickard

    ReplyDelete
  32. Martin,
    0.23 watts is barely worthy of note and due to the nature of its use I would use a plain 1/4 watt resistor on the input. Maybe all round, I forgot what the load is.

    I have a question if anyone would like to answer.
    The internal resistance of these cells. Is there a variation along the battery's charge curve? I've looked for a suitable charge graph to no avail or am I blind?

    The reason I'm asking is if internal resistance drops when the batteries voltage rises near the top of the charge... Thats not good if it means fiddling and faffing to ensure other batteries catch up to equalize.

    I'm not a person who likes to fiddle. Maybe it is why man'frs will install BMS to avoid the nut behind the wheel messing around.

    ReplyDelete
  33. Internal resistance does increase as you discharge, and as you note, correspondingly decreases to a point on charge.

    We have not done a curve of this as it has looked like information of a rather limited use.

    Jack Rickard

    ReplyDelete
  34. The entire battery pack is only as good as the "worst" cell in the pack?

    If that is the case, doesn’t that fact alone tell you that top balancing a pack will "stretch" the worst cell to its limits towards fully depleting the pack?


    Padraic

    ReplyDelete
  35. Hi Jack,

    Your views on batteries are interesting - if, er, perhaps stated with rather more vigour than would be considered normative in my generation over here in the UK!

    I think you are right to challenge received wisdom though: if nothing else, some of the failure modes of shunt balancers look nasty (a few mA leakage would total a cell and a few Amps leak could start a fire)

    One point does make me pause however - I guess your cars are quite "young" from a cycle life viewpoint i.e. the equivalent of 35 - 100 full pack cycles with cells that are specc'd for 2000. Suppose balance starts to drift after a certain percentage of cell cycle life (as happens with AGMs)?

    ReplyDelete
  36. Padraic: It doesn't matter if you are top, bottom or NOT balancing a pack if you have a cells that if very different from the rest. The faulty cell will either be over-charged or dischaged or both depending on the situation. This is why a cell _monitoring_ system might come in handy. Forget the balancing type. But you would still need to know what to with the data in order to protect the battery/car. The only reason I balance my pack to be able to see long term drift of the cells and act upon it.

    Martin.

    ReplyDelete
  37. Hello Martin:

    Have you ever seen a discharge curve?
    And if so. Have you produced a discharge curve for your cells?
    Why do you think that your cells are drifting?
    Have you monitored the cells with and without a BMS for comparison?

    Kind regards,

    Padraic

    ReplyDelete
  38. http://abclocal.go.com/kgo/story?section=news/local/peninsula&id=7773276

    Please review this gentlemen. I conferred with Mr. Young at his behest often and at length on this project nearly a year ago. I implored both he and Perrone to NOT put a BMS in this system. I was outvoted by an awful lot of very smart people.....who were talking with great authority about their OPINIONS on what they knew NOTHING about......

    Here's the result. Certainly one of hte most expensive EV builds in the country, and I thought very beautiful.

    Jack Rickard

    ReplyDelete
  39. (Suppose balance starts to drift after a certain percentage of cell cycle life (as happens with AGMs))

    You are making a huge assumption that these Lithium Cells ACT the same as LEAD.

    Pete :)

    ReplyDelete
  40. So high power electrical systems are designed without the ability to spell "fault tree analysis"?
    Thats frightening.

    ReplyDelete
  41. propellator: FTA's are not better than the people doing the analysis. But yes, many EV's are built, used and maintained by people without any great experience with high power electrical system design and fault analysis.
    That said, I have also seen what professionals can do - and it is not always pretty and analyzed all the way.
    Three Mile Island and Chernobyl still happened...

    Martin.

    ReplyDelete
  42. So...
    Don't store your expensive, flammable keepsakes in the same building as you BMS "protected" electric car.

    JMS

    ReplyDelete
  43. Bottom balancing and top balancing.
    I believe Jacks approach to charging is good. There are too many variables that simply "work well enough" to be a pedant.

    But,
    I have a belief about BMS, beyond the extra money made by a manu'fr. There are a lot of people who simply want an automatic and fool proof charging and or discharging system. I would accept their fears. Vivre la Difference!

    On internal resistance dropping:
    Jack says "We have not done a curve of this as it has looked like information of a rather limited use."

