Tabless design cylindrical cells tests

[Pajda]

EVE 40PL sample:


EVE 40PL disassembly





BAK 45D sample:





Test setup:
- Samples was dressed into standard shrink tube with plastic isolator on the positive terminal.
- ca 5.5 mm^2 cross-section copper strips (11 mm wide with a thickness of 0.5 mm) were welded with the use of Micro-TIG welder to all three competitors (Molicel P30B, P45B and EVE 40PL)
- 6 mm^2 (power) and 0.5mm^2 (sense) Cu wires, ended witch cable lugs with M5 screw, were used for 4-Wire connection with the battery tester.
- Only EVE 40PL is shown as example.



Data Table:


Cycle life Tabless design vs. Tab design from Molicel & Samsung
Results after 1000 cycles at 5C discharge at 100% DOD



Graph below shows dT = Tend - Tstart (The difference between the cell body temperature at the beginning of the discharge which is RT: 20°C with up to 3°C tolerance and the body temperature reached at the end of the discharge). As a Temperature sensor is used NEWARE T-type miniature (0.2mm) thermocouple, attached with kapton tape.



Results after 1000 cycles at 1C discharge at 90% DOD (4.15-3.00V)



Results for 20 min Fast charge 3C-3C rate at 100% DOD



Discharge Voltage charts
EVE and BAK tabless cells vs. Molicel tab cells

 

[Pajda]

Disclaimer: The tested EVE INR4695P sample is most likely still a pre-production sample. This one has been disassembled from the customer battery pack. I only got this sample from Vapcell Company

Cycle life EVE 4680 vs 4695

 

[Pajda]

Disclaimer: The tested EVE INR4680 sample is most likely still a pre-production sample. This one has been disassembled from the customer battery pack. I only got this sample from Vapcell Company

 

[BlueSwordM]

How did you manage to get the EVE 40PL?

I've been trying to get my hands on it for months at this point.

Still, I'm very excited about such a cell. It seems like it's the tabless(foil all tab) design that's giving it the edge, and not the chemistry all that much honestly.

 

[riba2233]

Wow I am so excited for this, thanks! I would be so happy to see you test murata VX40 cells, they should perform even better than these EVEs imho.

 

[BlueSwordM]

Yeah. As it turns out, the 100A claim by EVE might not be feasible without liquid cooling.
One thing I've noticed though is that beyond 5C, the capacity loss becomes much smaller, indicating that the hotter cell has much lower internal resistance.

Wow I am so excited for this, thanks! I would be so happy to see you test murata VX40 cells, they should perform even better than these EVEs imho.

This may indicate that Murata may have an even more capable electrolyte/cathode mix that would enable even higher performance.

 

[Pajda]

I already ran the tests at 100A continuos with 80°C Tcut at the cell body side for both 21700 competitors. I hope that full data with complete test setting will be available this weekend but no promise. In short:

40PL runtime at 100A is: AC (without forced air cooling) @ 62 sec and FC (forced air cooling) @ 86 sec, both to the 80 °C Tcut
P45B runtime at 100A is: AC (without forced aircooling ) @ 44 sec and FC (forced air cooling) @ 58 sec , both to the 80 °C Tcut

40PL manage to stay above 3.00V at 100A load, P45B above 2.78V

 

[spinningmagnets]

I am a huge fan of the tabless design. Of course the focus right now is on using these for cars, but I feel that one of the most important characteristics of the design is that it reduces localized hot-spots in the common tabs, which in the past have acted as bottle-necks to current.

Heat is one of the factors that decays the range of any battery chemistry. The tabless cell design significantly reduces one of the physical causes of hot spots inside cells.

I have not yet seen an example of a tabless LiFePO4 cell, but I feel that the tabless design will allow tabless LiFePO4 cells to produce more current, and to last even longer than they do now.

 

[BlueSwordM]

Yes, and all-tab foils (tabless designs) also have the unintended side effect of improving cycle life even at lower currents since it improves internal cell thermal homogeneity and initial build stresses. This can be seen even in non classical foil tab designs:

Increase of Cycling Stability in Pilot-Scale 21700 Format Li-Ion Cells by Foil Tab Design

Li-ion cells of the industrially-relevant 21700 format were built on pilot-scale with tabs made from (a) the electrodes’ current collecting foils (Al and Cu, “foil tabs”) in comparison with (b) conventional tabs (Al and Ni) welded to the electrodes’ current collecting foils (“welded tabs”). Both...

www.mdpi.com

IMO, the biggest benefits of tabless designs will be seen in HE cells where tabs are sacrificed for maximum energy density. With our current cathode chemitries, just getting the current out efficiently is a big bottleneck, so decreasing resistance there has huge benefits.

