If you are building packs today, which cells to choose?

Anyone can explain why mg1 could be better than pf apart from the price? I had good experience with pf and even punctured some of them. No issue apart from voltage loss.

Never heard of issue with pf either.

But heard some issue in the past with lg cells
 
PF are getting old, manufacturing seems to improve. I still like some of the older cells, but i believe they get better every year.
 
Wheazel said:
cal3thousand said:
I think I would also include NCR18650b. Properly sized into 6P+ groupings, they will easily provide 40A continuous at 2C and be happy about it. If you keep these cells with 20% to 80% SOC, you can exceed 28,000 cycles.

There's many a good reason that Tesla uses them in the Model S. You just have to make sure you size the pack large enough that 2C will cover your Current requirements.

This sounds like pure BS. The second statement about Tesla is as already pointed out not correct, but please provide some evidence for the first claim.
My guess is that the B cell will suffer from extended use at anything higher than 0,5C.

http://lygte-info.dk/review/batteries2012/Panasonic NCR18650B 3400mAh (Green) UK.html

I got the information from this site: http://blog.evandmore.com/lets-talk-about-the-panasonic-ncr18650b/

He links to a study that is difficult to retrieve. Not calling it proof, but that's where I read it.
 
is this a good option ?
https://www.imrbatteries.com/samsung-48g-21700-4800mah-flat-top-battery/

the size and capacity seems good compare with 30Q. only that the discharge rate is low? 9.6 vs 15a
 
rojitor said:
4800 is impressive. The amps are not that bad.
These have my attention now:
https://eu.nkon.nl/rechargeable/other/sanyo-ncr20700b-lithium-battery.html

Yes, the 20700b have my attention too. Correct discharge rate and the high capacity could lead to approx 20% less cells needed in a pack.

Is there any test reviews or other manufacturing info on these cells? edited: just found this... http://lygte-info.dk/review/batteries2012/Sanyo NCR20700B 4000mAh (Red) UK.html
 
From my previous tests more than 8 types of modern cell format 18650 and 20700 I can make a general conclusion that if you design your battery for 50% DoD in average with up to 1C discharge current, it really does not matter what cell you actually use. All of samples from Samsung, LG, Panasonic and Sanyo (most of them are from Nkon) still have more than 87% of their initial capacity after 1000 cycles under those conditions (Even High Energy cells like LG MJ1, Samsung 35E and Sanyo GA).

So significant difference in cycle life can only be observed over 50% of the DoD average. Right now I have tests only for full 100% DoD under the same conditions 0,5C charge and 1C discharge. In genreal all tested NCA chemsitry cells (Panasonic and Sanyo NCR type) have faster initial capacity drop but most of them still remain more than 70% of initial capacity after 1000cycles (like Sanyo GA and Panasonic PF). Almost all other cells are based on NMC chemistry (INR type). They have slower capacity drop but some of them reach 70% of initial capacity just before 500 cycles. The biggest dissapointment for me are Samsung 30Q and 35E where both of them were practicaly dead after finished 500 cycles. On the other hand Samsung 29E seems to me very good and even better than Panasonic PF, but I have now only 200cycles run at 100%DoD and 750cycles at 50% DoD. LG cells also shows relatively good results at 100% DoD.

I hope that I can make data with graphs available soon.
 
Pajda said:
0,5C charge and 1C discharge.
...
The biggest dissapointment for me are Samsung 30Q and 35E where both of them were practicaly dead after finished 500 cycles.
Correct me if I'm wrong here but 1C discharge for 35E would be ~35A which is about 3,5 times their rated discharge current.
 
minimum said:
Pajda said:
0,5C charge and 1C discharge.
...
The biggest dissapointment for me are Samsung 30Q and 35E where both of them were practicaly dead after finished 500 cycles.
Correct me if I'm wrong here but 1C discharge for 35E would be ~35A which is about 3,5 times their rated discharge current.

