Making a Battery Pack using NCR18650

Hi All,

In reading up about battery issues, I've decided to skip buying a pack from some far away land that may or may not be what I think I'm buying and which will, either way, cost me a lot of money to ship. ...I'm very handy, so why the heck not?! Shipping only a few cells at a time is any easy way to help keep the cost / hazard down and I can get them in the USA, I presume, perhaps direct from Panasonic through the company I work for (a scientific tools company).

I have At Least Two fundamental Questions here:

1) Cell count to make a "48V pack"? And;

2) What pre-made enclosures may be available to put my own cells into? Mounting to a tube features?

So, Panasonic has set the standard, apparently, they NCR18650. I found the standard specification sheet:
https://na.industrial.panasonic.com/sites/default/pidsa/files/ncr18650.pdf

And I've ALSO found three other specification sheets for similar batteries -
Higher density of storage, 3.07Ah typical vs 2.9Ah typical
https://na.industrial.panasonic.com/sites/default/pidsa/files/ncr18650a.pdf
And this one has much better cold weather service:
https://na.industrial.panasonic.com/sites/default/pidsa/files/ncr18650f.pdf

But then, there's this one, also purported to be by Panasonic:
https://www.imrbatteries.com/content/sanyo_ncr18650ga.pdf
(Note that it's NOT on the Panasonic web site! The Panasonic site doesn't seem to know about this one but the data sheet has the Panasonic name on it. I suspect this one is from the Elon Musk / Tesla project, but I'm not sure. I will be calling Panasonic to ask.)

NOTE THE GREAT INCREASE IN STORAGE CAPACITY with the NCR18650GA! ... The nominal NRC18650 is around 2.9Ah for a typical cell and this one is up at 3450! That's quite an increase and obviously the choice to be made if they can be had!

All four types of cell have the same nominal voltage and very close to identical dimensions. And all have 3.6V nominal voltage. But how many makes a "48V pack". The math doesn't work out. You'd need 13 1/3 of them. So, considering charging voltages of about 4.2V max, a standing voltage of about 4.1 on a freshly charged cell, I figured the actual number was 12 since 12 X 4 = 48. However, I found this interesting set of comments:

https://electricbike-blog.com/2016/03/18/is-it-wise-to-power-my-48-volt-mid-drive-system-with-a-52-volt-battery-master-yoda/

Here's a pertinent excerpt that has me very puzzled:

"First, you should figure out what battery you have, as you might think you have a 48v battery that is really a 52v battery. When your battery pack is fresh off the charger just hook it up to a volt meter. If it reads around 54v then it is a true 48v battery, if it reads 58v or 59v then you have a 52v battery. LifePo4 batteries often come in a 16S (a string of 16 cells in series) configuration which equals about 52v volts nominal, although these batteries for whatever reason are often sold on the Chinese ebike market as 48v batteries. If you are running LifePo4 a ‘real’ 48v pack will only be a 15S configuration. 18650 cells that are 52v run in a 14S pack configuration although some Chinese vendors (like Paul) call these 50v packs for whatever reason, further adding to the confusion. ‘Real’ 48v 18650 packs are sold in a 13S configuration with a BMS that will be around 54.4v fresh off the charger. If you are running HK lipos they are about the same voltage as the 18650 cells, although there are people who run their BBS02 systems off a 12S system (44v nominal) , I really do not recommend it."

I'm very grateful to Karl Gesslein for writing that article, but now I'm also very puzzled... He's saying the 12 cell stack is "44v nominal." Hmmm... So, can anyone confirm his assertions? Again, he's saying:

12 cells = "44V not recommended".
13 cells = " 'Real' 48V", freshly charged = 54.4V
14 cells = "50V" he says 52V, presumably freshly charged, "sometimes called 50v"
15 cells = "real 48V", freshly charged = 54V
16 cells = "52V", freshly charged = 58V / 59V, called "48V" by Chinese for some unknown reason.

What A Mess!

