Anybody posted this video here before? A lot of enlightening things about LiFePo4 batterries.
http://www.ri.cmu.edu/video_view.html?video_id=60&menu_id=387
http://www.ri.cmu.edu/video_view.html?video_id=60&menu_id=387
Lecture on LifePo4 cells
1 - 20 of 63 Posts
Joined
·
1,379 Posts
It's a great video, highly recommended if you want to learn about how a lithium ion battery works, plus some practical metrics for judging a battery.
Joined
·
590 Posts
Watched it this afternoon. First-class information from someone who clearly knows his subject. I wish I'd been able to watch it a couple of years ago. He explains how lithium phosphate batteries are put together and why they behave the way they do. The information on upper and lower voltage and temperature limits is reassuring, and surprising in some ways.
I see Jack Rickard is claiming this video supports his arguments against using a BMS – I guess he pressed the stop button as soon as he'd heard what he wanted to hear. I'd recommend watching the whole thing, but if you're in a hurry the relevant part is from 65 minutes on.
I see Jack Rickard is claiming this video supports his arguments against using a BMS – I guess he pressed the stop button as soon as he'd heard what he wanted to hear. I'd recommend watching the whole thing, but if you're in a hurry the relevant part is from 65 minutes on.
Do you mean that this video does not support Jacks claims ?
I can pretty much hear Jay saing that it is a waste of money to put single cell management in a battery pack, and that the most cost effective way of doing pack management is String management of well matched cells that has been taken to the same state of charge, that being either top, bottom or middle (if you can find a good way of doing that).
Regards
/Per Eklund
Joined
·
590 Posts
Hi Per,
He supports Jack's argument to some extent, but that's not the whole story. Put simply, the message I got is that cell-level management doesn't make economic sense to manufacturers of EVs, so they rely instead on closely matching cells, so that they don't drift out of balance. But he also says that cell-level management is the correct engineering approach.
As DIY builders, most of us don't have the luxury of being able to pick and choose cells that are closely matched in capacity and resistance, so we need to manage cells individually, or at least monitor cell voltages and occasionally chivvy the stragglers back into line.
I'm not trying to start another pro/anti BMS argument, just setting things straight. I'd personally love to eliminate the complications of a BMS, and I'm sure that a few years down the line, as production ramps up and quality control improves, they won't be necessary.
He supports Jack's argument to some extent, but that's not the whole story. Put simply, the message I got is that cell-level management doesn't make economic sense to manufacturers of EVs, so they rely instead on closely matching cells, so that they don't drift out of balance. But he also says that cell-level management is the correct engineering approach.
As DIY builders, most of us don't have the luxury of being able to pick and choose cells that are closely matched in capacity and resistance, so we need to manage cells individually, or at least monitor cell voltages and occasionally chivvy the stragglers back into line.
I'm not trying to start another pro/anti BMS argument, just setting things straight. I'd personally love to eliminate the complications of a BMS, and I'm sure that a few years down the line, as production ramps up and quality control improves, they won't be necessary.
Joined
·
2,504 Posts
Took me a couple of cracks to get through it... but REALLY GOOD.... especially the last third that was more about application... Thank you!
Joined
·
2,123 Posts
It all depends on terminology used and how we understand this terminology. He was way too brief on the subject and the audience was too mesmerized to ask the right questions
Here is how I understood it. Single cell management is the ability to remove individual cell from the circuit and/or reroute energy around it, such that single cell failure does not stop the whole pack. This requires power electronics ( like IGBT modules ) for each cell, to be able to route energy around bad/empty cell. Some people proposed such BMS schemes, even on this forum, but for anyone familiar with electronics involved its 100% clear that such solution is not cost effective. I think this is what Jay was referring to, he actually mentioned bypassing a cell.
String management, the way I understood him, is monitoring cell level voltages and taking action upon entire pack, i.e. stopping charge/discharge when any cell goes outside of norm. He actually mentioned cell level monitoring using simple cheap circuits. This type of management is precicely what most affordable BMS systems do. Its a string management, based on cell level monitoring.
