Fantasic interview Jordan. So glad you managed to get hold of David. Way too much FUD and misconceptions out there, and way too little information coming out of companies like Tesla. Wonderful to have a true authority on the subject explain in such a clear way. Thanks for adding the hysteresis and other diagrams.
Yes, the diagrams were very helpful. And this interview was excellent as it filled in a lot of holes in my knowledge. As an electrical engineer myself I just eat up every little detail that's given out. But, as this was so far away from my area of expertise (lithography process effects on transistor formation), I think I was ignorant enough to be able to evaluate the content as a novice. And from this perspective I really got a lot out of this (as I do from most of your content). You've really turned me into a battery hobbyist. 😁
Thanks for clarification on LFP charging. Owning an EV with a LFP battery I'm actively trying to preserve the life of it,there's a plethora of confusing information out there! Thanks again 👍
So the initial release of a vehicle/battery model is likely to have a more conservative charge curves. This explains in part why the Tesla software updates have been able to improve range as more real world battery operation data becomes available over time. It also makes the ability to do over the air updates vital for optimal battery management in an electric vehicle.
Actually the opposite happened with the early Model S. Initially they let it charge at like 120kW, but then Tesla replaced lots of batteries out of warranty and lowered the power to under 100kW, maybe around 70kW? Not sure where it's at now.
SOC and balancing in a nutshell: SOC of a cell in general can't be measured in a practical way, it has to be estimated. (to measure it you have to either fully charge, fully discharge or disassemble the cell) Simple SOC estimators only work well for a few simple and highly predictable use cases. You need at least one sensor (usually voltage) to estimate SOC. Most batteries have three types of sensors (voltage, current, temperature). The more sophisticated your battery model (meaning mathematical description of behaviour) is, the better your SOC estimation will be. Aging/degradation changes the battery behaviour, different types of aging change behaviour in different ways. A very sophisticated model can try to predict aging or can try to automatically adapt its parameters to always best decribe the changing battery cell. SOC estimators can and do drift. For chemistries with a steep voltage curve, a good estimator will permanently calibrate itself and only drift a little. (my iPhone has a poor estimator, if I don't charge to 100% it drifts a lot despite having an easy-to-calibrate LiCoO2 battery). For LFP, the only good way to calibrate is a full or nearly full charge. If your SOC estimator has drifted, your car may show 20% SOC when it's actually 5%, but you don't lose capacity. Battery packs need to be balanced, because the cells are never exactly the same. In order to balance, you need to know the SOC of each cell. The best way to make sure you know the SOC is a full or nearly full charge. For this reason, balancing usually only starts near the end of a full charge. Sometimes, you may notice that charging a device from 98% to 100% takes hours. If that happens, it means that balancing is going on. If your battery pack is out of balance, you do lose useable capacity! But you regain it when you balance the pack.
in the cpu world they have something called binning or silicon lottery, can they run any tests or benchmarks on cells so they could also be binned, so to speak
@@jy5960 If you want your battery pack to last a really long time, you should do careful binning before assembling it. For any battery pack you should try to use cells that came off the same production line on the same day, otherwise you will run into balancing trouble.
@@jy5960 Some common properties used for binning are cell internal resistance and capacity. Measuring resistance/impedance is quick. Measuring capacity takes a long time (couple of hours).
@@w0ttheh3llYour original comment starting this thread is absolutely fantastic! Took me a long time to figure out what you basically started succinctly. Great job.
I know this channel is nerdy, but it should have over 1M subscribers. Jordan is a boss. Thanks for giving me a level of battery nerd I could not find anywhere else.
Jordan handled Prof Howey like a pro, teasing out his deep knowledge and blending it together with his own knowledge to produce information that the rest of us as viewers can understand and maybe use. Excellent, and thank you, Jordan.
You have an excellent eye for finding where there may be gaps in understanding for us lay people and then crafting experiences where we can gain knowledge. So thankful for you and what you do!
With batteries/packs I tend to think that: Unbalanced: When a cell (or cell-group) in series is out of step with the series voltage. This can cause cell damage at high currents as the (typically low voltage) unbalanced cell (or cell-group) can be reverse charged and destroyed. WRT state-of-charge for LFP: The pack is best fully discharged and the fully recharged to recalibrate the lower and upper bounds as well as pack capacity. Great content as usual, thank you.
15:30 min: Sodium-ion batteries generally do not have the problem of a wide voltage range. The positive electrode materials made of layered oxide for sodium-ion batteries do have this problem, that is true. But Prussian White analogues, for example, show a very flat voltage profile like LFP. So again, we have to be careful when generalising.
What a great interview! Fascinating! David obviously has a wealth of knowledge but also has that rare talent of being able to explain things in an easy to understand way. Also really enjoyed your interviewing skills Jordan. Quite excellent!
Great interview. I guess one takeaway is that SOC errors and battery pack balancing are independent problems that only share the same partial solution -- occasional 100% charging.
