Looking at building a new pack with Samsung 50E

dukestar

1 mW
Joined
May 31, 2017
Messages
17
I have a 20s7p 30Q pack. Am at 250 cycles since 06/2017. I typically charge to 80.1V (95%). I'd say 90% of the 250 cycles were to 80.1V. I'd typically use around 12Ah per ride. I'd guess I could not get more than 15Ah out of this pack now. I also notice the voltage drop seems to be more severe. I'd say from a 80.1 charge if I draw ~40A it drops to say 70V but bounces back if I get of the throttle. So I think this pack will last me this summer but it's time to consider building a new one.

I'm looking at building a new pack with 21700 cells and am wondering if the 50E would be a good choice. My concern is the max 9.8A rating. I max draw around 65A for short periods so I'm wondering if the 50E may be rated too low. I really like the idea of having a 35Ah pack but wondering if the 50E will not last too many cycles if I'm pulling close to the 9.8A per cell. Maybe the 40T is a better choice but at 30A that's way more than I need.

Any comments or suggestions?
 
How many Samsung 50Es in parallel were you considering? Take a look at this site to give you an idea how they would perform vs 30Qs or 40Ts etc...:
https://lygte-info.dk/review/batteries2012/Common18650comparator.php
 
Same as current pack 20s7p so 7. So max A would be 68.6. So my max of 65 is imo stressing these cells and hence my concern. Thanks for the link I'll check it out.
 
Short power peaks up to 10 sec actually does not have significant effect on the cell life. Most important is continuous or average charge and discharge current. 65A peak is only 2C rate for 7p pack of 50E and this short peak rate is actually no problem for almost any cell. The question is your average current. If it is lower than 1C, the 50E is today most probably the best choice in 21700 size in term of cycle life. Samsung 40T has excellent DCIR but it suffers a lot with poor cycle life (but it is better than 30Q).
 
Thanks Pajda...I'd say on a typical ride I average ~20A so the 50E may actually be a good choice after all.
 
dukestar said:
I have a 20s7p 30Q pack. Am at 250 cycles since 06/2017. I typically charge to 80.1V (95%). I'd say 90% of the 250 cycles were to 80.1V. I'd typically use around 12Ah per ride. I'd guess I could not get more than 15Ah out of this pack now. I also notice the voltage drop seems to be more severe. I'd say from a 80.1 charge if I draw ~40A it drops to say 70V but bounces back if I get of the throttle. So I think this pack will last me this summer but it's time to consider building a new one.

I'm looking at building a new pack with 21700 cells and am wondering if the 50E would be a good choice. My concern is the max 9.8A rating. I max draw around 65A for short periods so I'm wondering if the 50E may be rated too low. I really like the idea of having a 35Ah pack but wondering if the 50E will not last too many cycles if I'm pulling close to the 9.8A per cell. Maybe the 40T is a better choice but at 30A that's way more than I need.

Any comments or suggestions?


1. Unless you plan to run the cells at 0.2C on your eBike... Forget it you wont be getting anywhere close to 35Ah.

2. If you gonna pull 65A on a 7P setup. Forget the 50E cell setup... I personally would consider this battery only if my eBike would max out pulling 30A Max... Otherwise you battery will age very prematurely.

3. The samsung 40T is a way better candidate!! People, stop say you dont kneed all this power. Battery amp rating is a discrete value. ex:max 30A. But in reality it's not about a threashold value ... Ot a continous gradient value: the lower the internal resistance of the cell, the longer they will live for a same given amount of load. The manufacture amp rating gives you sort of an idea of the internal resistance. The higher the amp rating, the lower the internal resistance, thus they cooler it runs and thr longer they live.... You don't HAVE to run cells at 30A even though they are CAPABLE of doing it continously... But THEY will just greatly outlast the 50E cells for ebike application.

I would not even consider just thinking about using 50E cells for ebike applications. They just seem too weak to me. Nobody will ever exctract 5000mAh out of these cella unless they run a 15A controller an a 100P pack which is just not practical...
 
Thanks Matador. This was my concern with the 50E, that the max of 9.8A is too low for my application. The 40T looks like the 21700 version that is closest to the 30Q.

The issue I have with the 30Q is that I didn't get the cycles I was expecting even though I charged it to 90% most of the time. I want to avoid that with the new cells that I choose.

Are there any other 21700 cells I should consider?

Thanks again...
 
