18650 vs prismatic - which is safer?

Desertprep

1 kW
Joined
Oct 27, 2007
Messages
352
Location
United States
Are the 18650 cell packs safer than prismatic? Last night, my scooter caught on fire while it was being charged. Fire burned...all that was left was the steel in the frame and the hubs. I had a new battery that was prismatic instead of 18650 :(
 
pengyou said:
Are the 18650 cell packs safer than prismatic? Last night, my scooter caught on fire while it was being charged. Fire burned...all that was left was the steel in the frame and the hubs. I had a new battery that was prismatic instead of 18650 :(

Chemistry make the difference... Not form factor.

A 18650 and Prismatic and pouch form factor may be safe chemistry or dangerous chemistry.

You can aquire lifepo in pouch, prismatic or cylindrical. Safest.

You can aquire lion in pouch, prismatic or cylindrical. Flammable.

You can aquire lipo in pouch, prismatic or cylindrical. Very flammable.
 
The knowledge of the operator

and thus the care provided

is just as important if not more

than the chemistries involved.

But yes, LFP and LTO are **much** more forgiving of ignorance and carelessness

than the higher voltage li-ion chemistries at 3.6 - 3.7V nominal.

Using appropriate hardware to **assist** in oversight, monitoring, testing is part of the process,

but never think anything should be fully automated letting you walk away especially charging

as are documented scheduled and checklisted standard procedures

you should always **know** the SoH of the cells

no surprises

proactively replacing with new long before obvious wear symptoms of EoL start to appear
 
As above, other factors predominate

But all else equal, cylindrical cells are safer
 
fatty said:
As above, other factors predominate

But all else equal, cylindrical cells are safer

Versus automotive prismatics? How?

Even if the cells are equally reliable, there are lots more opportunities for something to be seriously wrong with packs that have hundreds of interconnections instead of tens.
 
For fire-safety, the LiFePO4 chemistry remains a widely well-regarded chemistry.

They have a lower nominal voltage, so it takes more cells. For instance 18650 cells would be 13S for 48V, and LiFePO4 would be 16 cells in series...
 
Chalo said:
Even if the cells are equally reliable, there are lots more opportunities for something to be seriously wrong with packs that have hundreds of interconnections instead of tens.
Agreed, but we're holding all else --including quality of interconnects-- equal. Spot-welding cylindrical cells is done reliably in third-world countries by the hundreds of millions, maybe billions? Non-solderable foil-tabbed automotive prismatics that are to be laser welded have at least as many opportunities for something to be seriously wrong. We have members here recommending foil tape and paperclips to connect tabs :shock:

Chalo said:
Versus automotive prismatics? How?
-Steel shells and cell holders. How many ebike pack housings are crash-rated?
-Packs of cylindrical cells necessarily include voids, and cell holders increase void space further.
Both cylindrical and prismatics can fail, but cylindricals are mechanically more robust and cylindrical packs are less likely to cascade into full-pack failure.
 
fatty said:
"Non-solderable foil-tabbed automotive prismatics "


Chalo said:
Versus automotive prismatics? How?

?!?!?!?! What is that? ?!?!?!?!?! "Non-solderable foil-tabbed automotive prismatics "

I know of solder-able pouch cells, with tabs, but I never saw a "Prismatic" with "tabs".. I have seen them with studs, and or pads, for laser welding.. but not any with a sort of tab.

A pouch cell has tabs, but a pouch cell is not prismatic. A pouch cell is a pouch cell and has layers, not a jelly roll inside ( like the Ford C-Max prismatic, the Volt pouch, very different, one has solder-able tabs and one uses bolts.. ) ( neither have " non solder-able tabs" and the " foil tab" (pouch) is NOT prismatic. )


How many ebike pack housings are crash-rated?
-Packs of cylindrical cells necessarily include voids, and cell holders increase void space further.
Both cylindrical and prismatics can fail, but cylindricals are mechanically more robust and cylindrical packs are less likely to cascade into full-pack failure.
I would like to see actual tabulated stats on this. Claim.


Trust me, any highway going lipo or lithium cell goes through MUCH more stringent safety standard testing than ANY ebike cell. Ever. Period. In the factor of billions of dollars more R&D, certifications, than any ebike manufacturer gets the certifications for. Every highwa...

Highway going pouch cell systems are bolted too, not only laser or ultrasonically welded. I have seen bolts, laser and ultrasound,.

Prismatic is 100% fill efficiency.

Cylindrical is only 91% fill efficiency.

Both cylindrical and prismatics can fail, but cylindricals are mechanically more robust and cylindrical packs are less likely to cascade into full-pack failure.