Really good content in this video! if I can add one more answer to the question "what causes rapid change in SoC" , based on my tech. experience in the control automation for industrial BESS projects, the BMS developed from different companies uses also pseudo real time algorithm that adjust SoC value at specific condition typically after cell voltage relaxation around one hour and by looking at the voltage in those conditions the SoC estimation is adjusted consequently; this may happen stepwise but really depends on how elaborated and refined is the algorithm. BTW LFP voltage profile plateau i'is really an hot topic and is causing many headeach to industrial customers that needs to perform these so called SoC Calibration quite often; there are implications on the availability of BESS system and unfortunately this also causes a loss of energy due to BESS components efficiencies and losses. Anyhow, looking and working ontop of those issues , its really stimulating since at all level, from the battery researchers up to who develop control automation to provide BESS secondary functionalities like me, there are multi-layer challenges that makes the job interesting every day. Ciao and Thanks for the great content you always publish in your channel!
Really good interview and questions on highly specific problems. Much appreciated. Though: would it be possible to clean up the audio on prof. Howey? Sometimes it's rather hard to make out what he says, esp. for non native speakers like myself.
@@thelimitingfactor Well, in that case, thank you for trying! To be honest, I didn't know that having a setup like yours makes this much of a differnce. So I'm glad you do.
While a professional microphone and sound-dampening-material-coated studio are ideal, I suspect simply having an inexpensive microphone a few inches from his mouth would have gone a long way to help with the echo for Dr. Howey (though of course, the interviewer has little control over this except for in-person interviews). Using the microphones built into webcams is like using a speakerphone for a conversation, yet I see this even in national news interviews. I've wondered whether it might help if an interviewer offered "if you don't have anything better you intend to use, I'll mail you a $20 set of earbuds as a gift"
Great interview, thank you. I have a question concerning a damage to a battery pack when the cells are out of balance 45:45. In the simple example of lots of cells in series, each unbalanced cell should have different internal resiatance. Thus, when passing a high current through the serie of such cells wouldn't there be an issue with uneven Joule heating? And thus greater degradation and unbalancing?
Great interview, the flat curve is the main reason to why you need to wait util you are over 3.3V before you want to do the balancing on battery, if you try to balance midways it can easily be that they take energy from a 40%SoC cell and move it into a 70%SOC cell and make them even more out of balance. just because a cell at a lot higher SoC can still have a little lower voltage at that flat part of curve. This is very important to think about if you build DIY storage.
Generally the homebuild community tends to balance at 3.45 volts. Newer setups with cells at even resistance can even balance while charging from 3.4 volts on. But then again, it's comparing small cells for ev with diy blocks of 280Ah. Thats not ideal.
Great and very interesting talk, thank you!!! I'm curious if drift of estimated SoC is somewhat predictable. For example: If car not charged to 100% since 4 months, estimate X% of less total capacity and give user a notification - or something along those lines. Part of the answer to the "changing SoC without usage" was a little bit unsatisfying cause what's going on if my car's SoC is increasing and not decreasing? I think it has to do with temperature but I'm not 100% sure.
The hysteresis effect is very interesting to hear about. I assumed cells had an internal resistance that would lead to a voltage drop from the cell internal voltage to the battery system output on discharge, and vice versa when charging. I had seen this hysteresis behavior on my LFP battery system, and was honestly confused by it as it did not seem resistive. Thank you so much!
Really fabulous interview. Probably more useful information in 1 hour than any other battery or EV video I've ever seen. Amazing how Howey immediately grasps the crux of every question and already has the answer formulated in his head. But I have a question related to my own vintage EV. Is it deleterious to an LFP battery to have the voltage dip below 2.7 volts/cell in cold weather when the state of charge is say 30-40% and the only reason for low voltage is a high current load combined with the high internal resistance of the battery at low temperature? My Orion BMS sounds an alarm when this happens because even though it calculates an open circuit voltage, the alarm is based on the instantaneous voltage,
Thanks so much! I was a physics major but I knew very little of this. Some time I would like someone to ask when new battery tech will improve phones, laptops, and power tools.
Excellent video. Apart from charging the Tesla Y to 100% for calibration, I would like to know what level I should drop the charge rate before recharging i.e. 30%, 50% ??
Great stuff...for the most part....I think. I am old, I have my original ears, hence they are old. David's mic/sound, with a touch of that fabulous accent, left me struggling at times to string each syllable together. 😭 Also, don't be afraid to coach your interviews to over explain ( for your level) but explain it in a way that an 80 y/o with the mental capacity of a 6 y/o will understand. I noticed he back-peddled a few times, commenting that he knew you knew this or that. Great interview, enjoyable chemistry, in all regards!
Great interview! One thing that is still not clear to me is whether the cells in a battery can only be balanced at 100% charge. It seems logical that a BMS would keep cell voltages the same throughout the charge cycle rather than only at the top end, especially given that a battery will usually have a buffer so it never charges to a true 100%. I would welcome any thoughts on this.
Listed to it twice now. I listed carefully to hear about REGENTERATIVE BRAKING. I always wondered how efficient regen is, given that the battery temp is managed for discharge when driving (not charging) when slowing down. Wouldn't his be a good application of a super capacitor? Treat it as charge first, discharge first. When decelerating or going down hill, first charge the capacitor with overflow going to the battery. The most likely event after deceleration is acceleration. The capacitor is there to give a burst of power, before the battery takes over. Since capacitors are very fast and efficient, more energy is saved. Correct me please?
