Chriull Posted September 29, 2021 Share Posted September 29, 2021 5 minutes ago, mrelwood said: Why don’t we discuss the technical matters in it’s own topic? Link below. These won’t be answered in a short single posting. Splitted these posts over to here - unfortionately it is sorted in by date... On 9/27/2021 at 8:13 PM, EUC Endurist said: would be interested in what happens when the cells are fully outbalanced and the charger is still plugged it? would there be some kind of bypass? If cells are fully outbalanced they'll stay at exactly one 24th of the charger voltage (in case of a 100.8V wheel). If everything is adjusted perfectly this would be 4.2V per cell, just before bleeding resistors are "applied". So they'll trickle charge forever and build up some platings/dendrits over time. As nothing is 100% perfect the charge voltage will be at bit above or below this threshold. If all resistors are applied trickle charging current will be a bit lower for the cells as some current is bypassed. As all resistor voltage threshold are not 100% equal there also could be the case that some resistors are applied and some not. This could lead to some very slight balance by different trickle charge currents. But should not really change too much... 1 Quote Link to comment Share on other sites More sharing options...
mrelwood Posted September 29, 2021 Author Share Posted September 29, 2021 On 9/27/2021 at 9:13 PM, EUC Endurist said: would be interested in what happens when the cells are fully outbalanced and the charger is still plugged it? would there be some kind of bypass? I’m not sure what you mean by “outbalanced”. Do you mean fully balanced out, or fully out of balance? Quote Link to comment Share on other sites More sharing options...
Popular Post Jason McNeil Posted September 30, 2021 Popular Post Share Posted September 30, 2021 (edited) It's a good video, but the explanation of the causes of cell/pack failure is to simplisitc & far from complete. Over the past 5 years, we've sold many thousands of Wheels, of the couple hundreds of pack failures that we have dealt with, >90% had been using the stock charger & following the standard manufacturing practice of charging to 100%. The strongest correlation evidence of premature pack failure is the cell-brand, followed by BMS fault & then shock/trama events—where single cells break from the nickel strips in the series. Another consideration, that is somewhat unique to the electric unicycle, is that during hard braking, the amount of regenerative transient surge current will be far & beyond that of a 'normal' lab cycle longevity test. Partial charging will give your Wheel some buffer space for packs to absorb the converted kinetic energy without over charging beyond spec. This effect is also offset by the number of parallel packs in the Wheel & is not linnear in nature, i.e. it is logarithmic over a certain threshold. A host of other factors come into play as well: ambient temperature measured against the individual load factor of the cells, a cell's internal resistance, charge rate, depth of discharge frequency, etc. In response to Mrelwood's concerns last year, I had instructed our charger supplier to implement a CV phase into the partial charging modes to potentially minimize cell drift, & a Wh counter into the LCD output for Customers to be able to monitor the pack health/capacity. I agree that this topic needs further research to try to address the questions of: 1) What is the degree of cell imbalance if the charge cycle is terminated during the CC/CV transition? 2) Does tapering down the current in partial 80/90% modes reduce drift? In theory it should, but it ought to be demonstrated 3) Collect pack samples from both a control group (100%) & frequent partial charge usage to see if there's any descernible longevity effects, measure the IR of each cell 4) Future BMSs ought to have an interface to view individual cells & cell-groups for monitoring those that are out of sync—this is something we're working on with our battery partner LiTech on the Begodes 5) Evaluation of active-balancing so that cells remain balanced over the whole charge cycle In the larger context of targets for mfgrs to improve, battery issues consistute a minor rounding error to controller & motors issues. Edited September 30, 2021 by Jason McNeil 19 Quote Link to comment Share on other sites More sharing options...
meepmeepmayer Posted September 30, 2021 Share Posted September 30, 2021 @Jason McNeilWhat are the chances of convincing the manufacturers to use a BMS that always balances the cells (at least when the wheel is off) regardless of charge state? Assuming that would make sense, but balancing only in some states seems like an arbitrary and unnecessary technical holdover. 3 Quote Link to comment Share on other sites More sharing options...