I would love to see actual referenced data as to this ( fail rate you refer). There is only one manufacturer who uses the cylindrical cell in OEM highway going vehicles. There s a reason for that.

I know of highway going certified NHTSA cells that are much safer than any ebike cell.... ( much more expensive too.. ) No fire, No conflagrations, can be punctured, and are listed in the datasheets as " NOn Flamable". More power density, safer. That is why they use them in vehicles for roadgoing use. Oh hell yea they tested teh hell out of highway cars, but not the ebike junk.

Ebike fires are much more common than ... Ford or Chevy or Tesla fires. IDk. I just dont want people to confuse pouch cells ( with tabs) and prismatic cells ( with ... jelly rolls) .
 
Sorry hear about your scooter, any idea what might have happened, accident damage, old cells, badly out of balance, ect , this second battery fire on here in last couple days, be interested see picture of scooter :es:
 
Cylindrical are great if you're only doing small capacity.

For use cases where 100Ah is small, then just a single string of big prismatics is far easier, denser and more reliable.

Easiest to just buy the ones with posts built in, or threaded sockets for bolts and use marine wire with ring terminators.

Yes the case needs to compress them against swelling, but lots easier than spotwelding literally hundreds of the bitty ones
 
Greendog said:
Sorry hear about your scooter, any idea what might have happened, accident damage, old cells, badly out of balance, ect , this second battery fire on here in last couple days, be interested see picture of scooter :es:
The battery was only 4 weeks old. My school hauled away the carcass (the only thing remaining was the steel frame) before I even had a chance to take photos, so it looks like I am going to get screwed at that end.
I was charging the battery when it happened. The battery was low, down to 20% or so, and usually takes 6 hours to charge fully from that point. I have charged the battery before for as long as 8 hours with no ill side effects. This is VERY unsettling.
 
I build batteries out of both and I think they are both more dangerous and 'safe' in different ways.

Safety wise;
-Cylindrical cells like 18650's are great for being robust, easy to handle and easy to spot weld. They also won't change shape/size like pouch cells.
-Pouch cells can be large capacity, and various shapes so require less connections, thereby making less failure points. They can also dissipate heat much better due to the larger surface area.

Danger wise;
-Cylindrical cells become bombs/rockets when under thermal runaway or in a fire. They can also be miss spot welded on the posotive side leading to a dead short.
-Pouch cells are easy to puncture if not protected well, and will puff up when run hard. The puffing up is probably the worst thing about pouch cells, but it can be dealt with through compression, and even reversed through controlled venting and sealing, neither of which is ideal.

I can't say I have a preference for either as I use both depending on the circumstance. I find pouch cells can be a bit more stressful to assemble though as the tab's can be pulled or twisted enough to rupture open the cells if not careful when crimping or bolting them together. That being said, having a bad 18650 going into thermal runnaway in the middle of a large pack being assembled is quite the scary situation. Extracting that cell without loosing the whole pack is one hell of a challenge!!!

Cheers
 
Contactless, no-weld options for cylindricals is a great option to allow quick atomization, testing each cell, tossing the too-worn ones and reassembly.

Annual maintenance process.

But as with bitty size cells in general, only suitable for small packs, as in under 100Ah or so.
 
fatty said:
Chalo said:
Even if the cells are equally reliable, there are lots more opportunities for something to be seriously wrong with packs that have hundreds of interconnections instead of tens.
Agreed, but we're holding all else --including quality of interconnects-- equal.

If equal means the same odds of failure per cell connection, then more cells give you proportionally higher chance of failure.

Spot-welding cylindrical cells is done reliably in third-world countries by the hundreds of millions, maybe billions?

And many of them fail. Those of us who've seen a decent sample have experienced it directly. Six sigma or better manufacturing facilities have a pretty good record. None of us are that.
 
Chalo said:
If equal means the same odds of failure per cell connection, then more cells give you proportionally higher chance of failure.
I don't think it's that simple. And maybe I'm mentally oversimplifying, but I think the risk of high number of simple, reliable connections versus the low number of complex, failure-prone connections tends to balance out. I think it's a wash, and not a significant factor in answering the question.

Chalo said:
And many of them fail. Those of us who've seen a decent sample have experienced it directly. Six sigma or better manufacturing facilities have a pretty good record. None of us are that.
The failure rate is lower than competing solution, but there's orders of magnitude more cylindrical cell packs in use, so you just hear about those failures more.