Jordan, great job as usual. Would love a historical review of petrol development versus electric to dispel the FUD out there. Electric vehicles have been around basically as long as petrol, but had nothing like the funding petrol had and EVs faced technical limits that no amount of money would have likely fixed way back then. I am old enough to remember leaded gas and the early days of mass produced turbo charged engines that fried because of no real cooling management system so if you shut off the vehicle whole the turbo melted down. People who stared driving in the last few decades do not appreciate or realize how much petrol tech has changed and how mature of a tech it is. Battery tech alone has seen massive improvements in the past twenty years and reminds us battery tech is a much younger or less developed tech. This is why we see massive jumps in EV range and performance every five years or less. Those damning EV tech as a failure and something that will never work seem poorly informed and it would be nice to have a ten minute video that just lays out the facts. No politics, no hype, just laying out where things started and the reality of how EVs really are just hitting their stride and EV tech will have likely be the dominat force in all types of transportion before 2040. I see nothing from petrol or hydrogen tech that will make them the logical choice for transportation once EV tech and infrastructure is built out more. This is not question of if EVs are the near future, as in next ten years, best choice, but when in the next ten years will the EV tech hit the level of performance and cost that will draw most buyers in and gain public trust. Not that you need another project, but your delivery style and well thought out videos will likely be viewed by people on the fence about EVs who might be getting sucked in by the FUD pushed by some on media. It will be shared a ton by those wanting people to know the reality of EVs. Could be a good viral video.
Occasionally charging to as lttle as 3.45 volts and have it start balance cells that reach that voltage (absorb for 30 to 45 minutes) will do for most diy setups with larger lfp cell blocks. It is common practice from what i have been learning in the diy blue cell block community. The idea i get from this video is that EVs with many smaller cells might sooner or later suffer from "runaway" cells that have less internal resistance due to cell degradation , causing different temperature compared to the rest of the pack, causing unpredicted SOC States after charging(increase absorption time?) or high load(prevent high loads to not have fast voltage drops of lower resistance cell, eg faster discharge of single cell) . From this i would say for stationary setups one could prefer oversized cells with the same c rates they will produce more power or to put it in other words, they will not be maxed out so soon as lower Ah cells and preserving longevity is what we all want. Also i oversized the bms. Where i typically use no more than 100Amps at peak usage from my battery the bms is rated 200 amps. The max is for short bursts/start up. I tend to reduce wear and tear. I can run 18Km/h but i seldom do that at the groceries. For what it's worth, after seeing and reading many sources, a jkbms can do a fine job but don't trust manufacturer settings. We need admin rights. For problematic(older) batteries an extra neeey active balancer can be a real help in maintaining cells balanced. Eventually we want flexible Busbars, compression, temperature within limits and replaceable cells. Wether diy or plug and play. We certainly don't want black boxes. My 2 LFP cents. Cheers.
Is it normal that my Felicity 48 v battery voltage fluctuates between 55.6 and 54 after it been charged to 100 % ? Is my inverter starting new charging cycles or its the battery performing balancing ?
I've been working with LFPs and the SoC-voltage behavior is an absolute nightmare. If you give someone an LFP battery and ask them to determine its SoC based on the voltage without charging or discharging it and the voltage is let's say 3.35V, the uncertainty is like 50 SoC percentage points. It's ridiculous and it takes a lot of effort to get around this in a usage pattern where you don't reach the high or low SoC points to reset your coulomb counting.
Edited: There have been some EV owners that had their car die on the road when it said it had 7% SoC left. This is most likely due, in some cases, to cell imbalance being forced to adhere to the minimum voltage protection from that one cell (or parallel module) falling far below the others. Would love to see that covered in your follow-up video you mentioned you were going to make.
@@RCdiy I guess I shouldn’t say “all” lol. It was something that would occur and pop up in the forums and Facebook groups for the older 2012-2018 Tesla Model S’. Once there were data tools available to monitor the cell imbalances it became more clear that there were many users with heavily imbalanced packs that never charged to (or near) 100% and triggered the balancing process.
Aha! That's one battery cell rather than thousands, so it's a different story. No balancing is needed. It does probably need to calibrate, but I wouldn't know how often that needs to happen. Great question
The RC community, especially at the competitive drone racing level, has a lot of experience with state of charge and pushing the limits of what a full cell is. The problem is the cells catch fire.
Interesting conversation. I typically don't have a need and have only charged my EV to 100% once, typically staying within 90%-10% range. It sound like that might not be best practice.
AI summary: Highlights 💡 State of charge in a battery is determined by factors such as voltage, current, thermal management, lithium plating, and different battery chemistries. 💡 Charging curve is created based on maximum power limits, thermal constraints, and limitations to avoid lithium plating. 💡 Discharging of a battery is less of a concern compared to charging, as it does not lead to lithium plating. 💡 Charging a battery to 100% can degrade it more rapidly, especially in certain chemistries, and there’s a trade-off between maximizing battery life and capacity versus the need for accurate state of charge measurement. 💡 Advancements in AI, machine learning, and battery modeling can help with material discovery, optimization, and anomaly detection to improve battery performance and longevity. Not the best summary... the AI missed a few juicy points. The key question everyone probably wants to know is addressed at 37:50 (how should you charge your LFP batteries). As typical with this channel, the journey to the key points is a long, technical one. If you're highly into battery tech details then I'm sure you'll enjoy all the details. Thanks
Probably a good thing Jordan wasn't interviewing Dr. Euan McTurk, another battery electrochemist, because he has a particularly strong Scottish accent which would send US watchers into a right tizz! 😁 But he's certainly worth following, he has his own channel and often talks in EV battery discussions and appears on panels at Fully Charged live shows.
