EV West Too Good to be True? 50¢ per cell, new

[wheelbarrow]

[UPDATE: The modules are cheap because the cell's tech is from 2011, and the capacity per cell is not that large. Could still be a decent performer but with a bad energy to weight/size ratio. Worth it for $0.50 per cell IMO. Test results of these cells linked by docw009 (reply #8): Test of Samsung INR18650-20R 2000mAh (Green). Going with a JK 150 or 200A bms to keep my options open, only $10 more than a 100A version.]

Hey there! I have some questions before I buy my first battery. Your answers are hugely helpful.

1. Are there any obvious downsides to the battery module from EV West below? Samsung INR 13S 54 Volt 20R. (Link at the end of this paragraph.)
   
   
2. Picking my first BMS, what do you think would suit this battery? I'll connect it to a Fardriver 72V50A controller.
   1. I think active balancing would be great to have, is 0.6A great enough for balancing or should I go for 1 or 2 amps? I hear good things about ANT BMS, but it looks like JK BMS are the ones with active balancing.
   2. Would I be wasting money by purchasing something like the JK-BD6A24S12P rated at 8-24s and 120A for $46.19? Should I just run with a generic $20 40A BMS from Daly or similar?
3. Could I wire up two of these 13s6p modules (10.5A 505 Wh each) in parallel to create a 13s12p battery with a capacity of 21A and ~1,000 Wh? Could I use one BMS for the whole thing, or would it be better for each module to have its own BMS? Would I need a 200A BMS if combining both batteries?

Trying to get the most bang for my buck without buying a cheapo bomb-style downtube battery. If more info is needed, please let me know. Thanks again!

Samsung INR 13S 54 Volt Lithium Ion Battery 20R Scooter and eBike Battery Modules

 

[BlueSwordM]

Hey there! I have some questions before I buy my first battery. Your answers are hugely helpful.

1. Are there any obvious downsides to the battery module from EV West below? Samsung INR 13S 54 Volt 20R. (Link at the end of this paragraph.)
   
   
2. Picking my first BMS, what do you think would suit this battery? I'll connect it to a Fardriver 72V50A controller.
   1. I think active balancing would be great to have, is 0.6A great enough for balancing or should I go for 1 or 2 amps? I hear good things about ANT BMS, but it looks like JK BMS are the ones with active balancing.
   2. Would I be wasting money by purchasing something like the JK-BD6A24S12P rated at 8-24s and 120A for $46.19? Should I just run with a generic $20 40A BMS from Daly or similar?
3. Could I wire up two of these 13s6p modules (10.5A 505 Wh each) in parallel to create a 13s12p battery with a capacity of 21A and ~1,000 Wh? Could I use one BMS for the whole thing, or would it be better for each module to have its own BMS? Would I need a 200A BMS if combining both batteries?

Trying to get the most bang for my buck without buying a cheapo bomb-style downtube battery. If more info is needed, please let me know. Thanks again!

Samsung INR 13S 54 Volt Lithium Ion Battery 20R Scooter and eBike Battery Modules

View attachment 380487

Well, the main downside is that these cells are outdated and that a newly built pack using 3Ah 18650s would contain 50% more energy per volume.

However, at 0.50$USD per cell for a fully built pack, I wouldn't care at all lol.

I'd completely remove the old BMS if you plan to parallel 2 packs together and use a singular high power one.

Furthermore, if you want to connect them in parallel, equalize their voltages before connecting them together.

 

[Nix Liou]

22A, especially in a module, might be a bit of a stretch. It's probably capable of 15A per cell CDR.

The main reason this module is cheap is probably because the cell has been out since 2013 iirc, and is basically obsolete vs modern power cells with much better energy density. These modules might be surplus and or old stock as well since they don't specify build date that I can see. The price seems like a good deal though.

Don't buy a DALY BMS though. If these cells are older, you might want active balancing, otherwise passive is probably okay for new cells. Get a bms capable of the power you want to use, either JK, Ant, or JBD.

 

[wheelbarrow]

Well, the main downside is that these cells are outdated and that a newly built pack using 3Ah 18650s would contain 50% more energy per volume.

