Lithium mini batteries

[s/v Jedi]

The worries about unbalanced parallel batteries is unfounded. Once they reach the upper knee in the SOC graph during charging they balance automatically.

This means it also doesn’t matter if they connect to the same busbar or not. Because when they reach the upper knee, current comes down so much that voltage drop is eliminated. I do recommend a bit longer absorption time to give the batteries farther away from the charger enough time to complete their charge. You can simply test how long they need with a multimeter and a clock, then set your absorption time to that number.

[CaptainRivet]

Originally Posted by rabbi

I fail to see a chain of events that lead to failure of the 16s, as there was no chain of events. It was a drop in with it's own BMS that failed during normal operation.

The only thing known is that all of a sudden one battery pack started to gas and minutes later the gas exploded.

Is LFP relatively safe technology? I think so.
Is it 100% safe? Nope.

read again my post...the chain of events is having 16 bad quality pooch cells in series where the tollerances added up, a basic 16S BMS with insufficent passive balancer not been able to correct the runner that these 16s pooch cells caused over time, no proper securtiy functions of BMS, no existing last resort cut off with eg a BMV712 midpoint monitioring which would have dedcted that runner and shutoff before overcharge can happen. then gasing of the overcharged cell and when its empty cell short that gives the spark to ignite the explosion is very very very bad construction of these cells.

Explosion can only happen if the batteries where in a small room with no proper ventalation. and well one pooch cell cannot create so much gas for such an explosion so multiple cells must have gased.
that is a long chain of events.

one measure to interrupt and to prevent this chain: use prismatic cells where the layers are wrapped around anode and catode as these cannot get a cell short due to mechanical failure of the terminals eg using fixed busbars in 16 cell in a row without claming. the reason i used the 272AH lishen and not the 280AH EVE who has the z-layer setup internally.

[CaptainRivet]

Originally Posted by s/v Jedi

I do recommend a bit longer absorption time to give the batteries farther away from the charger enough time to complete their charge. You can simply test how long they need with a multimeter and a clock, then set your absorption time to that number.

that is the only proper reason you need absorption on Lifepo4 batteries otherwise the risk of overcharging through absorption is a much higher risk then the 1-2% of capacity gain you get when doing so.
If you or other is interested Eric bertscher of Nordkyn wrote a very good article about that and Andy from offgrid garage made several videos about absorption...bottom line is for LFP you don't need it but can use it. Additional the charge current on a lot of vessels with a big LFP bank is 0.2C at optimum harvest per day, early or late in day that is already basically absorb charge in bulk mode of the MPPT the bank when charging 1000AH with 15-20A...different in offgrid, there you are massivly overpaneled compared to bank size like eg Andy.


And well if the balancer is passive and too small for the imbalance 16 low quality cells in series can and will create, you can absorb till year 3000 and it won't be able to fix the problem, it will get bigger and bigger till....depends on installation. same cells, same BMS in 8S it won't be a problem.
thats why its also a big difference if you buy budget dropins in 4 or 8S config then 16S. A risk not many people are aware. this is the huge downside of 48V...and why 8S is the best compromise here. And thats why there are in 16S special batteries made for propulsion as this uses higher currents which excallerate the 16S proplematic, same do EV manufacturer.

in 4S to use parallel cells like i did if they are of very good quality is not an issue and they won't drift apart over time as that battery is generally stable. In 8S that is for sure an issue over time if not winston cells and 16S doing that is screaming for having huge trouble.
the burning tesla EVs is exactly here the root cause of the many small 18650 cells in series and parallel, you can fix a big part of it with active balancing on the LI.-ION which you can't do with Lifepo4 due to charge curve. A isssue tesla had/have on model 3 with the non Li-ion cells .

[Rockinar]

Originally Posted by rabbi

BS? Nope

No video but a first hand story of a fellow cruiser including some pictures, followed by a discussion about the most likely cause:
https://www.akkudoktor.net/forum/akk...ne-erklaerung/

Thousands and thousands of lithium battery on boats. All you got against them is "One time this guy in a internet chat room...."