Grid scale batteries - It might be that you drive the batteries very hard because when you replace them, the new batteries will be better and you don't want to miss that.
We don't wait until one cell reaches a voltage limit before we start balancing. That would make no sense. because charging would have to stop to balance. We can balance cell voltage at any state of charge and even while charging.
I know they are very different from cells used in EVs but I've been flying electric model airplanes for years and our chargers balance the cells from the start of the charge. We can usually charge from 20% to 100% in 15 min or so and we discharge back to 20% in 4 to 10 minutes depending on the application so we are very hard on the cells. Packs are generally 6 to 8 cells max.
Great convo - thanks for sharing! But, please, could we stop hyping V2G? If we keep going, people start buying Lightnings and Cybertrucks for them to play Powerwall!! 🤦
If you can get close enough to the state of charge, it's OK. Your gas gauge us not very accurate. When you get to about 4.2V per cell, that's 100% and 3V is 0% Knowing the internal resistance of the cells, lets you correct for the terminal voltage. But once you get a good enough model, you can use it all the time.
If a Tesla LFP vehicle is charged to "100%" is that really 100% or the charge state defined by Tesla as 100%. Ah David just said as much. I guess LFP pretend 100% is much closer to real 100% than for an NMC
Strictly speaking, a "real" 100% doesn't exist. You can take any commercial Lithium Ion battery and charge it higher than the manufacter recommends. That way you will store more energy, but also have more degradation. The 100% and 0% criteria as given by the battery manufacturer are just another set of tradeoffs their engineers decided on. If you look at data sheets, you will see that these criteria sometimes vary from brand to brand and cell to cell, especially for LFP.
Take a look at the data collected by Bjørn Nyland. He's been testing the various Tesla packs for years using "Scan my Tesla" software which gives theoretical and actual states for the system.
100% in an EV is simply the point when it stops charging. if you want it to know what the actual capacity of your car's battery is, you would need to run the battery flat (that is, running the car until the power cuts out) and then find out how much energy it receives when recharging. But that energy reading might not be the actual capacity because it might also include any thermal energy that is generated in the battery and on-board power electronics as the car is charging. So to minimise thermal energy losses, you should recharge with a low-power (2-20 kW) charger.
If it's long range, it's nickel, if it's short range, it's probably LFP. However, from memory there were some changes made in the past year where they were using nickel based cells in short range vehicles. Either way, there's a number of ways to work it out for sure, just Google it.
Go to Controls > Software > Additional Vehicle Information in the car - it should tell you what type of HV battery you have. It could only be LFP if it was 2021-2022 standard range 3 or Y, and even then maybe not.
(With all due respect to the guest) I can't put my finger on if this guy has the actual knowledge and is simply mediocre at communication, or the other way around. He manages to explain foundational engineering work in an impressively vague manner. Instead of naming the degrees of freedom and methodically walking through them step by step, explaining the precise solutions (of which there are many), he convolutes things by mentioning seemingly handy-wavy solutions, such as with the SoC estimation.
Fantasic interview Jordan. So glad you managed to get hold of David. Way too much FUD and misconceptions out there, and way too little information coming out of companies like Tesla. Wonderful to have a true authority on the subject explain in such a clear way. Thanks for adding the hysteresis and other diagrams.
I'm glad you enjoyed it and thanks for the comment!
Yes, the diagrams were very helpful. And this interview was excellent as it filled in a lot of holes in my knowledge. As an electrical engineer myself I just eat up every little detail that's given out. But, as this was so far away from my area of expertise (lithography process effects on transistor formation), I think I was ignorant enough to be able to evaluate the content as a novice. And from this perspective I really got a lot out of this (as I do from most of your content). You've really turned me into a battery hobbyist. 😁
Thanks for clarification on LFP charging. Owning an EV with a LFP battery I'm actively trying to preserve the life of it,there's a plethora of confusing information out there! Thanks again 👍
So the initial release of a vehicle/battery model is likely to have a more conservative charge curves. This explains in part why the Tesla software updates have been able to improve range as more real world battery operation data becomes available over time. It also makes the ability to do over the air updates vital for optimal battery management in an electric vehicle.
Correct 🤠
Actually the opposite happened with the early Model S. Initially they let it charge at like 120kW, but then Tesla replaced lots of batteries out of warranty and lowered the power to under 100kW, maybe around 70kW? Not sure where it's at now.
SOC and balancing in a nutshell:
SOC of a cell in general can't be measured in a practical way, it has to be estimated. (to measure it you have to either fully charge, fully discharge or disassemble the cell)
Simple SOC estimators only work well for a few simple and highly predictable use cases.
You need at least one sensor (usually voltage) to estimate SOC. Most batteries have three types of sensors (voltage, current, temperature). The more sophisticated your battery model (meaning mathematical description of behaviour) is, the better your SOC estimation will be. Aging/degradation changes the battery behaviour, different types of aging change behaviour in different ways. A very sophisticated model can try to predict aging or can try to automatically adapt its parameters to always best decribe the changing battery cell.
SOC estimators can and do drift. For chemistries with a steep voltage curve, a good estimator will permanently calibrate itself and only drift a little. (my iPhone has a poor estimator, if I don't charge to 100% it drifts a lot despite having an easy-to-calibrate LiCoO2 battery).