However, at 0.50$USD per cell for a fully built pack, I wouldn't care at all lol.

I'd completely remove the old BMS if you plan to parallel 2 packs together and use a singular high power one.

Furthermore, if you want to connect them in parallel, equalize their voltages before connecting them together.

That makes a lot of sense. All the other cheaper (less than $1.50 ea) cells I've found each have continuous discharge ratings of 10A or less, so the 22A rating on these guys should be much better suited for an ebike.

There is no BMS that comes on the pack, so I suppose that could be a downside to some. I'm glad I'm not paying for some cheap built in BMS in the battery's cost and can choose my own superior BMS.

Equalizing the voltages is a good move for sure. Not trying to flash boil my shiny $0.50 cells.

 

[wheelbarrow]

22A, especially in a module, might be a bit of a stretch. It's probably capable of 15A per cell CDR.

The main reason this module is cheap is probably because the cell has been out since 2013 iirc, and is basically obsolete vs modern power cells with much better energy density. These modules might be surplus and or old stock as well since they don't specify build date that I can see. The price seems like a good deal though.

Don't buy a DALY BMS though. If these cells are older, you might want active balancing, otherwise passive is probably okay for new cells. Get a bms capable of the power you want to use, either JK, Ant, or JBD.

Gotcha, old stock is almost guaranteed. What capacities are new 18650s up to these days? I thought only 21700 cells went farther than ~3Ah. I know new cells have insane discharge ratings though.

If each cell could discharge 15A continuously * 6 cells in parallel, that would mean a pack rating of 90 amps, right? In that case, would an 80A BMS be sufficient, or should the BMS have headroom above the cells themselves, say 120A BMS?

 

[Nix Liou]

Gotcha, old stock is almost guaranteed. What capacities are new 18650s up to these days? I thought only 21700 cells went farther than ~3Ah. I know new cells have insane discharge ratings though.

If each cell could discharge 15A continuously * 6 cells in parallel, that would mean a pack rating of 90 amps, right? In that case, would an 80A BMS be sufficient, or should the BMS have headroom above the cells themselves, say 120A BMS?

New 18650 cells go up to 4A, but usually only ~10a discharge from those. 3.2Ah 18650 cells are cheap (EVE 33v), and 3.5Ah ones like the samsung 35e or eve 35e are pretty good prices as well. 21700 cells go up to 6Ah, but tabless 21700 cells are 5Ah and capable of 50A discharge as well (EVE 50PL and Tenpower 50XG are preproduction currently, or the Reliance RS50).

Even LiFePO4 will probably have better gravimetric energy density than these bricks probably.

From old stock, you may have lower than rated capacity, especially depending on how well it was stored. So keep that in mind as well.

The BMS should at least have overdischarge limits based on what the cells can do. A 100A BMS with the max discharge set to 90A would be fine, and an 80A BMS would be fine as well (You would just be limited to 80A).

 

[Nix Liou]

Batemo predicts 7.46 Wh per cell for the 20R while these battery modules only claim 6.5Wh. Even just off 2Ah at 3.6v, you would expect more. Seems like they are only claiming 90% original capacity.

 

[docw009]

Some test results for the 20R cell. It tested as advertised, like all Samsungs do, If you pull 12A on the modufe, that's 2A per cell, and you get a 1 hour run time. That's the rub. You have to balance your run time with the current. Or make the battery big enough so you get both.

If the cells haven't been used, but sat in storage, they probably haven't lost much capacity, I have some leftover Samsung 30Q's that I bought four years ago. They're still testing at the original rated 3000 mah in my tester. Bear in mind they were stored at their original shipping voltage (30% SOC or 3.43 V) too, The untested cells are still holding close to the shipping voltage too.

 

[wheelbarrow]

Those are some impressive specs on new cells. If I were building from raw cells I'd go with tabless 21700 for sure.

Lower than rated capacity makes sense for the age, and I also see how EV West claims 90% of the original capacity. Nice to see they are trying to be transparent.

15A seems like a good idea, the datasheet says they don't live that long at 22A discharge rates:

 

[wheelbarrow]

Some test results for the 20R cell. It tested as advertised, like all Samsungs do, If you pull 12A on the modufe, that's 2A per cell, and you get a 1 hour run time. That's the rub. You have to balance your run time with the current. Or make the battery big enough so you get both.