This is more of a case for them than against them.

[rabbi]

Originally Posted by Rockinar

Thousands and thousands of lithium battery on boats. All you got against them is "One time this guy in a internet chat room...."

This is more of a case for them than against them.

It helps to read the other posts.

[CaptainRivet]

Originally Posted by Rockinar

Thousands and thousands of lithium battery on boats. All you got against them is "One time this guy in a internet chat room...."

This is more of a case for them than against them.

So many things wrong in this thread....

And carring a gasing and burning cell, well celebrate your 2nd birthday...
Never heard of this brand...sounds like grade D EVE cells...name ...LF 280 is EVE rewrapped.
Just stay away and use the known ones like EVE, Lishen, calb...

[donradcliffe]

Originally Posted by CaptainRivet

Yes they do because they have mosfets BMS and the mosfets are NOT calibrated means for the first milliseconds one poor weakest mosfet has to carry all the load and the more in parallel the more load is on that one mosfet during that milliseconds and it doesn't even need a big load to fry them=slow death of the BMS as the mosfets will die one after each other over time leaving less mosfets to carry the loads. The mosfets are oversized so till smoe load one can carry that load but if you put more in parallel the load on that weakest mosfets exceeds its current spec even during a light load on the 6x100Ah. Thats why limit on 3 or 4 in parallel.
Thats why i always recommend the biggest capacity with the biggest cappacity BMS as the mosfets are bigger in current rating.

Rule of thumb if a 12V 200A mosfet BMS (not the total battery with cells, just BMS) is below 350Euro it cannot have calibrated mosfets even if purely chinese build and sold via its on webshop.
Calibrated means all mosfets are identical and switch in the exact milliseconds synchronized, not calibrated mosfets don't and there is always one weakest who get the total load in the first milliseconds.

Eg the mueller energy BMS 250A build by jk for mueller is based on the 300A JK BMS and has calibrated mosfets and cost 470Euro. The 300A JK equivalent is 180-230Euro depending on balancer with uncalibrated mosfets.
I wouldn't hesitate a sec to put 750A on 3 mueller energy 250A BMS in parallel but not more then 450A on 3x300A jk bms because of that.

You really need to take a course in dynamic circuit theory. Where do you come up with nonsense like 1 mosfet out of 60 in parallel is going to carry all the load?

[s/v Jedi]

Originally Posted by donradcliffe

You really need to take a course in dynamic circuit theory. Where do you come up with nonsense like 1 mosfet out of 60 in parallel is going to carry all the load?

On what do you base your assertion that he is wrong? All the parallel MOSFET’s switch at the exact same nanosecond?

[CaptainRivet]

Originally Posted by donradcliffe

You really need to take a course in dynamic circuit theory. Where do you come up with nonsense like 1 mosfet out of 60 in parallel is going to carry all the load?

Jedi told you why..
And what i state comes a) from my own experience based on practical use and not theortical courses and b) from my buddy who has an NPO producing commercial EV and powerwalls and eg for the cheap budget powerwall series he uses mosfet JK BMS and tested (means killed) dozen of them to see what they can do and whats the limit he can really use them to figure if he develops his own with JK or use their standard ones. Also in parallel.
And c) thats what JK R&D told me and my buddy how to spec and install parallel BMS, which they state you shouldn't do due to mosfets are not calibrated.
And d) you can talk to Mueller Energy in Australia who developped together with JK a mosfet BMS with calibrated mosfets that you can officially parallel and their 250A Mueller BMS produced by JK cost 470Euro, so triple of the standard 200A with uncalibrated mosfets they use as base.

[donradcliffe]

What to you think kills mosfets? It ain't voltage and it ain't current. It's temperature. The time it takes to overheat is an awful lot of nanoseconds.

Why do you think BMS ratings are a combination of current and time?

PS how many mosfets are already in parallel in the typical BMS?