For LFP, the only good way to calibrate is a full or nearly full charge.
If your SOC estimator has drifted, your car may show 20% SOC when it's actually 5%, but you don't lose capacity.
Battery packs need to be balanced, because the cells are never exactly the same. In order to balance, you need to know the SOC of each cell. The best way to make sure you know the SOC is a full or nearly full charge. For this reason, balancing usually only starts near the end of a full charge. Sometimes, you may notice that charging a device from 98% to 100% takes hours. If that happens, it means that balancing is going on.
If your battery pack is out of balance, you do lose useable capacity! But you regain it when you balance the pack.
Thanks for the well considered comment!
in the cpu world they have something called binning or silicon lottery, can they run any tests or benchmarks on cells so they could also be binned, so to speak
@@jy5960 If you want your battery pack to last a really long time, you should do careful binning before assembling it. For any battery pack you should try to use cells that came off the same production line on the same day, otherwise you will run into balancing trouble.
@@jy5960 Some common properties used for binning are cell internal resistance and capacity. Measuring resistance/impedance is quick. Measuring capacity takes a long time (couple of hours).
@@w0ttheh3llYour original comment starting this thread is absolutely fantastic! Took me a long time to figure out what you basically started succinctly. Great job.
I know this channel is nerdy, but it should have over 1M subscribers. Jordan is a boss. Thanks for giving me a level of battery nerd I could not find anywhere else.
😊 Thanks for the support!
Jordan handled Prof Howey like a pro, teasing out his deep knowledge and blending it together with his own knowledge to produce information that the rest of us as viewers can understand and maybe use. Excellent, and thank you, Jordan.
😊😊😊 Thanks for the kind words!
You have an excellent eye for finding where there may be gaps in understanding for us lay people and then crafting experiences where we can gain knowledge. So thankful for you and what you do!
It's really satisfying to hear that!
I try 😁✊🏼
With batteries/packs I tend to think that:
Unbalanced: When a cell (or cell-group) in series is out of step with the series voltage. This can cause cell damage at high currents as the (typically low voltage) unbalanced cell (or cell-group) can be reverse charged and destroyed.
WRT state-of-charge for LFP: The pack is best fully discharged and the fully recharged to recalibrate the lower and upper bounds as well as pack capacity.
Great content as usual, thank you.
Thanks for the support and the well-considered comment Clark! ✊🏼
15:30 min: Sodium-ion batteries generally do not have the problem of a wide voltage range. The positive electrode materials made of layered oxide for sodium-ion batteries do have this problem, that is true. But Prussian White analogues, for example, show a very flat voltage profile like LFP. So again, we have to be careful when generalising.
Nevertheless, great video
Good point of clarification!
True, sodium ion batteries are not monolithic, just like lithium-ion batteries are not.
Great insights for people like me who would be clueless if it wasn't for your interest, hard work and flair for sharing.
What a great interview! Fascinating! David obviously has a wealth of knowledge but also has that rare talent of being able to explain things in an easy to understand way. Also really enjoyed your interviewing skills Jordan. Quite excellent!
Love the weeds. Now I want to know about what they are now doing to isolate bad cells in parallel and series.
Great video, Jordan. One of your best. Thanks.
Thanks Man! ☺️
Great interview. I guess one takeaway is that SOC errors and battery pack balancing are independent problems that only share the same partial solution -- occasional 100% charging.
Thanks man! I appreciate the support.
Really good content in this video! if I can add one more answer to the question "what causes rapid change in SoC" , based on my tech. experience in the control automation for industrial BESS projects, the BMS developed from different companies uses also pseudo real time algorithm that adjust SoC value at specific condition typically after cell voltage relaxation around one hour and by looking at the voltage in those conditions the SoC estimation is adjusted consequently; this may happen stepwise but really depends on how elaborated and refined is the algorithm.
BTW LFP voltage profile plateau i'is really an hot topic and is causing many headeach to industrial customers that needs to perform these so called SoC Calibration quite often; there are implications on the availability of BESS system and unfortunately this also causes a loss of energy due to BESS components efficiencies and losses.
Anyhow, looking and working ontop of those issues , its really stimulating since at all level, from the battery researchers up to who develop control automation to provide BESS secondary functionalities like me, there are multi-layer challenges that makes the job interesting every day.
Ciao and Thanks for the great content you always publish in your channel!
Thanks a lot for this very informative video
Excellent interview. It certainly improved my understanding of Lithium battery technology.
Thank you, Professor Howey. Fantastic job. Jordan...👒
Brilliant interview Jordan. I would love to see more content like this.
Really good interview and questions on highly specific problems. Much appreciated.
Though: would it be possible to clean up the audio on prof. Howey? Sometimes it's rather hard to make out what he says, esp. for non native speakers like myself.
Glad you enjoyed it!
I tried, but removing the echo introduced more distortion
He's not a TH-camr so he's not set up like I am
@@thelimitingfactor Well, in that case, thank you for trying!
To be honest, I didn't know that having a setup like yours makes this much of a differnce. So I'm glad you do.
While a professional microphone and sound-dampening-material-coated studio are ideal, I suspect simply having an inexpensive microphone a few inches from his mouth would have gone a long way to help with the echo for Dr. Howey (though of course, the interviewer has little control over this except for in-person interviews). Using the microphones built into webcams is like using a speakerphone for a conversation, yet I see this even in national news interviews. I've wondered whether it might help if an interviewer offered "if you don't have anything better you intend to use, I'll mail you a $20 set of earbuds as a gift"
Thanks a lot! Many new things for me, explained on a nice level. And a nice choice on the factsheets/graphs to support the understanding.