If the cells haven't been used, but sat in storage, they probably haven't lost much capacity, I have some leftover Samsung 30Q's that I bought four years ago. They're still testing at the original rated 3000 mah in my tester. Bear in mind they were stored at their original shipping voltage (30% SOC or 3.43 V) too, The untested cells are still holding close to the shipping voltage too.

What a great resource, thank you. With a 1kW motor I don't have the intuition of how many amps it'll draw at a set speed or how fast I could get at something like 20A, but your explanation of amps vs discharge time is starting to pull things together.

That's great to hear about your old cells and how they still had life in them. I hope these modules were stored around 3.6-ish volts. 13 years in storage could be rough, there are humans younger than that riding these ebikes around.

The test you linked charged the cells at .5C (1 amp) as the data sheet recommends, and it was shown to have taken ~130 minutes which is much faster than the claimed 180 minutes at that rate on the datasheet. I'm curious as to what the real world charging speed at 2C (4 amps) would be, as the datasheet claims it would take 50 minutes, which seems a little long. I probably wouldn't push them past 1.5C myself considering their age.

Thanks again.

 

[amberwolf]

. With a 1kW motor I don't have the intuition of how many amps it'll draw at a set speed

That depends on your controller, which is the part of the system that converts battery power to motor power, and does any current limiting the system has. So the controller will have a current limit on it, which is the max current your system shouldl be drawing worst-case. Any battery you use must be more capable than that, by enough so that as it ages it will still be at least capable enough to handle that current without being stressed.

or how fast I could get at something like 20A,

You can use the simulators at ebikes.ca and elsewhere to find out how much of what it will take to do a certain thing under your riding conditions. Put in the things you already know, such as voltage, controller current limits, terrain, hill slopes, wind, weight, speed, etc., and you can find the other things, like how many amps it takes to go a certain speed, or how fast it will go when a certain battery (or motor, which will be different) current is being drawn, etc.

 

[Nix Liou]

With a 1kW motor I don't have the intuition of how many amps it'll draw at a set speed or how fast I could get at something like 20A

Well, with a 1kw motor is only going to need 20 or so amps, and you can expect that to get you to 35 mph. Generally a battery should be rated for 125% of the load or more, so 30-40A would be plenty for the whole pack.

These cells may still be a good deal and convenient, but if you wanted to build a ~1KWh battery pack (equivalent of 2 of these modules) with brand new cells, 52 5Ah cells would get you ~940Wh, and most solid new cells can handle 2C discharge rating, so it would be capable of 40A (EVE 50e, Samsung 50e or 50g, BAK N21700cg, LG M50LT, etc). The total weight of those cells would be similar to just one of these modules, at the caveat of costing ~130 (EVE 50e) or so, and needing to be assembled.

I have bought surplus modules (with EVE 33v) and taken them apart for the cells, and as doc says, usually they perform fine. Even after a year or two under good conditions, they may test above the minimum spec. But if you look at that spec sheet, you will notice that the shelf life at raised temperatures drops off a cliff, and the problem with surplus modules and similar is that you don't know if they were stored in good conditions or in a toasty warehouse. Hopefully EV west knows they are >1.5 years old and just put the 90% spec from the sheet on it? They may have tested some and gotten 505Wh from it though. They're likely not 13 years old, at least. Most cells would be cooked at that point I think.

than the claimed 180 minutes at that rate on the datasheet. I'm curious as to what the real world charging speed at 2C (4 amps) would be, as the datasheet claims it would take 50 minutes, which seems a little long. I probably wouldn't push them past 1.5C myself considering their age.

It's 50 minutes because of the constant voltage phase at the end of the charging. I would generally stick to 1C charging on any cells even if they are rated for more, just because charging is the most stressful condition for the cells. Note that many cycle life tests at high discharge still keeps the charge rate down at 1C or even 0.5C

Overall I think it's not too much of an investment to buy em and see. If they turn out to degrade rapidly and die in a month, oh well, get some new cells and reuse the BMS.