Popular Post Jason McNeil Posted September 30, 2021 Popular Post Share Posted September 30, 2021 3 minutes ago, meepmeepmayer said: What are the chances of convincing the manufacturers to use a BMS that always balances the cells (at least when the wheel is off) regardless of charge state? Assuming that would make sense, but balancing only in some states seems like an arbitrary and unnecessary technical holdover. With LiTech, almost certain, but these projects take time, especially if it's something that's relatively novel. 6 Quote Link to comment Share on other sites More sharing options...
Marty Backe Posted September 30, 2021 Share Posted September 30, 2021 On 9/24/2021 at 12:26 AM, mrelwood said: I finally got this video made. I hope that the crucially important information spreads better this way. Late to the game, but I really enjoyed your video. I've been charging to 100% for years on all my wheels, and I typically don't get around to unplugging the charger for probably 8+ hours after the charging is complete. I am guilty of leaving all my wheels at 100% and some of them will sit fully charged for over a month. I do notice many of the newer wheels are consuming some amount of power even when off. Many of the control boards have at least one tiny LED that will be flashing all the time. For instance, a Sherman that's left alone for over a month may be down ~1-volt. So I wonder if the "store at 40-60%" is as important now??? 3 Quote Link to comment Share on other sites More sharing options...
Popular Post RockyTop Posted September 30, 2021 Popular Post Share Posted September 30, 2021 I am old school knowledge on batteries, yet I don’t see why this would not apply here. When you run batteries in series they are only as strong as the weakest cell. From day one, one of the cells is the weakest link. That cell is then damaged by the good health of the other batteries. The weak cell is over drained and charged every cycle. The BMS does what it can to minimize the damage. The point is that any abuse will be directed to the weakest cell. The damage is compounded. This also backs up Jason’s statement on poor quality batteries as being a major factor in failure. 8 Quote Link to comment Share on other sites More sharing options...
Popular Post RagingGrandpa Posted September 30, 2021 Popular Post Share Posted September 30, 2021 Cells are a chemical sandwich that cannot be manufactured in an identical matter. The same way that ice cream sandwiches cannot be manufactured with identical bubbles and overflow Variation in capacity and resistance are normal, and our designs must tolerate it. -- There are always (slightly) weak cells -- 6 minutes ago, RockyTop said: The weak cell is over drained and charged every cycle. Not really. As an example: Assume one 'weak' cell in a series has 10% less stored energy than the rest. Assume we begin with all cells 'balanced': every cell starts from 4.20 V. During discharge, the EUC will bring the healthy cells down to 3.15V. The 'weak' cell will become discharged to about 3.0V. It's lower than the others, but still a perfectly healthy and non-damaging voltage. Then while recharging, the 'weak' cell will return to 4.2V almost perfectly. Resistive balancing will resolve any small offset. --> This is the normal design approach that EUC's use to tolerate cell-to-cell variation. And bear in mind, the thresholds for 'cell damage' are very far outside the normal voltage ranges seen in EUC applications. Things have to go very wrong to result in a fire. Cells do not suddenly explode at 4.3V 3 hours ago, Jason McNeil said: The strongest correlation evidence of premature pack failure is: the cell-brand, followed by BMS fault & then shock/trama events—where single cells break from the nickel strips in the series. This seems quite plausible and applies to any large Li-Ion pack, EUC or otherwise. #1 "cell defects" imply catastrophic developments of internal shorts, which become 'heaters' and cannot be stopped by BMS or fusing. #2 "BMS faults" imply the normal design to tolerate cell-to-cell variation (discussed above) cannot function properly, and so cells exit their safe voltage ranges after many cycles of use. (#3 "physical damage" hopefully requires no further explanation) Love it when suppliers participate directly in community conversation, thx for posting @Jason McNeil 8 Quote Link to comment Share on other sites More sharing options...
RockyTop Posted September 30, 2021 Share Posted September 30, 2021 13 minutes ago, RagingGrandpa said: The 'weak' cell will become discharged to about 3.0V. It's lower than the others, but still a perfectly healthy and non-damaging voltage. So using one cell, the weakest cell more than the others, draining just a bit further is NOT going to weaken it even quicker than the good healthy cells? 1 Quote Link to comment Share on other sites More sharing options...