Great interview, thank you. I have a question concerning a damage to a battery pack when the cells are out of balance 45:45. In the simple example of lots of cells in series, each unbalanced cell should have different internal resiatance. Thus, when passing a high current through the serie of such cells wouldn't there be an issue with uneven Joule heating? And thus greater degradation and unbalancing?
Jordan you always have incredible guests and this was another banger!
💕
Great interview, the flat curve is the main reason to why you need to wait util you are over 3.3V before you want to do the balancing on battery, if you try to balance midways it can easily be that they take energy from a 40%SoC cell and move it into a 70%SOC cell and make them even more out of balance. just because a cell at a lot higher SoC can still have a little lower voltage at that flat part of curve. This is very important to think about if you build DIY storage.
Generally the homebuild community tends to balance at 3.45 volts.
Newer setups with cells at even resistance can even balance while charging from 3.4 volts on.
But then again, it's comparing small cells for ev with diy blocks of 280Ah. Thats not ideal.
Great interview!
Learned a lot!
Love this so far, about 15 min in, would appreciate you leaving the charts up for just a bit longer or putting links in the comments.
Exceptional episode, thanks!
🤠✊🏼
Great and very interesting talk, thank you!!!
I'm curious if drift of estimated SoC is somewhat predictable. For example: If car not charged to 100% since 4 months, estimate X% of less total capacity and give user a notification - or something along those lines.
Part of the answer to the "changing SoC without usage" was a little bit unsatisfying cause what's going on if my car's SoC is increasing and not decreasing? I think it has to do with temperature but I'm not 100% sure.
The hysteresis effect is very interesting to hear about. I assumed cells had an internal resistance that would lead to a voltage drop from the cell internal voltage to the battery system output on discharge, and vice versa when charging. I had seen this hysteresis behavior on my LFP battery system, and was honestly confused by it as it did not seem resistive. Thank you so much!
Thanks
You're most welcome!
Super informative video… Thanks.
Thanks for the info. It was confusing if the information towards the end of the video related to just LFP or other battery chemistries also.
Really fabulous interview. Probably more useful information in 1 hour than any other battery or EV video I've ever seen. Amazing how Howey immediately grasps the crux of every question and already has the answer formulated in his head. But I have a question related to my own vintage EV. Is it deleterious to an LFP battery to have the voltage dip below 2.7 volts/cell in cold weather when the state of charge is say 30-40% and the only reason for low voltage is a high current load combined with the high internal resistance of the battery at low temperature? My Orion BMS sounds an alarm when this happens because even though it calculates an open circuit voltage, the alarm is based on the instantaneous voltage,
Thanks so much! I was a physics major but I knew very little of this. Some time I would like someone to ask when new battery tech will improve phones, laptops, and power tools.
Superb! Such helpful content for us interested but otherwise uninformed nerds.
Great source of info.
Excellent video. Apart from charging the Tesla Y to 100% for calibration, I would like to know what level I should drop the charge rate before recharging i.e. 30%, 50% ??
Way over my head but I loved it. Thank you!
Terrific interview ... I think I'm getting it !! Thanks
How often should we recalibrate our Tesla NMC?
I'm not sure what you mean by recalibrate
Nickel based chemistries don't usually go out of balance as much as lfp
I go months without charging to 100%
Well done again. Great vid
Great stuff...for the most part....I think. I am old, I have my original ears, hence they are old. David's mic/sound, with a touch of that fabulous accent, left me struggling at times to string each syllable together. 😭 Also, don't be afraid to coach your interviews to over explain ( for your level) but explain it in a way that an 80 y/o with the mental capacity of a 6 y/o will understand. I noticed he back-peddled a few times, commenting that he knew you knew this or that.
Great interview, enjoyable chemistry, in all regards!
Thank you!
Proud owner of a few vintage battery hydrometers😎
😂
Great interview! One thing that is still not clear to me is whether the cells in a battery can only be balanced at 100% charge. It seems logical that a BMS would keep cell voltages the same throughout the charge cycle rather than only at the top end, especially given that a battery will usually have a buffer so it never charges to a true 100%. I would welcome any thoughts on this.
Listed to it twice now. I listed carefully to hear about REGENTERATIVE BRAKING.
I always wondered how efficient regen is, given that the battery temp is managed for discharge when driving (not charging) when slowing down.
Wouldn't his be a good application of a super capacitor? Treat it as charge first, discharge first.
When decelerating or going down hill, first charge the capacitor with overflow going to the battery.
The most likely event after deceleration is acceleration. The capacitor is there to give a burst of power, before the battery takes over.
Since capacitors are very fast and efficient, more energy is saved.
Correct me please?
Just about to settle down to Jordan at 2X speed but first, there's something about the opening theme music that I love.
🤣 I watch my own videos at 2x too
Jordan, great job as usual.
Would love a historical review of petrol development versus electric to dispel the FUD out there. Electric vehicles have been around basically as long as petrol, but had nothing like the funding petrol had and EVs faced technical limits that no amount of money would have likely fixed way back then.