I agree go for 40T cells.
I am going to use them for my next battery pack.
At the moment I have a 20S-9P made by 30Q cells, still works but it sags a lot and my concern is that it doesn't have lot of cycles. Generally speaking 30Qs cycle life is very poor.
let's hope 40T will last longer.
So there is not a reason to go for weaker cells this time.
There is one more high power, high quality cell with good cycle life but this is 20700 and it has only 3500mah. If you add more P its a possible solution.

https://www.nkon.nl/rechargeable/21700-20700-size/sanyo-ncr2070c-3500mah-30a.html

you can take a look.
 
@icherouveim, do you know how likely the 40T cycle life will be similar to the 30Q's?

I'm not happy with how my 30Q pack lasted. I definitely don't want to build a new pack with cells that have a cycle life profile similar to the 30Q.
 
dukestar said:
@icherouveim, do you know how likely the 40T cycle life will be similar to the 30Q's?

I'm not happy with how my 30Q pack lasted. I definitely don't want to build a new pack with cells that have a cycle life profile similar to the 30Q.
With simple maths:
For 30Qs you have maximum named current 7P x 15A = 105A
For 40Ts 7P x 25A (from real tests) = 175A
So for exactly the same setup 40T cells will stress less they will run cooler and they are going to live longer.
 
If I recall correctly, the 40T are rated 35A continuous.
If there is a temperature cutoff, they are then rated 45A continuous.

The 30Q are rated for 15A...

An 18650 has a volume of 16.54 mL, whereas a 21700 has a volume of 24.24 mL...
The 40T cell has a volumetric power density of 1.444 A/mL
By comparison, the 30Q cell only has a volumetric power density of 0.907 A/mL

An 18650 has a mass of 48g, whereas a 21700 has a mass of 67g...
The 40T cell has a gravimetric power density of 0.5224 A/g
By comparison, the 30Q cell only has a gravimetric power density of 0.3125 A/g

So really, with all other things being equal (if the 30Q was scaled up to the size of a 40T cell), the 40T seems much more power dense then the 30Q cells.
 
@icherouveim and @matador thanks for the replies...looks like I'll build a 20s7p 40T pack...50E off the table.

I'm starting to gather the parts to build it. I got a JP spot welder, BMS, nickel strip, shrink wrap, glue gun etc, just need the cells now.

Will post my results when done. Probably won't be until end of summer.

Thanks again...
 
dukestar said:

Wow this ebike looks fantastic! I have some questions.
Is this a conversion ? you have matched the rim color and everything so nicely, it looks complete!
The thing with the sparks above the motor is it just for design or it does something ?
Is the battery pack into the gas tank ? Are you sure you can fit 20S-7P 21700 there ?
 
I hooked up with Pat at Sportsman Flyer (https://sportsmanflyer.com/) and worked with him to put together an electric version of the gas repros he makes. The ebikes aren't on his website but are available if you reach out to him. He has tons of pics on his Flickr here (https://www.flickr.com/photos/sportsmanflyer/). You'll see many ebike builds buried in amongst all his pics.

Battery in tank, controller in toolbox. Motor is a big block from LightningRods. The V-twin with spark plugs, pipes and carb is for show.

I wanted a bike that looked like the old Harley/Indian board track racers. In one of the V-twin cylinders I have a buck converter to step down the voltage and run my lights.

4000 miles and all I've needed to do is oil the chain every once in a while and charge and go. Burn through brake pads every 500 miles and rear tire every 1000 miles. And now a battery...lol...
 
dukestar said:
I hooked up with Pat at Sportsman Flyer (https://sportsmanflyer.com/) and worked with him to put together an electric version of the gas repros he makes. The ebikes aren't on his website but are available if you reach out to him. He has tons of pics on his Flickr here (https://www.flickr.com/photos/sportsmanflyer/). You'll see many ebike builds buried in amongst all his pics.

Battery in tank, controller in toolbox. Motor is a big block from LightningRods. The V-twin with spark plugs, pipes and carb is for show.

I wanted a bike that looked like the old Harley/Indian board track racers. In one of the V-twin cylinders I have a buck converter to step down the voltage and run my lights.

4000 miles and all I've needed to do is oil the chain every once in a while and charge and go. Burn through brake pads every 500 miles and rear tire every 1000 miles. And now a battery...lol...

Very nice job! if you use ebike brakes like tektro dorado the ones with the thicker discs you won't have to replace pads so often. Also regenerative braking helps to reduce the problem.
 