The fact is that spot-welded cylindrical cell packs are not just one common or leading solution for ebike applications, but the overwhelming solution, despite the otherwise unattractive qualities of cylindrical cell packs. In fact, cylindrical cells don't distinguish themselves on any other metric: as noted, construction is laborious, charge and discharge rates are limited, even energy density is middling when accounting for holders and connections.

But there are kids spot-welding salvage packs outside in third-world countries that do well enough most of the time. In fact, this subforum reports more failures from BMSs than from the cylindrical cells or connections.

Frankly, the only redeeming quality of cylindrical cells is their inherent mechanical robustness.
 
Such cells are used because of their energy density, and ability to fit Ah into small oddly shaped spaces.

Prismatics, with hard shell and bolt posts as opposed to raw pouches, have much easier and MUCH more reliable connections.

LiPo packs as well, but there the downside is cost.
 
When you guys figure this out, contact GM. We are the proud owners of a 2017 Chevy Bolt with 39K miles on it so far. GM issued two battery management software updates in 2018 to remedy a situation where a number of Bolts, with failed cells, stopped the car dead in the highway.

Over the last year there have been over a dozen spontaneous fires after charging. The last two had just had the latest of two software updates related to the fires, using existing onboard sensors, to monitor cell voltage, and six of the ten battery module ends for temperature...to blow the horn, and flash the lights when it goes up in flames.

Now they are telling us to only charge to 90%, and to not discharge below 30%, only charge during the day while watching it charge. After charging, we are to move the car away from any structures. They say they will get back to us in September when they will begin to replace some, most, or all batteries in 60K Bolts built from late 2016 to early 2019. Oh!...and have a nice day. lol

Chalo, these are 288 pouch cells 96s3p, for what it's worth.

https://allev.info/category/bolt/
 
I guess GM/LG have a good idea of the exact problem with those Bolt packs, but i have not seen a detailed explanation .
Is there any other detailed reports to identify potential causes. ?
Remember the fires with the Boeing 777 YUSA lithium prismatic packs, and the detailed investigations there..?
Changing out the pack is obviously a “catch all” to eliminate the issue, but it does not tell us much about the root cause.
It will be good to find out if it is a “cell” problem... ( materials, manufacturing quality, basic chemistry, etc etc)..or if it is a “cell management” issue ( BMS, charging, assembly, software, etc).
 
Hillhater said:
I guess GM/LG have a good idea of the exact problem with those Bolt packs, but i have not seen a detailed explanation .
Is there any other detailed reports to identify potential causes. ?
Remember the fires with the Boeing 777 YUSA lithium prismatic packs, and the detailed investigations there..?
Changing out the pack is obviously a “catch all” to eliminate the issue, but it does not tell us much about the root cause.
It will be good to find out if it is a “cell” problem... ( materials, manufacturing quality, basic chemistry, etc etc)..or if it is a “cell management” issue ( BMS, charging, assembly, software, etc).

from the safety recall notice:
"experts from GM and LG have identified the simultaneous presence of two rare manufacturing defects in the same battery cell as the root cause of battery fires in certain Chevrolet Bolt EVs"
 
PRW said:
"experts from GM and LG have identified the simultaneous presence of two rare manufacturing defects in the same battery cell as the root cause of battery fires in certain Chevrolet Bolt EVs"

Yes. They just discovered the second defect recently. Of course, they aren't saying if these defects can be detected with nondestructive tests.
 
I reckon I have a pretty good idea...seeing as I've built a powerwall using these exact LG cells salvaged from an as new Bolt, imported over to Australia.

I had one cell module (2 cells) brand new ever so slightly venting through a hole I could not even see, but only under load. The only way I knew about it was a noise it made when I was testing the pack under load.
At the time I did not know which exact cell module of 3 it was making the noise so I disassembled the pack and momentarily dead shorted each module to see which one made noise. Sure enough one of them did, but only when dead shorted. I have a video of this somewhere I can re-post if you want.
The crazy thing is, apart from the noise, this cell module performed almost identical to the others. Maybe 5% less capacity.

Anyway, my theory is that these cells have a manufacturing defect where they can be open to the elements. Over time air gets in causing oxidation, shorting, corrosion etc and excess heat. Because it can happen to just one cell and not every cell is temperature monitored, it can catch fire easily as the cells around it absorb excess heat during charge/discharge.

I'm still using my powerwall, but have reduced output/charge rates and keep a close eye on it. I would say that the majority of Bolts are ok, but because this is a random fault that cannot be fixed, they will have to end up replacing all the packs. When they do, guy's like me will be waiting for the used cells. :mrgreen:

Cheers
 
Back
Top