I am old enough to remember leaded gas and the early days of mass produced turbo charged engines that fried because of no real cooling management system so if you shut off the vehicle whole the turbo melted down.
People who stared driving in the last few decades do not appreciate or realize how much petrol tech has changed and how mature of a tech it is.
Battery tech alone has seen massive improvements in the past twenty years and reminds us battery tech is a much younger or less developed tech. This is why we see massive jumps in EV range and performance every five years or less.
Those damning EV tech as a failure and something that will never work seem poorly informed and it would be nice to have a ten minute video that just lays out the facts. No politics, no hype, just laying out where things started and the reality of how EVs really are just hitting their stride and EV tech will have likely be the dominat force in all types of transportion before 2040. I see nothing from petrol or hydrogen tech that will make them the logical choice for transportation once EV tech and infrastructure is built out more. This is not question of if EVs are the near future, as in next ten years, best choice, but when in the next ten years will the
EV tech hit the level of performance and cost that will draw most buyers in and gain public trust.
Not that you need another project, but your delivery style and well thought out videos will likely be viewed by people on the fence about EVs who might be getting sucked in by the FUD pushed by some on media. It will be shared a ton by those wanting people to know the reality of EVs. Could be a good viral video.
Agree, great video.
Occasionally charging to as lttle as 3.45 volts and have it start balance cells that reach that voltage (absorb for 30 to 45 minutes) will do for most diy setups with larger lfp cell blocks.
It is common practice from what i have been learning in the diy blue cell block community.
The idea i get from this video is that EVs with many smaller cells might sooner or later suffer from "runaway" cells that have less internal resistance due to cell degradation , causing different temperature compared to the rest of the pack, causing unpredicted SOC States after charging(increase absorption time?) or high load(prevent high loads to not have fast voltage drops of lower resistance cell, eg faster discharge of single cell) .
From this i would say for stationary setups one could prefer oversized cells with the same c rates they will produce more power or to put it in other words, they will not be maxed out so soon as lower Ah cells and preserving longevity is what we all want.
Also i oversized the bms. Where i typically use no more than 100Amps at peak usage from my battery the bms is rated 200 amps.
The max is for short bursts/start up.
I tend to reduce wear and tear.
I can run 18Km/h but i seldom do that at the groceries.
For what it's worth, after seeing and reading many sources, a jkbms can do a fine job but don't trust manufacturer settings. We need admin rights.
For problematic(older) batteries an extra neeey active balancer can be a real help in maintaining cells balanced.
Eventually we want flexible Busbars, compression, temperature within limits and replaceable cells.
Wether diy or plug and play.
We certainly don't want black boxes. My 2 LFP cents.
Cheers.
Is it normal that my Felicity 48 v battery voltage fluctuates between 55.6 and 54 after it been charged to 100 % ? Is my inverter starting new charging cycles or its the battery performing balancing ?
I've been working with LFPs and the SoC-voltage behavior is an absolute nightmare. If you give someone an LFP battery and ask them to determine its SoC based on the voltage without charging or discharging it and the voltage is let's say 3.35V, the uncertainty is like 50 SoC percentage points. It's ridiculous and it takes a lot of effort to get around this in a usage pattern where you don't reach the high or low SoC points to reset your coulomb counting.
Great anecdote!
Edited: There have been some EV owners that had their car die on the road when it said it had 7% SoC left. This is most likely due, in some cases, to cell imbalance being forced to adhere to the minimum voltage protection from that one cell (or parallel module) falling far below the others. Would love to see that covered in your follow-up video you mentioned you were going to make.
No. Not seen not heard of this happening.
@@RCdiy I guess I shouldn’t say “all” lol. It was something that would occur and pop up in the forums and Facebook groups for the older 2012-2018 Tesla Model S’. Once there were data tools available to monitor the cell imbalances it became more clear that there were many users with heavily imbalanced packs that never charged to (or near) 100% and triggered the balancing process.
What is the limiting factor in implementing V2G then? The car charger and or the software to limit degradation?
More interesting than I thought. 👍
Great video. Im wondeeing if i should charge my phone to 100% once a quarter? It has been limited to 85% (now 80) for the last two years.
Aha! That's one battery cell rather than thousands, so it's a different story.
No balancing is needed. It does probably need to calibrate, but I wouldn't know how often that needs to happen.
Great question
When is a fuel tank full? When the gas gauge touches the F line? When the pump handle stops filling? After spilling a little fuel due to over filling?
The RC community, especially at the competitive drone racing level, has a lot of experience with state of charge and pushing the limits of what a full cell is.
The problem is the cells catch fire.
Interesting conversation. I typically don't have a need and have only charged my EV to 100% once, typically staying within 90%-10% range. It sound like that might not be best practice.
Depend on the chemical of your battery. What you should do is written in the owner manual, you should read it and follow its instructions.
Interest rates as temperature is a brilliant metaphor
AI summary:
Highlights
💡 State of charge in a battery is determined by factors such as voltage, current, thermal management, lithium plating, and different battery chemistries.
💡 Charging curve is created based on maximum power limits, thermal constraints, and limitations to avoid lithium plating.
💡 Discharging of a battery is less of a concern compared to charging, as it does not lead to lithium plating.
💡 Charging a battery to 100% can degrade it more rapidly, especially in certain chemistries, and there’s a trade-off between maximizing battery life and capacity versus the need for accurate state of charge measurement.