Sure, if you have the luxury of hand picking cells based on AH and IR match from a large pile of tested cells, and also size the pack such that you never reach top or bottom, then BMS requirement is reduced. However, us DIY folks don't have such luxury.
I'm expecting some rocks being thrown my way because I am the evil BMS pusher
, but I strongly believe there are real benefits in a simple affordable BMS, despite what EVTV crowd is lead to believe by their fearless leader BTW, the video was awesome, by far the best info about Lithium batteries in a simple easy to understand form, although it probably helps to have some background in electrochemistry.
I also liked how simple in 1-2 sentences he explained how to balance a new pack before first use. There are pages and pages of forum threads here discussing complicated ways of doing the same thing that can be done in few simple steps. Parallel all cells, charge or discharge to get them outside of flat area, let them sit for a while, then connect in series, done, end of story.
Joined
·
2,504 Posts
I understood pretty much the same thing as Dimitri. I think he was talking about bypassing a cell when referring to the cost and complexity of a BMS. That said, it was very comforting to see most of what this community has already discovered in practice, explained in theory. The application portion of this discussion was really great and supports a lot of what we already have collectively discussed. I think BMS's like Dimitri's are good thing once proven (like his is) and especially for those who want more peace of mind and a less active roll in the technical side of an EV. Personally, I am going without one...at least to start with...who knows, maybe I will change my mind. I recognize the benefits. I've ordered 67 Calb 180 Ah cells (two spares) and am glad that I really pushed for cells that are closely matched in capacity and internal resistance. I hope they give me more than just the words....
Joined
·
1,379 Posts
Something I would really like to see is someone wire up all the odd numbered cells in their pack with a BMS, and do Jack Rickard's recommendations on the even numbered cells.
I'm not a chemist, nor do I play one on TV -- so I'm going to flail about for a minute here. Something I gathered from the video is the chemical reaction at the discharged end of the curve is different from the chemcial reaction at the overcharged portion of the curve. Maybe this is something fundamental that would make bottom balancing better than top balancing? Still flailing here, for instance maybe top balancing runs all the cells into an "overcharged" condition, but bottom balancing (and stopping as soon as any cell hits "full") keeps the cells away from a harmful regime?
I'm not a chemist, nor do I play one on TV -- so I'm going to flail about for a minute here. Something I gathered from the video is the chemical reaction at the discharged end of the curve is different from the chemcial reaction at the overcharged portion of the curve. Maybe this is something fundamental that would make bottom balancing better than top balancing? Still flailing here, for instance maybe top balancing runs all the cells into an "overcharged" condition, but bottom balancing (and stopping as soon as any cell hits "full") keeps the cells away from a harmful regime?
Hi !It all depends on terminology used and how we understand this terminology. He was way too brief on the subject and the audience was too mesmerized to ask the right questions
Here is how I understood it. Single cell management is the ability to remove individual cell from the circuit and/or reroute energy around it, such that single cell failure does not stop the whole pack. This requires power electronics ( like IGBT modules ) for each cell, to be able to route energy around bad/empty cell. Some people proposed such BMS schemes, even on this forum, but for anyone familiar with electronics involved its 100% clear that such solution is not cost effective. I think this is what Jay was referring to, he actually mentioned bypassing a cell.
String management, the way I understood him, is monitoring cell level voltages and taking action upon entire pack, i.e. stopping charge/discharge when any cell goes outside of norm. He actually mentioned cell level monitoring using simple cheap circuits. This type of management is precicely what most affordable BMS systems do. Its a string management, based on cell level monitoring.
Sure, if you have the luxury of hand picking cells based on AH and IR match from a large pile of tested cells, and also size the pack such that you never reach top or bottom, then BMS requirement is reduced. However, us DIY folks don't have such luxury.
I'm expecting some rocks being thrown my way because I am the evil BMS pusher
BTW, the video was awesome, by far the best info about Lithium batteries in a simple easy to understand form, although it probably helps to have some background in electrochemistry.