RagingGrandpa Posted September 30, 2021 Share Posted September 30, 2021 51 minutes ago, RockyTop said: using the weakest cell more than the others, draining it just a bit further, is going to weaken it even quicker than the good healthy cells ^ this, yes, but only slightly It means: for every 100 full-cycles of the 'weak' cell, there are only 90 equivalent full-cycles of the strong cells. So the weak cell ages faster. This is a small effect, because problems resulting from this usually don't matter until we're nearing 500 cycles. Since EUC's are not a high-cycle-count application, it's rare to find an example where this normal aging process makes a pack unusable. (500 cycles is 12,500 miles ridden for Tesla, and 33,000 miles ridden for Sherman!) So when there's a failure, it's more likely that one of Jason's 1-2-3 is to blame 2 Quote Link to comment Share on other sites More sharing options...
Tawpie Posted September 30, 2021 Share Posted September 30, 2021 Would your 'weak' cell be in danger of getting pulled below the 2.5V damage threshold during large current draw events? Or do the individual cells have under voltage protection (in the olden days of primary lithium cells, each cell had over and under voltage protection and some had fused PTC—I don't know about individual LiIon cells). Quote Link to comment Share on other sites More sharing options...
Planemo Posted September 30, 2021 Share Posted September 30, 2021 17 minutes ago, Tawpie said: Or do the individual cells have under voltage protection I don't believe any of our EUC's use packs with protected cells. Presumably the idea being that when used with a BMS, you shouldn't need them. Of course, the problem with this is that we are up to so many cells in our EUC's now, with silly amounts of them being paralleled, that each string monitored by the BMS has too many cells hanging off it for my liking. 1 Quote Link to comment Share on other sites More sharing options...
Popular Post mrelwood Posted October 1, 2021 Author Popular Post Share Posted October 1, 2021 Thank you Jason for taking part in the conversation! This backs up everything positive I said about eWheels even further. 13 hours ago, Jason McNeil said: It's a good video, but the explanation of the causes of cell/pack failure is to simplisitc & far from complete. I’m absolutely sure that it is. But it’s the first video talking about the issue. I definitely wish to see people more knowledgeable about the tech or even just the business side to also make videos on the subject. 13 hours ago, Jason McNeil said: Over the past 5 years, we've sold many thousands of Wheels, of the couple hundreds of pack failures that we have dealt with And herein lies the very reason we see so drastically different sides of any EUC issues. The battery warranty is generally 6 months (haven’t checked yours). If a pack fails to charge above 70% after 9 or 12 months, what does the owner do? Based on what I’ve read and what I’d do myself, contacting the seller at that point is already way down the list. More attractive wheels come available every year, and I’ve seen many owners write that they’ll rather just ride a few months with the failed battery and then buy a new wheel, than to ship the wheel back to the seller and not ride at all for a few weeks. Once the new wheel arrives, the one with the failed pack is simply tossed, or fraudulently or unknowingly sold 2nd hand. Therefore it stands to reason that the battery issues you have dealt with are mainly actual manufacturing faults, and the ones we read about here at the forum (or face locally) are drastically different. Neither of which presents the correct ratio of issues. Many times I’ve even tried to convince a rider to even utilize the warranty in the first place when their wheels were having clear manufacturing issues. 13 hours ago, Jason McNeil said: In response to Mrelwood's concerns last year, I had instructed our charger supplier to implement a CV phase into the partial charging modes to potentially minimize cell drift, & a Wh counter into the LCD output for Customers to be able to monitor the pack health/capacity. I am stoked to hear this! Unfortunately it might not do much to help the cause, since it won’t engage the balancing function of the BMS. But at least it could help the user to monitor the capacity. 13 hours ago, Jason McNeil said: I agree that this topic needs further research to try to address the questions of: 1) What is the degree of cell imbalance if the charge cycle is terminated during the CC/CV transition? Tests like this might require a budget and a sample size that is simply not possible for EUCs for several years. And btw, charging to 80% still leaves a good chunk of the CC phase to go before reaching the CC/CV transition. 13 hours ago, Jason McNeil said: 2) Does tapering down the current in partial 80/90% modes reduce drift? In theory it should, but it ought to be demonstrated I agree, it should be demonstrated. Since the charging current is the same for all cell groups, I’m not sure why there would be a meaningful difference in charging the low and high voltage groups. I recall seeing charts of the charging process to show a perfectly linear increase in battery voltage during the whole CC phase. 