After hundred of thousands of cycles on modern small cylindrical cells I can say, that I did not found direct relationship between DCIR value and cycle/calendar life. In fact it is very problematic to find any general rule when considering parameters of new/unknown cell. It is always necessary to test the cell to be sure about its cycle/calendar life.

There are cells with high initial DCIR (>40mohm in 18650) which is growing rapidly during cycling/age, but they have excellent cycle/calendar life (ca 85% of initial capacity after 1000 cycles at 100% DoD)
There are cells with low initial DCIR (<30mohm in 18650) which remains almost constant during cycling/age, but the cells suffer with high capacity loss during use (ca 60% of initial capacity after 1000 cycles at 100% DoD)
.. and you can always find few cells with low DCIR with high cycle life and a lot of cells with high DCIR combined with low cycle life :wink:

If we talk about a specific cells. 40T and 30Q definitelly belongs to the 2. category. They lose their capacity almost regardless on load. Both are designed for continuous operation at 1.3C charge and 5C discharge and so the are best suited for powertools. 30Q has shitty cycle life (only ca 400cycles no matter when discharged at 1C or 3C continuous, there is also no benefit when the 30Q is charged to 4.1V, the capacity loss is the same as for 4.2V charging) 40T is significntly better than 30Q and can do 1000cycles with 63% of initial capacity at 100% DoD.

But for the comparison 50E will do 1000cycles with 85% of its initial capacity and so it totally outperform 40T at 100% DoD in the cycle life if your average load is lower than 1C and you do not need faster charging that 0.5C. So I can only repeat that 50E is actually the best cell for traction application in 21700 size. LG cells excells in 18650 size, but in 21700 their actual production has poor cycle life and I still did not find reason why.
 
But for the comparison 50E will do 1000cycles with 85% of its initial capacity and so it totally outperform 40T at 100% DoD in the cycle life if your average load is lower than 1C and you do not need faster charging that 0.5C.
@pajda On a typical ride I use say 15A in 60min so I average is 15Ah or around .25C if using 50E in a 7p config. However I'll have hill climbs where I pull 50A for say 1 minute or around 1.5C. Do you think the 50E could handle this?
 
Came across this cell: Lishen LR2170SF 13.5A 4500mAh 21700. Does anyone have any comments on them? Looks like a possible compromise between the 40T and 50E. Haven't seen much on Lishen cells so not sure if they are legit or not.
 
dukestar said:
But for the comparison 50E will do 1000cycles with 85% of its initial capacity and so it totally outperform 40T at 100% DoD in the cycle life if your average load is lower than 1C and you do not need faster charging that 0.5C.
@pajda On a typical ride I use say 15A in 60min so I average is 15Ah or around .25C if using 50E in a 7p config. However I'll have hill climbs where I pull 50A for say 1 minute or around 1.5C. Do you think the 50E could handle this?

There is no significant problem with cycle life of high quality HE cells even at higher continuous discharge rate than 1C but you have to check the temperature inside the pack. 50E have in its datasheet maximum continuous discharge 2C (9800mA) and 3C (14700mA) not for continuous discharge. I recently added torture test to my cycle life database with 3C continuous discharge and 50E did very well (92% of its nominal capacity after 150cycles). Of course cell is going hot as hell on free air, that means that it can easily toast itself inside the pack if there is no temperature sense.
 
As I previously mentioned it is almost impossible to give relevant information about cycle life of unknown cell without proper measurement. Lishen is big cell producer in China and I think that Tesla asked them about their 21700 cell production for using it in TM3, produced here. But still for me it means nothing and this particular cell is not interesting for me right now.

I am more focused on LG INR21700H40 as a direct (and slightly cheaper) comepetitor to Samsung 40T.
 
Pajda said:
... I am more focused on LG INR21700H40 as a direct (and slightly cheaper) comepetitor to Samsung 40T.

That should be an interesting cell. Have you found any datasheet's or discharge graphs for that cell? What price did you see for the H40?
 
pwd said:
Pajda said:
... I am more focused on LG INR21700H40 as a direct (and slightly cheaper) comepetitor to Samsung 40T.

That should be an interesting cell. Have you found any datasheet's or discharge graphs for that cell? What price did you see for the H40?

No I did not already find any datasheet or graphs of LG INR21700H40 (but I give only 5sec of my time for googling it). I found this cell listed in QeenBattery price list for 3.9USD, where the INR21700-40T is listed for 4.5USD. But you should count with significant shipping cost from QB and there is also minimum order amount.
 
Back
Top