Yes, you are wrong.!
1C discharge would be 3.5amps ..
 
minimum said:
But 0,5C charge would be 1,75A which is still over rated current.
Yes this is good point. In my test procedure I try to test all cell samples under the same testing conditions. The charging current of 0.5C is recomended as a standard charging rate (even for cycle life) by almost all manufacturers for most modern cells. But there are some exclusions like above mentioned Samsung 35E, Panasonic BE, BM or LG F1L with lower 0.3C charging current, recomended for cycle life.

In Samsung 35E datasheet is as standard charge value 1.7A (0.5C), which I have used im my tests. But there is also stated that for manufacturer guaranteed cycle life (60% of standard capacity >= 2010mAh after 500 cycles at 100% DoD) is charging current only 1.020 mA. By the way my result (with 0.5C charge) is 2177mAh of standard capacity after 500 cycles. But the next course is already steeply falling.

By the way, I badly expressed. I have no as big problem with Samsung 35E cycle life, but I am totally dissapointed with 30Q cycle life at 100% DoD (if I did not have a luck for bad sample) especially when compared to its LG HG2 competitor.
 
Instead of the Sanyo 20700B's I'm trying to build my next pack out of Sanyo 20700A cells. Obviously they have less capacity, just 3100mAh... but they are also rated at 30A continuous discharge per a cell, and from what I have heard, that is a realistic number. My goal is to build basically the smallest/lightest 20s pack I could that would be able to utilize the 3000w power of my new Cyclone. I'm going to try to pull this off with a 20s2p pack, for my short distance rides that will be the majority of my use of the bike.

Now for the cool part about this. I'm getting these Sanyo 20700A cells from 18v 6.3ah Bosch battery packs. Each oack is costing roughly $55 shipped, and contains 10 cells in a 5s2p pack. On top of that, the cells are already inside a nice holder that's supposed to be some kind of cooling plastic (considering these are Bosch, I wouldn't doubt this being true... though I am not too familiar with plastic). ALSO, the cells are already connected by a copper plate that is laser welded to the cells, with copper tabs at the beginning and end of each pack, making it pretty damn easy to connect these packs into a 20s pack (could really do it pretty easily with a simple soldering iron... WITHOUT the issue of causing damage due to over heating the battery by directly soldering to the cells). Given this 20s pack will only be a 6.3ah pack... but that should be plenty for 90% of my rides, and I plan to make a second or possibly 3rd to run in parallel for longer rides.

If you ask me, for the price, and the fact that these cells are pretty much all set up, good to go... could easily be done with minimal cost in tools (I assume anyone making their own pack already owns a halfway decent soldering iron)... and the 40 cells + holder + copper plate laser welded (better than any home welder is gonna do) only cost about $220 so far.

Like I said, trying to go small and lightweight, but still make 3000w of power. If I was going for capacity.. and could find a good source for them, I'd go with the Sanyo 20700B most likely... tho I currently don't know any good place to get them at a good price. I suspect they'll be showing up inside some power tool's high capacity power pack soon enough... or something comparable. And for me, not being able to find battery holders the right size would be pretty trivial... I can easily 3D model and print them to the exact size and layout I want (given I already have multiple 3d printers... but even if I didn't... could get one for under $300 that would be more than adequate to do the job).
 
Sanyo NCR20700B in my standard 0,5C-1C cycle life test behaves pretty same like others NCA chemistry cells, such as Panasonic PF or an old NCR18650B. So I do not see any technical advance in this specific parameter.