13 hours ago, Jason McNeil said: 3) Collect pack samples from both a control group (100%) & frequent partial charge usage to see if there's any descernible longevity effects, measure the IR of each cell The thing is, the cells themselves only get damaged if the imbalance is drastic. Case #1: A variance of 0.5-1V that should still be able to be salvaged by either pumping the BMS feature (100%-90%-100% looping) or manually charging the individual low groups. Case #2: The cell groups are already down to 0V like they were on my 16S, way beyond saving. If the pack is only charged to 80%, the user has absolutely no way of knowing whether or not they’re having either of the above two cases. In the test you suggest, the case #1 probably wouldn’t show any actual cell aging or damage, although the pack is well on its way to the case #2, fast. 13 hours ago, Jason McNeil said: 4) Future BMSs ought to have an interface to view individual cells & cell-groups This would be a HUGE jump forward! Experienced users would easily be able to keep their packs in good health for the lifetime of their EUCs. 13 hours ago, Jason McNeil said: 5) Evaluation of active-balancing so that cells remain balanced over the whole charge cycle And THIS would completely obliviate the battery related issues we have seen time and time again, even for the ones who don’t know what to do with the cell group monitoring in the step 4)! This would be a PERFECT solution. Even if only on Begodes bought from eWheels at first, it would surely drive all manufacturers to catch up with the progress. 13 hours ago, Jason McNeil said: In the larger context of targets for mfgrs to improve, battery issues consistute a minor rounding error to controller & motors issues. I do agree that especially during the past year, burnt controllers have kind of made a comeback. I definitely agree that they should be improved first, as they are an imminent safety issue to the rider. But as I pointed out above, eWheels’ rounding error is a much much larger issue in the whole community. 5 hours ago, Planemo said: I don't believe any of our EUC's use packs with protected cells. Presumably the idea being that when used with a BMS, you shouldn't need them. I think a bigger factor might be that the transient current peaks are probably too high for the protection to handle. 5 Quote Link to comment Share on other sites More sharing options...
Dogmaticjoe Posted October 1, 2021 Share Posted October 1, 2021 so if I charge my wheel (the KS18XL) to 100% with my stock charger how long after it hits 100% should I leave it plugged in to let it balance? Quote Link to comment Share on other sites More sharing options...
Chriull Posted October 1, 2021 Share Posted October 1, 2021 1 hour ago, Dogmaticjoe said: so if I charge my wheel (the KS18XL) to 100% with my stock charger how long after it hits 100% should I leave it plugged in to let it balance? TLDR: once the charger shows the green light. If the KS18XL is (still?) using the NCR18650GA panasonic recommends: "(5) Termination of Charging The system will determine that the battery is full by detecting the charge current. Stop charging once the current has reached 0.1 It to 0.07 It. Note that there will be some degree of variation for each individual battery. (6) Charge Timer A total charge timer and a charge completion timer should be included" in https://industrial.panasonic.com/cdbs/www-data/pdf/ACA4000/ACA4000PE4.pdf For this specific cell they use/recommend for a full charge in their cell datasheet in one diagram 67mA as charge current threshold, 100mA in another. With the KS18XL having 6 cells in parallel this translates to 6 x 67...100mA~400...600mA. It's not secure how the charge current threshold of the individual chargers to show the green light are adjusted, but normally, from what us reported here this green led has chances to show at some a bit lower current thresholds. The effects of charging downto 600, 400 or even 250mA should imho be minimal, but it is interesting and notable that panasonic in their dazasheet charges downto 67mA for the "performance graph", but only downto 100mA for their "cycle life graph"! 1 Quote Link to comment Share on other sites More sharing options...
div Posted October 1, 2021 Share Posted October 1, 2021 (edited) I don’t have the 18xl at hand but in the case of the 16S (it’s about the same) the charger goes green before the voltage stabilises; so in the lack of current measuring I like going for the second. (Feels more right, no scientific backing) (The selection point on the graph is when the charger goes green, at 67,43V, it stabilises at 67,59V) Ultimately people don’t stand by their charger to see the second it goes green, so as a general rule (considering most won’t notice it until a bit later) it’s probably fine. Edit: Chriull is certainly correct here, only as he mentions we’re not sure of the threshold of the individual charger to go green. Hence me estimating stable voltage as end of CV. Edited October 1, 2021 by div 3 Quote Link to comment Share on other sites More sharing options...