[CaptainRivet]

Originally Posted by donradcliffe

What to you think kills mosfets? It ain't voltage and it ain't current. It's temperature. The time it takes to overheat is an awful lot of nanoseconds.

Why do you think BMS ratings are a combination of current and time?

PS how many mosfets are already in parallel in the typical BMS?

All nice theories, i gave you my sources and all line up...
Don't believe it buy 3x 200A JK BMS, connect them in parallel without a class T fuse and put 500A on it and watch the mosfets popping like popcorn...record if they don't i pay them, if they do its on you, deal?

[wholybee]

Originally Posted by CaptainRivet

All nice theories, i gave you my sources and all line up...
Don't believe it buy 3x 200A JK BMS, connect them in parallel without a class T fuse and put 500A on it and watch the mosfets popping like popcorn...record if they don't i pay them, if they do its on you, deal?

Have you tested this? I have tested my JBD with a direct short, multiple times. The MOSFETS are very, very fast. They are fast enough to both save the BMS from damage, and save the fuse from blowing. It takes more heat to blow a fuse than it does to damage a MOSFET.

Every time I hear about how fragile MOSFETS are I wonder where all the examples of blown BMS's are. Then I remember how frequently I had to change relays in high power applications decades ago before everything changed to solid state. MOSFETS are far more reliable, and if the circuit is well designed to protect itself from overcurrent you don't need to worry about damage from overcurrent. Sure, they can and will fail. But not as often as relays.

And, because of the time/heat curve, the instantaneous peak current is much higher than the continuous rating. Even if a BMS is rated for only 100A, with a peak rating of 150A for 30 seconds, if you are talking peaks of less than 1 second, the capacity goes up even more. I mention this because I know of several boats with cheap 100A daly BMS's (which I generally don't recommended) working just fine with 1000W windlasses, several years in. One of those boats installed the Daly after his load relay failed closed and his battery was drained to 0V and ruined. The Daly and 4 Lishen cells from Amazon were the only things he could get in 48 hours. I did the install. That was 3 years ago, still working fine, full time cruising and anchoring every week.

So, not theory, and not sources from the Internet, but my personal tests and experience.

Will you also pay for the LFP battery as the power source for this test? I don't otherwise have one to use for it. But damn sure, from previous tests, that I would win that bet.

[CaptainRivet]

Originally Posted by wholybee

Have you tested this? I have tested my JBD with a direct short, multiple times. The MOSFETS are very, very fast. They are fast enough to both save the BMS from damage, and save the fuse from blowing. It takes more heat to blow a fuse than it does to damage a MOSFET.

Every time I hear about how fragile MOSFETS are I wonder where all the examples of blown BMS's are. Then I remember how frequently I had to change relays in high power applications decades ago before everything changed to solid state. MOSFETS are far more reliable, and if the circuit is well designed to protect itself from overcurrent you don't need to worry about damage from overcurrent. Sure, they can and will fail. But not as often as relays.

And, because of the time/heat curve, the instantaneous peak current is much higher than the continuous rating. Even if a BMS is rated for only 100A, with a peak rating of 150A for 30 seconds, if you are talking peaks of less than 1 second, the capacity goes up even more. I mention this because I know of several boats with cheap 100A daly BMS's (which I generally don't recommended) working just fine with 1000W windlasses, several years in. One of those boats installed the Daly after his load relay failed closed and his battery was drained to 0V and ruined. The Daly and 4 Lishen cells from Amazon were the only things he could get in 48 hours. I did the install. That was 3 years ago, still working fine, full time cruising and anchoring every week.

So, not theory, and not sources from the Internet, but my personal tests and experience.

Will you also pay for the LFP battery as the power source for this test? I don't otherwise have one to use for it. But damn sure, from previous tests, that I would win that bet.

I gave you my sources already.
My buddy tested that. If i remember correctly the 200A JK has 12 x 30A mosfets in parallel. So JK already takes that into account for their BMS by giving a 200A rating and officially state you shouldn't parallel them.