    My take is the lower resistance batteries which already contain more charge will attempt to charge quicker as they beat the others. The errors increasing after each cycle... But the axe swings both ways if it's matched by a similar quicker discharge due to it's lower internal resistance. Then who needs a BMS?

    I believe from what I can find out, this is so.
    Jacks use of a quality charger to remove the chance of over voltage charging and letting the batteries relax at 80% capacity seems passable.

    But the thought of ducking inside every so often galls me with what should be a hermetically sealed up, fit and forget item. Especially in dank and soggy UK/Ireland.

    The cold run to the big garage and the cooler batteries in the wind doing less well on battery capacity gives good reason for a dry, insulated, climate controlled battery box. I would attached a wire to every battery jumper for remote monitoring of loosened bolts and internal resistances. Maybe tickle the opposing performing batteries through that line if need be on a timely, low current basis. Then every one can be happy.

    ReplyDelete
  44. AAARGH!
    Please note my last paragraph. Do not do it! Not by direct connection or you'll be shorting across other batteries. The "Earthy end" will need to be raised to a certain voltage or be entirely independent of the rest of the pack.

    ReplyDelete
  45. Andy:

    Precisely. In a number of cases, it isn't the BMS that caused the problem, it was the BMS WIRING.

    The LESS you mess with these cells, the happier they seem to be. That said, we are noting significant differences in operation under temperature extremes. And we're starting to look at what would be necessary to heat batteries. They do not get terribly warm interestingly, and have a LOT of overhead on temperature. But they do not appear to like cold, and even operate significantly better at 120 degrees under stress for example, than they do at room temperatures. This all points to some form of heating system, particularly in the winters here.

    We're looking at something like that for the Escalade project.

    Jack Rickard

    ReplyDelete
  46. Would wireless sensors/transmitters atacted to cells work.

    ReplyDelete
  47. They might possibly. There is a Zigbee device called an XBee available from Sparkfun. They are used to form mesh networks and they are relatively inexpensive, - under $25.

    They have little A/D converters and communicate wirelessly of course. They happen to operate at an interesting voltage level - 2.6 to 3.4 volts. Better, they automatically report their battery voltage. This is basically our cell range.

    I would epoxy dip them and I have no idea on how this will affect their heat dissipation and wireless characteristics. They use very little power.

    During acceleration, they would drop out as the cell voltage dips below 2.6 volts. But they should come back up as you stop.

    I think that might make for an interesting BMS with no real spaghetti wiring.

    That said, you are connecting a device across the terminals of your cell. If it shorts, it will burn. If it burns, the otherwise quite safe LiFePo4 cells make great fuel.

    Jack RIckard

    ReplyDelete
  48. I will mount a battery _monitoring_ system on my battery. A fuse will be added to each connection from the cells with a wire which is not long enough to reach from one terminal to another. The will be a standard 5A automotive type http://www.biltema.dk/da/Bil---MC/Bil-tilbehor/Bil-el/Sikring/Sikringsholder6/

    Martin.

    ReplyDelete
  49. Yay, more sparks were flying. The mount looks good, hopefully it will work well. Thanks for stating the metal thicknesses used.

    I live where they salt the roads in winter. I sand and/or sand blast parts I made, primer and paint them with Rustoleum spray paint. I've had good luck with my parts not rusting.

    ReplyDelete
  50. I like the idea of a RFID chip on each cell, fused of course, reporting temperature and voltage. No one's told me it can't work, though some have suggested noise could be an issue. There may already be existing RFID chips that can do this.

    For rust prevention I've been experimenting with spray on truck bed liner over the paint. No long term results yet but I like the thick coating for chip resistance.

    JRP3

    ReplyDelete
  51. Interesting, I was just looking at the ZigBee products and then read Jacks comment. "Functional Devices Inc." makes a wireless relay (used for building automation and control) They also might be useful for De-Spaghettifing.

    ReplyDelete
  52. On internal resistance dropping:
    Jack says "We have not done a curve of this as it has looked like information of a rather limited use."

    I would use it this way: knowing the SOC from the Ahr counter and measuring the instantaneous battery current, I could then calculate the expected battery pack voltage. Comparing this to the measured voltage would give an early warning of faulty or loose connections.
    Gerhard

    ReplyDelete