Chriull Posted October 1, 2021 Share Posted October 1, 2021 1 hour ago, div said: (The selection point on the graph is when the charger goes green, at 67,43V, it stabilises at 67,59V) Seems your green light really comes a bit early. For comparison the charge graphs (with EUC Worlds "estimated" charge current via the HS110 plug) are from my KS16S https://forum.electricunicycle.org/topic/18317-battery-faq/?do=findComment&comment=307641 . Our batteries very likely have some different "internal state" by now, but in my graphs this last ~0,15 happen from ~0,5..0,7A on... ... and sounds plausible. Perfect internal 16s4p resistance (48mOhm per cell) would be about 0.2 Ohm. 0.7A * 0.2Ohm = 0.14V 2 Quote Link to comment Share on other sites More sharing options...
Popular Post Chriull Posted October 1, 2021 Popular Post Share Posted October 1, 2021 17 hours ago, RagingGrandpa said: Not really. As an example: Assume one 'weak' cell in a series has 10% less stored energy than the rest. Assume we begin with all cells 'balanced': every cell starts from 4.20 V. During discharge, the EUC will bring the healthy cells down to 3.15V. The 'weak' cell will become discharged to about 3.0V. It's lower than the others, but still a perfectly healthy and non-damaging voltage. Li Ion Cells drop voltage very fast in the last state of charge. The Discharge rate characteristics of NCR18650GA show for 2A discharge current: ~3000mAh used until 3.15V So such a cell with 10% less capacity should be at a voltage were a "normal 100%" cell used up ~3000mAh/0,9~3333mAh which would be 2,5V or even below - it's more or less the end of the graph... I know numbers like this 10% are just some assumptions but the nonlinearity in cell discharge characteristics could be a degradation reason for a battery pack. 17 hours ago, RagingGrandpa said: Then while recharging, the 'weak' cell will return to 4.2V almost perfectly. Resistive balancing will resolve any small offset. Could be that this, in combination with 80% charging is the reson for imbalanced packs. Burdening a wheel till empty - the one weak cell gets a little bit degraded and just reaches a little bit less voltage as the others. Since no balancing happens to remove this small offset it will get even lower in voltage the next time the pack is emptied, degraded bit more, offset gets a bit higher and is not balanced again. Sounds like a positive feedback loop... ;( So with if one does not use his whole battery capacity for ones rides it's better to recharge before every ride than empty the battery fully and then recharge... 5 Quote Link to comment Share on other sites More sharing options...
mrelwood Posted October 1, 2021 Author Share Posted October 1, 2021 2 hours ago, Chriull said: If the KS18XL is (still?) using the NCR18650GA panasonic recommends: Panasonic doesn’t have recommendations for large packs in series with a passive top balancing systems though… I think that balancing even just a bit longer easily overrides the microscopic difference that the charging current cutoff can ever bring. 1 Quote Link to comment Share on other sites More sharing options...