Right now I am expecting delivery of the Samsung INR21700-48G from Fasttech and finally some samples of original Tesla 18650 cells. So I hope that I can start with their cycle life tests soon.
 
if I were building battery pack today ....
the best cells are never sold to the DIY people.
SAFT would not sell to you
LG Chem would not sell to you
KOKAM would not sell to you
NEC /Nissan Leaf/ would never sell to you
etc, etc.
so buying modules from wrecked EV is the way to go with me.
I am building 2 packs from 41Ah Leaf cells from wrecked 2016
 
progrock said:
Instead of the Sanyo 20700B's I'm trying to build my next pack out of Sanyo 20700A cells. Obviously they have less capacity, just 3100mAh... but they are also rated at 30A continuous discharge per a cell, and from what I have heard, that is a realistic number. My goal is to build basically the smallest/lightest 20s pack I could that would be able to utilize the 3000w power of my new Cyclone. I'm going to try to pull this off with a 20s2p pack, for my short distance rides that will be the majority of my use of the bike.

Now for the cool part about this. I'm getting these Sanyo 20700A cells from 18v 6.3ah Bosch battery packs. Each oack is costing roughly $55 shipped, and contains 10 cells in a 5s2p pack. On top of that, the cells are already inside a nice holder that's supposed to be some kind of cooling plastic (considering these are Bosch, I wouldn't doubt this being true... though I am not too familiar with plastic). ALSO, the cells are already connected by a copper plate that is laser welded to the cells, with copper tabs at the beginning and end of each pack, making it pretty damn easy to connect these packs into a 20s pack (could really do it pretty easily with a simple soldering iron... WITHOUT the issue of causing damage due to over heating the battery by directly soldering to the cells). Given this 20s pack will only be a 6.3ah pack... but that should be plenty for 90% of my rides, and I plan to make a second or possibly 3rd to run in parallel for longer rides.

If you ask me, for the price, and the fact that these cells are pretty much all set up, good to go... could easily be done with minimal cost in tools (I assume anyone making their own pack already owns a halfway decent soldering iron)... and the 40 cells + holder + copper plate laser welded (better than any home welder is gonna do) only cost about $220 so far.

Like I said, trying to go small and lightweight, but still make 3000w of power. If I was going for capacity.. and could find a good source for them, I'd go with the Sanyo 20700B most likely... tho I currently don't know any good place to get them at a good price. I suspect they'll be showing up inside some power tool's high capacity power pack soon enough... or something comparable. And for me, not being able to find battery holders the right size would be pretty trivial... I can easily 3D model and print them to the exact size and layout I want (given I already have multiple 3d printers... but even if I didn't... could get one for under $300 that would be more than adequate to do the job).

one issue you may run into is that they've likely got a bms inside each one of those packs - and that BMS is likely rated at less than 20s voltage. Just make sure you check before series'ing them all!
 
Samsung 30Q for excellent high-rate performance during the whole life of the cell
Panasonic 18650BM for excellent cycle life, safety, overall performance and value (Tesla cell)
Sanyo GA for best initial performance but heavy internal resistance degradation over time
 
Very good 35E price (3.1€) right now from nkon
https://eu.nkon.nl/rechargeable/18650-size/samsung-inr18650-35e.html
Very curios about cycle life vs PF cell.
In electric car, cycle life matters 7x more!
 
35E has very very bad cycle life ! And it sags a lot more under 10A discharge. Its initial capacity is the highet of the holy trinity GA MJ1 35E but after 20 cycles it becomes the worst.
 
I can confirm that Samsung 35E and 30Q both have very poor cycle life at 100% DoD. But a surprise for me is Samsung 29E cycle life which is significantly better after 300cycles than Panasonic PF, where PF is one of the best cells in this parameter. Unfortunatelly I do not have results for more than 300cycles of 29E right now. Also 29E have slightly higher IR than Panasonic PF.
 
Is there any news on your 29E cycle test?
 
Hello,

i'm about to order a battery pack for a TSDZ2 36V motor with max 16Amp consumption on battery side.
I'm hesitating between Samsung 30Q cells in 10S5P configuration for 15Ah in total and Sanyo NCR18650GA Cells in the same configuration for a total of 17.5Ah capacity.

My goal is to have a pack that will be durable. I'm never going to use the full capacity of the pack and will charge it to 80% most of the time.

Some advice, maybe other cells like Samsung 29E ??
 
Back
Top