💡 Advancements in AI, machine learning, and battery modeling can help with material discovery, optimization, and anomaly detection to improve battery performance and longevity.
Not the best summary... the AI missed a few juicy points. The key question everyone probably wants to know is addressed at 37:50 (how should you charge your LFP batteries).
As typical with this channel, the journey to the key points is a long, technical one. If you're highly into battery tech details then I'm sure you'll enjoy all the details.
Thanks
The reason for large change in soc on parking is due to temp change of the pack. When it cools off it has less energy potential and the bms knows it.
💯
Great video. Unfortunately the audio from the professor made it really hard to understand.
Very helpful thank you
Brilliant, thanks!
Probably a good thing Jordan wasn't interviewing Dr. Euan McTurk, another battery electrochemist, because he has a particularly strong Scottish accent which would send US watchers into a right tizz! 😁 But he's certainly worth following, he has his own channel and often talks in EV battery discussions and appears on panels at Fully Charged live shows.
Love the beard! Keep it mate!
Thank you!
Grid scale batteries - It might be that you drive the batteries very hard because when you replace them, the new batteries will be better and you don't want to miss that.
We don't wait until one cell reaches a voltage limit before we start balancing. That would make no sense. because charging would have to stop to balance.
We can balance cell voltage at any state of charge and even while charging.
There's more to it than that
That's why there's procedures to force the cells to balance
I know they are very different from cells used in EVs but I've been flying electric model airplanes for years and our chargers balance the cells from the start of the charge. We can usually charge from 20% to 100% in 15 min or so and we discharge back to 20% in 4 to 10 minutes depending on the application so we are very hard on the cells. Packs are generally 6 to 8 cells max.
How does Porche pull off the bananas curves in the new Taycan?
They designed it to do that, but every design decision has trade-offs
But what were the trade-offs? Was it cycle life?
I think Jordan is channel Vincen Van Gogh. Looks good IMHO
I set my Ora Good Cat charging to 90% but every five or so charges the car charges to 100%.
Glad this interview is in British accent, makes it sound more scientific 😊
🤣💯
Great convo - thanks for sharing!
But, please, could we stop hyping V2G? If we keep going, people start buying Lightnings and Cybertrucks for them to play Powerwall!! 🤦
If you can get close enough to the state of charge, it's OK. Your gas gauge us not very accurate.
When you get to about 4.2V per cell, that's 100% and 3V is 0%
Knowing the internal resistance of the cells, lets you correct for the terminal voltage.
But once you get a good enough model, you can use it all the time.
You could narrate audiobooks.
I might do that in my retirement, lol
ANOTHER "Two (+) Listen masterpiece"!!!!
.
😉
I should add, PLENTY of "FUD rebuttal" material! 👍
❤️😊
If a Tesla LFP vehicle is charged to "100%" is that really 100% or the charge state defined by Tesla as 100%. Ah David just said as much. I guess LFP pretend 100% is much closer to real 100% than for an NMC
🎯
Strictly speaking, a "real" 100% doesn't exist. You can take any commercial Lithium Ion battery and charge it higher than the manufacter recommends. That way you will store more energy, but also have more degradation.
The 100% and 0% criteria as given by the battery manufacturer are just another set of tradeoffs their engineers decided on. If you look at data sheets, you will see that these criteria sometimes vary from brand to brand and cell to cell, especially for LFP.
Take a look at the data collected by Bjørn Nyland.
He's been testing the various Tesla packs for years using "Scan my Tesla" software which gives theoretical and actual states for the system.
100% in an EV is simply the point when it stops charging.
if you want it to know what the actual capacity of your car's battery is, you would need to run the battery flat (that is, running the car until the power cuts out) and then find out how much energy it receives when recharging. But that energy reading might not be the actual capacity because it might also include any thermal energy that is generated in the battery and on-board power electronics as the car is charging. So to minimise thermal energy losses, you should recharge with a low-power (2-20 kW) charger.
🙋♂️ DON’T KNOW WHAT KIND OF BATTERIES I HAVE IN MY MY …MADE IN FREMONT IN JUNE 2023 🤷♂️💚💚💚
If it's long range, it's nickel, if it's short range, it's probably LFP.
However, from memory there were some changes made in the past year where they were using nickel based cells in short range vehicles.
Either way, there's a number of ways to work it out for sure, just Google it.
I would imagine the data is in the VIN Number.
Just did a quick search
"Tesla VIN decoder".
Plenty of options.
Go to Controls > Software > Additional Vehicle Information in the car - it should tell you what type of HV battery you have. It could only be LFP if it was 2021-2022 standard range 3 or Y, and even then maybe not.
@@thelimitingfactor THANKS JORDAN,LONG RANGE,AWD 👍
@@newsfromthefrunk 🙋♂️🤗THANKS I’LL CHECK IT OUT 👍
That helped me a lot
Glad to hear it! 🙌🤠
Like cocaine for battery knowledge junkies…
(With all due respect to the guest)
I can't put my finger on if this guy has the actual knowledge and is simply mediocre at communication, or the other way around.
He manages to explain foundational engineering work in an impressively vague manner. Instead of naming the degrees of freedom and methodically walking through them step by step, explaining the precise solutions (of which there are many), he convolutes things by mentioning seemingly handy-wavy solutions, such as with the SoC estimation.
To turn the hand-wavy stuff into detail would take a lot more than one hour!