Unventor Posted October 1, 2021 Share Posted October 1, 2021 I haven't been bothered with cell degradation due to I bought so many new wheels because the development went so fast. So far I have yet to see a problem. Well my V11 had a BMS state failure on one battery pack but beside that neither of my wheels showed any signs of degradation. Now going forward I suspect the upgrade cycle at least in my part will be much slower. I have the speed and in general the range I need. And the comfort of the V11 suspension too. That is not to say something better might make me want a new wheel. But having 2x V11 with winter/all season tire on I think they will outlast me for some time. And I can choose to use one for spare parts if any problems occur in one of them. But I am fairly sure most do not have a wheel similar just for backup. I chose this as I know how hard spare parts can be to come by these times. And I predict it will take a year maybe more before this get stabil. So far I charge as long my wheel get below 70% in battery. And I use EUC.WORLD to monitor the charging though a TP-link HS110 smart plug. It might degrade the battery over tid. On the other hand I am more worried of cells getting out of balance. I am not a battery or charging expert at all. What I am doing could be all wrong. What I do think is very very interesting is the @Jason McNeil joined in. I think he will hold more data in view perspective what causes problems. But like @mrelwood said one thing is warranty resolved issues another is what happens outside warranty. That leaves a blank data sheet in most cases compared to what I know from my line of work. As battery pack gets larger the impact of degradation might show more as it is less prone for people to buy a new wheel if the old one is sufficient if in working order. Also charge cycles might show different because of different charging behaviours. 2 Quote Link to comment Share on other sites More sharing options...
RockyTop Posted October 1, 2021 Share Posted October 1, 2021 I have watched @mrelwood rack up excessive miles on many new wheels in just a year or two. ( 16s > MSX > v11 ) The DarknessBot tends to set him near the top of the mileage list. So I am going to trust his observations. As a hobby rider, weekend explorer, I will never put the miles on a wheel like the commuters. I have always assumed that my batteries are going to age out before they wear out. I thought that max battery preservation included a cycle of 4:1 Four times at 80% then a full charge. Is this not true? 2 Quote Link to comment Share on other sites More sharing options...
Popular Post mrelwood Posted October 1, 2021 Author Popular Post Share Posted October 1, 2021 9 minutes ago, RockyTop said: The DarknessBot tends to set him near the top of the mileage list. Stalker. 9 minutes ago, RockyTop said: I thought that max battery preservation included a cycle of 4:1 Four times at 80% then a full charge. Is this not true? Could be. Or it could be 8:1. Or 1:1. Varies by battery manufacturer, cell matching, and usage. Until we get individual cell group voltage monitoring, we have no way of knowing which is the correct ratio for your specific battery. What we do know is that even with a 1:1 ratio you won’t lose any range until something like 20000km (1800Wh). We also know that if you don’t balance as often as your battery secretly requires, your range and power will plummet, and the battery becomes a fire hazard. So, anything other than 1:1 means huge risk with zero gain. 6 Quote Link to comment Share on other sites More sharing options...
RockyTop Posted October 1, 2021 Share Posted October 1, 2021 1 hour ago, mrelwood said: Stalker. I like your voice. Hey it's not me. Blame Darknessbot. You are a Darknessbot celebrity. RockyTop 136 miles this week. rank 995. So who is in the top 10? mr.elwood. Quote Link to comment Share on other sites More sharing options...
mrelwood Posted October 1, 2021 Author Share Posted October 1, 2021 1 hour ago, RockyTop said: I like your voice. Someone said that I sound like Arnold Schwarzenegger! (But thanks! I made it myself!) 1 hour ago, RockyTop said: You are a Darknessbot celebrity. RockyTop 136 miles this week. rank 995. So who is in the top 10? mr.elwood. It’s actually interesting, since I haven’t used DarknessBot for a year or two. I have connected it a few times though, so maybe it catches up for the long term charts. But It has been maybe a month since I last connected with it. 1 Quote Link to comment Share on other sites More sharing options...
EUC Endurist Posted October 3, 2021 Share Posted October 3, 2021 On 9/29/2021 at 10:06 PM, Chriull said: Splitted these posts over to here - unfortionately it is sorted in by date... If cells are fully outbalanced they'll stay at exactly one 24th of the charger voltage (in case of a 100.8V wheel). If everything is adjusted perfectly this would be 4.2V per cell, just before bleeding resistors are "applied". So they'll trickle charge forever and build up some platings/dendrits over time. As nothing is 100% perfect the charge voltage will be at bit above or below this threshold. If all resistors are applied trickle charging current will be a bit lower for the cells as some current is bypassed. As all resistor voltage threshold are not 100% equal there also could be the case that some resistors are applied and some not. This could lead to some very slight balance by different trickle charge currents. But should not really change too much... thanks really much. so ill opt for 1-2hours green led and then cut the charger. you helped me many times chriull 1 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.