 

[riba2233]

I am a huge fan of the tabless design. Of course the focus right now is on using these for cars, but I feel that one of the most important characteristics of the design is that it reduces localized hot-spots in the common tabs, which in the past have acted as bottle-necks to current.

Heat is one of the factors that decays the range of any battery chemistry. The tabless cell design significantly reduces one of the physical causes of hot spots inside cells.

I have not yet seen an example of a tabless LiFePO4 cell, but I feel that the tabless design will allow tabless LiFePO4 cells to produce more current, and to last even longer than they do now.

We are also seeing them in some tools packs, so they will leak to DIY market relatively soon

 

[Pajda]

added small update

 

[Pajda]

added small update

Unfortunately, the EVE 40PL sample I received shows a very poor cycle life.

 

[From-A-To-B]

I'm just absolutely floored that you can rip 100A out of a 21700-size cell in the first place. It's additionally surprising to me that the cells held up as well as they did (even for a single cycle!) in your 12.5C tests.

Time and again, Pajda, you find new and creative ways to put your tools to work and perform really interesting tests. Thank you for sharing your findings with ES.

 

[BlueSwordM]

Oof, very poor cycle life.
No wonder it's pre production

 

[riba2233]

Wow yeah that doesn't look good.

Now I want to see the vx40 even more lol

 

[Pajda]

Update for EVE4695P, this tested sample does not deserve a "P" like performance. Parametrically it belongs to the HE category. Cyclic life is not very impressive either.

 

[Pajda]

One more thought. Data from EVE 4680 sample will be available later, because unfortunately Vapcell this time messed the welding process of nickel strips. They they made the schoolboy mistake when used CD welder rather than laser used on 4695 sample, where CD welded nickel has no chance to properly stick on the aluminium positive terminal. So I am now waiting for my partners with laser welding machine to process aluminium strips.

Before got the EVE4695P results I was thinking about using of something like a 1mm thick and 15mm wide laser welded aluminum strip (15mm is the diameter of the Al positive terminal on 4680) = 15mm^2 which could be equivalent to circa 10mm^2 Cu to go up to 10C (250A), but now I am expecting similar results here as well.

 

[Pajda]

small update, Molicel P30B finished 1000cycles and added table data for both Molicels, hope I will add discharge graphs soon as well. EVE40PL is still alive but not for long.

 

[riba2233]

small update, Molicel P30B finished 1000cycles and added table data for both Molicels, hope I will add discharge graphs soon as well. EVE40PL is still alive but not for long.

Thanks, P30B doesn't look bad at all, 75% remaining capacity after 1000 cycles at 4.2v charge and 5C discharge.

 

[ray88]

Thanks, P30B doesn't look bad at all, 75% remaining capacity after 1000 cycles at 4.2v charge and 5C discharge.

How are Chinese cells doing ?

 

[riba2233]

Molicel are from china, but they make cells better than basically all other manufacturers from Japan or Korea. Really incredible stuff. They are a class for themselves.

 

[BlueSwordM]

Nah, Molicel is a Taiwanese company, not Chinese.

 

[riba2233]

Nah, Molicel is a Taiwanese company, not Chinese.

True! Although Taiwan calls themselves "Republic of China" but I get the point

 

[ray88]

It’s unfortunate that Vapcell encountered issues with the welding process for the nickel strips on the EVE 4680 sample. Using a CD welder instead of a laser welder can indeed lead to suboptimal results, especially when attaching nickel to aluminum. Waiting for your partners with a laser welding machine to process the aluminum strips sounds like a prudent approach. Laser welding tends to provide better precision and reliability.

Try jewellery welder and maybe cooper will do better ..

 

[Pajda]

The EVE 4680 sample cell thing has been postponed indefinitely, but recently my colleague tuned up Micro-TIG and somehow the Cu busbar on aluminium is holding up. Cu with Al is not the best combination, but it will do for tests. The nominal capacity of the EVE 4680 sample is only 22Ah out of 26Ah declared, which cannot be seen as anything other than disappointing. Let's see how they cycle. A competitor in the EVE 4695 format has already completed 1000 cycle and hope I will update the table and graph soon