I also liked how simple in 1-2 sentences he explained how to balance a new pack before first use. There are pages and pages of forum threads here discussing complicated ways of doing the same thing that can be done in few simple steps. Parallel all cells, charge or discharge to get them outside of flat area, let them sit for a while, then connect in series, done, end of story.
Well I'm no fanatic, and I'm not orthodox follower of MrRickard either. I have had BMS on my cells, but I don't have that now. I have done some stupid thing with my cells but most of them still lives.
I really think that your BMS is one of the best value for money that you can get in ev land right now. I have the outmost respect for your work both with the BMS and the Display. At the same time much of Jacks talk about bottom balancing makes sense to me.
BTW this video with Jay was really good and informative, it would have been nice to be there to ask some more question, but he seemed to be in a hurry, or it could be that people at his level simply talks faster because they think faster then most of us.
Bypassing a cell during chargeing is exactly what your BMS in doing to some extent when it shunts. The shunt lets some of the current past the Cell that is full. For large prismatic cells there is simply no way that it would be economical to be able to take one cell out of the string, for smaller cylindrical cells it would be possible, And as I recall Tesla is pretty much doing that I think with a Fuse on each cell or something like that.
Your BMS is really good in the aspect that you can choose if you want to go with top or bottom balancing, but I don't know if it handles a problems that I have faced with my ThunderSkys.
One of my cells became a busbar. 0v just plain conductor. The BMS would not spot it, because the voltage was to low. I don't know how long I had the cell that way, but it freaked me a little, that and a shunt that went haywire and started bleeding a cell that was in normal voltage.
From that day I have no BMS.
I haven't bottom balanced yet but I plan to do.
I used a ver2.6 of the Goodrum/Fechter BMS, Probably mostly my own fault for not using an ESD safe area when soldering and handling the components.
Best Regards
/Per Eklund
Joined
·
937 Posts
Hi Dimitri,
It is not completely clear to me, he meant just that. Shunting during charging could also be interpreted in this way.
And yes, I thought he also says that monitoring is a good thing. But with the nuance he tells earlier in the video about what can go wrong with adding wires and sensors to a pack.
That is pretty much in accordance with what I think Jack's opinion is.
You must read a lot between the flaming lines to understand Jacks real opinion. He actually is waiting on a wireless BM(onitoring)S. I think.
Joined
·
8,155 Posts
He was really approaching the BMS issue from a cost standpoint, which is in line with my thinking. He said that even 10% of cell cost wasn't effective, so if your cells are around 100ah or larger then Dimitri's MiniBMS makes the grade Obviously the perfect answer is matched cells and that's what we should be pushing for from our suppliers, even if we have to pay a little extra, maybe up to 10% or so, to get them.
What I found really interesting was his suggestion that pulsing would be beneficial to the cells. I wonder if the charge reversal of regen fits into this category or if he meant higher frequency pulses?
Obviously the perfect answer is matched cells and that's what we should be pushing for from our suppliers, even if we have to pay a little extra, maybe up to 10% or so, to get them. What I found really interesting was his suggestion that pulsing would be beneficial to the cells. I wonder if the charge reversal of regen fits into this category or if he meant higher frequency pulses?
Joined
·
937 Posts
Yep, that part interested me too, but I don't understand it.
I was thinking about Simon’s $200 charger I’m building. Isn’t the Inductor (that’s so hard to obtain) the part of that charger that tries to prevent just that?
And another thing: I doubt that the regen of an AC motor is smoothened like that.
Joined
·
2,123 Posts
First BMS I built was also based on G/F design and I quickly realized that normally open bus is prone to issue you describe and also to wiring failures, etc. This was the main reason I started MiniBMS design from scratch, with normally closed loop being the key feature. NC loop immediately detects "busbar" cell or loose wiring, this is why I laugh when people state loose wiring as the reason they hate/fear BMS.