Will the REAL 48V pack please stand up! (Flashback to "Truth Or Consequences")

In looking at the curves, I figure 3.0 is about the bottom cell voltage anyone could think much about - it's just about dead at that point. So, lets look at what the math tells us, first at the nominal of 3.6 they all claim to have and then their fully charged peak, and finally their dead voltage:

12 cells = 43.2, 49.2, 36
13 cells = 46.8, 53.3, 39
14 cells = 50.4, 57.4, 42
15 cells = 54.0, 61.5, 45
16 cells = 57.6, 65.5, 48

So... it looks like 12 cells is the right number if you look at the fully charged value and 16 if you look at the completely flat value, and 14 is in the dead center between them, so I can see why that seems like the right call, so maybe Karl was right? Anyone know for sure?

By the way, straight off the charger, batteries have a capacitance charge and will drop slightly after a little bit of time - they self-discharge back to their true peak - and THAT is your true target charge voltage. A 14 cell stack should be charged at 58.8v when flat and as the charging current drops off (as it gets full), the voltage should be dropped to 57.4. This will avoid "cooking" the batteries. Adjust accordingly; per cell charging voltage is 4.2 until topping off when it should be 4.1. ... I know a lot about batteries, just nothing about them in the context of bikes!

Now, back to packaging. Anybody know of any available packages to do this with? Presuming 14 cell stacks are correct, I'm thinking two stacks would be great. If the standard MCR18650, storage should be about 5.8 Ah, and with the 18650GA, it should get you a whopping 6.9 Ah per pack! Multiples of those would be easy to imagine... But I'd like to have a standardized attachment to the bike, plug in for charger and bike use, etc...

TIA,
RTIII

It looks like the Lunacycle people think the right number is 13, and that's also what someone has reported is the maximum number of cells the 48v TSDZ2 will take (my unit).

Here's a chart from the Lunacycle people:

https://cdn1.bigcommerce.com/server800/9vkjq73s/product_images/uploaded_images/48v-nominal.jpg?t=1467500767

According to them, dead is a bit higher than 3.0 volts! They say about 3.23 v per cell is the lower limit...

They ALSO say - and I heartily agree - that if you care about your bateries, the never-drop-below voltage for these batteries is (at least!) 20% above that, or 44.2v for a nominally "48v" pack (13S4P).

Go to Amazon buy Micah Tolls book.
Read it cover to cover.

https://www.amazon.com/dp/B06XRKD15B/re ... TF8&btkr=1

Use the sear h function here. Its all been asked and answered before. More than once.

Luna sells stuff start with learning from someone with nothing to sell.

tomjasz said:

Go to Amazon buy Micah Tolls book.
Read it cover to cover.

https://www.amazon.com/dp/B06XRKD15B/re ... TF8&btkr=1

Click to expand...

Thanks for the book link.

tomjasz said:

Use the sear h function here. Its all been asked and answered before. More than once.

Luna sells stuff start with learning from someone with nothing to sell.

Click to expand...

I have been searching, just not focused on here; another good tip to focus searches here. Got it.

And, I am starting by learning from someone with nothing to sell - the people at this site! ... Micah's book is, however, for sale! :lol:

The book was 99 cents for awhile and many ES members grabbed a copy electronically.

If you have access to a tab welder and good soldering equipment then you'll certainly save some money. Otherwise the cost adds up very quickly. Consider that you'll need a BMS or at least some way to access to balance the groups, keep in mind you'll have to buy little things like kapton tape, nickel strip, wires, heatshrink tubing, material to build an enclosure.

t_tberg said:

If you have access to a tab welder and good soldering equipment then you'll certainly save some money. Otherwise the cost adds up very quickly. Consider that you'll need a BMS or at least some way to access to balance the groups, keep in mind you'll have to buy little things like kapton tape, nickel strip, wires, heatshrink tubing, material to build an enclosure.

Click to expand...

AMEN

RTIII said:

In looking at the curves, I figure 3.0 is about the bottom cell voltage anyone could think much about - it's just about dead at that point. So, lets look at what the math tells us, first at the nominal of 3.6 they all claim to have and then their fully charged peak, and finally their dead voltage:

12 cells = 43.2, 49.2, 36
13 cells = 46.8, 53.3, 39
14 cells = 50.4, 57.4, 42
15 cells = 54.0, 61.5, 45
16 cells = 57.6, 65.5, 48

Click to expand...

The industry has mostly standardized on a 36V battery being 10 18650's in series ,10S, and a 48V battery is 13 in series. The 52V battery will be 14S.