So if a 30A mosfet gets full 500A it pops right away...and 500A from an inverter means effectively 500x1.41=700A on/off....

Search for Andy from Offgrid garage, he tested the 200A JK and with 200A the weakest mosfet blew.

[CaptainRivet]

Originally Posted by wholybee

Have you tested this? I have tested my JBD with a direct short, multiple times. The MOSFETS are very, very fast. They are fast enough to both save the BMS from damage, and save the fuse from blowing. It takes more heat to blow a fuse than it does to damage a MOSFET.

Every time I hear about how fragile MOSFETS are I wonder where all the examples of blown BMS's are. Then I remember how frequently I had to change relays in high power applications decades ago before everything changed to solid state. MOSFETS are far more reliable, and if the circuit is well designed to protect itself from overcurrent you don't need to worry about damage from overcurrent. Sure, they can and will fail. But not as often as relays.

And, because of the time/heat curve, the instantaneous peak current is much higher than the continuous rating. Even if a BMS is rated for only 100A, with a peak rating of 150A for 30 seconds, if you are talking peaks of less than 1 second, the capacity goes up even more. I mention this because I know of several boats with cheap 100A daly BMS's (which I generally don't recommended) working just fine with 1000W windlasses, several years in. One of those boats installed the Daly after his load relay failed closed and his battery was drained to 0V and ruined. The Daly and 4 Lishen cells from Amazon were the only things he could get in 48 hours. I did the install. That was 3 years ago, still working fine, full time cruising and anchoring every week.

So, not theory, and not sources from the Internet, but my personal tests and experience.

Will you also pay for the LFP battery as the power source for this test? I don't otherwise have one to use for it. But damn sure, from previous tests, that I would win that bet.

So if a 30A mosfet gets full 500A it pops right away...and 500A from an inverter means effectively 500x1.41=700A on/off....and simulate a LVC and then shut on again with full load, the peaks will be higher then 700A.
Its not fragile but with that overload of 20 or more times its rating it won't survive that a nanosecond.
Maybe the first 20 times but wears as its always this weak one and 21th time it blows. The mosfet pop not right away but some seconds afterwards, the nanosecond peak damages it and then the standard load give it the rest.

Search for Andy from Offgrid garage, he tested the 200A JK and with 200A via inverter the weakest mosfet blew, so here the mosfets saw effectively 200A×1.41=280A on/off.

You have to test a big lot of BMSs as you can be lucky and get one where the mosfets by luck are very close while the next one they can be far appart.
Thats why i said i pay for the 3 BMS and will take them after the test as if they in parallel survive that these 3 have basically by luck calibrated mosfets but the chance this happens is quite low. The peak rating of the mosfet BMS given is normally the combined continious rating of all parallel mosfets together. And well when you are lucky and these mosfets are all close a 100A cont speced BMS can do 150 or even 180A as they typically have 6x30A or 4x50A mosfets in parallel.

I run sometimes 500-550A loads and also start from house thats why i chose the non current carring ElectrodacusBMS. Otherwise i would have taken 3x JK 200A BMS.
Dalys are bad from their peoduction quality, software and balancing, the hardware is quite good and overspeced as they use mosfets with 5% deviation allowed while JK uses mostly the 30A ones with 1%. The 250A daly has 10x50A mosfets in parallel.

[more]

Originally Posted by CaptainRivet

So if a 30A mosfet gets full 500A it pops right away...and 500A from an inverter means effectively 500x1.41=700A on/off....and simulate a LVC and then shut on again with full load, the peaks will be higher then 700A.
.

what cable you use to carry 500A load.
you know is very,very stupid pull 100A from 12v battery. or use bigest inverter 2000w on 12v system.
watch from begining to end.couple time again,again also have video of your T-fuse-Fast acting fuse. when use ,why ecc
this for AC but similar is DC only different is arc arestor,and contactor.

https://www.youtube.com/watch?v=gqEu...neeringMindset