As for shunting, I always thought it to be the least important feature, basically a small bonus on top of actual BMS functions. Small amount of shunting accompanied by PTC fuse is very safe, so all that baloney about BMS burning cars down is just that, baloney. What burns cars down is people who built them lacking skills, attention to details and respect to physics. In my dealings with BMS customers I personally observed some scary things people do with their EVs, I am amazed we don't see people killed by their EVs on a regular basis. But its no different from amateur mechanics playing with ICE with gasoline and sparks in close vicinity, I wonder how many garages burned down without any batteries or BMSs in them, its all just a matter of perspective. Unfortunately some people with utter lack of perspective sport a pretty big mouth, as if one compensates for the other
Sorry for sidetracking, its just that you touched on important subject
Anyway, back to Lithium lecture topic. I also found it interesting to understand the difference between overcharge, which is destruction of electrolite, and overdischarge, which is plating of lithium on the graphite, both of which are permanent damages, but in a different way. Severity of damage of course depends on how much the cell is exposed to these conditions.
Joined
·
937 Posts
I believe you, without a doubt. But the professor in this video was talking about himself, burning down a battery with incorrect wiring. Don't know anymore what he did wrong or what happend exactly.
Point is, anyone can make a mistake. The stupid and the very smart. And a BMS with all its wiring in such a critical place, is just the part begging for it.
Joined
·
2,123 Posts
It could be interpreted this way, but I think professor Jay was focusing on more important points of pack management. I don't think he even thinks of shunting as a BMS function since in his field of large scale economics pack balancing is done at the factory, not in the car. Shunting represents such small part of overall pack management that its not worth all the attention it seems to get in EV forums. BMS spends 99% of the time managing/monitoring bulk charge and bulk discharge, everything else, including shunting/balancing is just icing on the cake. This is why I believe his references to cell management vs string management refers to removal/bypassing a cell during bulk usage, which is clearly not cost effective. String management is what most people do, when done cheaply and effectively it makes sense. When done expensive and complicated it makes no sense.
Joined
·
2,123 Posts
This is where we have a disconnect. I don't know specifically which wiring or BMS's you refer to, but BMS's that I consider effective have just one wire. 10,000 miles after installation this wire is in the same place where I left it, it didn't go anywhere, it didn't touch anything, it didn't beg for anything. It sits there just like 1000 other wires which exist in every one of 100,000,000 cars on the roads today. The wire argument is just silly. Its the same as saying if I piss in the gas tank my gas car will fail. Simple solution to the problem, stop pissing in the gas tank
Joined
·
937 Posts
Maybe. He touches the subject so shortly, we are all left speculating.
Dimitri, I like your minibms very much as my many comments about it on this site have shown. But in an effort to keep you from getting too big for your pants , I would remind you of one weakness: If the LM 339 chip (quad comparator) on the main board is taken out by a voltage spike, there is no indication the bms is no longer functioning and it will not trigger the relay for HVC or LVC. It happened on mine as you know, and I was lucky I caught it. Because I still don't trust any bms completely, I checked voltages near end of charge with my dvm and found one cell over 3.9V while the charger was still putting out high current. The green LEDs on the cell level boards were on, so everything looked ok. On the smaller pants side , you graciously sent me some optocouplers (4 of them were also taken out) free of charge, and told me the LM 339 was available at Radio Shack for about $2.00! Everything has worked fine since replacing them.
, I would remind you of one weakness: If the LM 339 chip (quad comparator) on the main board is taken out by a voltage spike, there is no indication the bms is no longer functioning and it will not trigger the relay for HVC or LVC. It happened on mine as you know, and I was lucky I caught it. Because I still don't trust any bms completely, I checked voltages near end of charge with my dvm and found one cell over 3.9V while the charger was still putting out high current. The green LEDs on the cell level boards were on, so everything looked ok. On the smaller pants side , you graciously sent me some optocouplers (4 of them were also taken out) free of charge, and told me the LM 339 was available at Radio Shack for about $2.00! Everything has worked fine since replacing them.1 - 20 of 63 Posts