The rated capacity of an 18650 cell is based on using it from a fully charged 4.2 volts down to 2.5 volts. It's not dead at 3.0 volts, but manufacturers have set the low voltage cutoff of the battery and the motor controller to about 3.0 volts. The result is that you won't ever get to use the full AH rating of a battery.

Also, the AH rating of the finished battery is the sum of the ratings of the cells in each parallel group, not the number of groups in series. A 36V and a 48V battery, each using the same cell type can have the same AH rating if they use the same number of cells in parallel, i.e, 36V 10AH and 48V 10AH result from using a 10S-4P or a 13S-4P, if the 4P is four 2.5AH cells in parallel.

With the GA cells, one can make a 10AH battery with only three cells in parallel, but it won't last as long as the 10AH battery using four 2.5AH cells in parallel. Less current needed per cell makes for longer life. Sure, one cell by itself can push 10 amps, but not for very many cycles.

Go big on your P arrangement.

Batteries can last if used within limits. Be reasonable. People want to use a 10amp 18650 at 10amp for 3,500mah ? Or 4p for 40amp controller ? Need 8p.

docw009 said:

RTIII said:

In looking at the curves, I figure 3.0 is about the bottom cell voltage anyone could think much about - it's just about dead at that point.

Click to expand...

ll is based on using it from a fully charged 4.2 volts down to 2.5 volts. It's not dead at 3.0 volts, but manufacturers have set the low voltage cutoff of the battery and the motor controller to about 3.0 volts.

Click to expand...

When I said 3.0v is "the bottom cell voltage anyone could think much about" I should have used the word SHOULD instead of the word COULD; I've studied the specifications for at least five different 18650 cells and for all but one they're dead for all practical purposes by that point, the exception being the LG MJ1 series made in Michigan. (There are more cells out there than this, but I haven't read ALL the spec sheets, only some of the more popular cells.) On the MJ1, they're the ONLY ONE that calls out a discharge to 2.5v AND they're the industry leader right now in terms of storage. It's my conjecture they get that #1 ranking by making a cell that will SURVIVE the 2.5v deep-discharge situation. IOW, that benefit is perhaps illusory (as you intimate but with other reasoning). As you point out, you probably can't get to that level of discharge anyway, so buying those cells may be not worth the premium price if you intend to use it, though could be worth it in terms of battery longevity.

Note also that 4.2v is NOT "fully charged" that's the CHARGING voltage. Cells have some capacitance so they'll look like 4.2v when freshly (and fully) charged but that's what's known as the "float" voltage and it will fairly quickly dissipate. Lithium ion batteries, by their chemistry, are completely full at 4.1v. This is very clear from the literature. They all have this trait.

Different versions of the 18650 have different charging criteria - which is unfortunate. The TYPICAL characteristic for charging requires 4.2 v to charge when deeply discharged and then later in the charging cycle the voltage should be dropped to 4.1v per cell. Again, this is all very clear from the specification sheets for the various cells - and again, they're not all identical in this regard. This difference among cells in method of charging is likely responsible for the mythology found on this topic; they DO survive a lot longer when not "cooked" by their charger implementing the wrong charging regime! But this stuff about an 80% or 90% charge is, in my view, based not on a real difference in charge, but based on a misunderstanding of the manufacturer's literature. It would be great to have a side-by-side-by-side experiment with one set of cells treated according to specification and another set treated in the standard way and a third and maybe fourth set managed to the 90 and 80% strategies, but of course this is not realistically possible for most of us as their service lives will necessarily be different. However, I'd bet you'd find that there's no practical difference between cells charged properly, according to their manufacturer's directions, and the 80 or 90% crowd for the simple reason that the 80 or 90% strategy is a poor-man's way of not over-charging as a stupid-charger will do.

BTW, this is why I'm investing in this charger:

http://www.ebikes.ca/documents/Satiator_Manual_V1.0FW.pdf

So far as I have seen, this is the only charger on the market like it! With this one, you can get the charging profile that your battery actually needs.

You make or you buy the first time you open it up put some sense wires on it for a cell meter so you get a check and know each group of cells voltages with 6s -8s meter (common) . I have the Need to know. A problem to avoid a problem ? That's with a bms. Helps with problems.