Tabless design cylindrical cells tests

[BlueSwordM]

Are there no 5000+mAh tabless cells ? It could be a good compromise to get a low ir so it can take high discharge peaks, but still get a a little extra capacity. For my use in an enduro bike I have short peaks of 40-50A/cell, but the average consumption is more like 4-5A/cell.

Not yet. However, don't let perfection be the enemy of the good.

Get the Molicel P50B and be happy for now.

 

[Pajda]

Updated first and second post Tabless design cylindrical cells tests

- added "Table results" for BAK 45D
- EVE4680 finished 1000 cycles in good order and I will continue with cycling. The chart was moved to the second post.

News:
- BAK 45D started 0.5C-4.44C HP cycling. (the 20A limitation is caused by my new BTS4000 tester, I can let the test run at 5C on an older ZKE, but its data file is not as good for further processing and especially it can't track temperature.)
- BAK 45D started 0.5C-1C @ 4.15-3.00V cycling
- EVE 40PL started 0.5C-5C @ 4.20-2.80V (to check if the poor cycle life is related with the DoD)

 

[BlueSwordM]

Very nice. On the topic of testing, I ordered a pulse driver and a MOSFET switch to test out pulse charging and test out a special circuit I'm working on.

I'll finally be able to start testing batteries with "next-gen" charging algos

 

[riba2233]

Very nice. On the topic of testing, I ordered a pulse driver and a MOSFET switch to test out pulse charging and test out a special circuit I'm working on.

I'll finally be able to start testing batteries with "next-gen" charging algos

nice, be sure to post some results!

 

[BlueSwordM]

For sure. I'll need to find a way to integrate my CC discharger with the charger though to make the testing automated however...

 

[MorbidlyObeseKoala]

For sure. I'll need to find a way to integrate my CC discharger with the charger though to make the testing automated however...

Derailing slightly here but is there much research into different charging types for lithium? I recall some old posts discussing NiMH using pulse charging, curious if there was an explanation out there for the chemistry or physical advantages to that method. Good luck with the project.

 

[BlueSwordM]

Derailing slightly here but is there much research into different charging types for lithium? I recall some old posts discussing NiMH using pulse charging, curious if there was an explanation out there for the chemistry or physical advantages to that method. Good luck with the project.

There are plenty of researched and implemented new charging algorithms.

There's the simple CPCV (Constant Power CV), MS-CCCV(Multi Stage CCCV), and the combined MS-CPCV, which is what most larger electric vehicles use to maximize charging speed while minimizing degradation vs CCCV.

There are more advanced versions of these algorithms like MS-CPCV with an IR bias, cathode monitoring, negative temperature coefficients, etc.

There are also more complex algorithms like optimal high frequency pulse charging and low complexity low frequency pulse charging, since middle of the road frequencies in pulse charging tend to be bad for battery lifetime. Reflex charging is also a thing (pulse charging-short discharge pulse). There has also been integrations with ML algos.

 

[Pajda]

It seems there are eager eaters for the results of the BAK 45D cycle life. So here are the first 50 cycles. So far it looks good, that's about the only thing to say right now. Tabless design cylindrical cells tests

 

[thermal_runaway]

Any chance of the 45D cutaway pictures soon?

 

[riba2233]

wow this cell looks unreal at its price point...

 

[Pajda]

Any chance of the 45D cutaway pictures soon?

I hope to make it today

Btw. BTCAP 40P and Lishen 58SK was sent, so in ca 14 days there can be first results.

 

[Pajda]

BAK 45D disassembly pictures added in the first post Tabless design cylindrical cells tests

 

[BlueSwordM]

BAK 45D disassembly pictures added in the first post Tabless design cylindrical cells tests

Damn it, they cheaped out on the cathode.
Imagine if they had gone with a 9x0.2mm tab

 

[jonb]

little update in the first post Tabless design cylindrical cells tests

Unfortunately, nothing has changed with the 40PL (MP - Mass Production) even in terms of cycle life.
EVE 4680 performs nicely

Why are your results so different to those of mooch? any idea?

 

[Pajda]

Why are your results so different to those of mooch? any idea?

Stay tuned for a few more weeks and we'll have the data for an answer. Today, the first 50 cycles of the EVE 40PL were completed with a 2.80V cut-off voltage setting instead of the standard 2.5V. There seems to be good potential in the yellow trend, at about 250 cycles, we will be almost sure if here is the problem Tabless design cylindrical cells tests

I think Mooch uses a higher cut-off voltage of about 3.20V in his tests.

 

[CamLight]

I think Mooch uses a higher cut-off voltage of about 3.20V in his tests.

Nope…the abusive cycle life testing uses a 2.50V cutoff, mentioned in my post:
Tabless design cylindrical cells tests. After 250 abusive cycles I’m at a run time loss of 7.5%.

The “E-Scores” (delivered Wh) for the conventional performance tests are measured at 3.2V though.

Ampace recommends not welding within a centered 6mm circle on the bottom (essentially, keep away from the spiral) so that can be a consideration. Other manufacturers for conventional “tabbed“ cells recommend staying out of a 3mm circle to avoid the internal weld so the larger contact patch for the internal welds of these “tabless” cells would seem to require a larger ”keep out” area IMO.

 

[BlueSwordM]

Nope…the abusive cycle life testing uses a 2.50V cutoff, mentioned in my post:
Tabless design cylindrical cells tests. After 250 abusive cycles I’m at a run time loss of 7.5%.

The “E-Scores” (delivered Wh) for the conventional performance tests are measured at 3.2V though.

Ampace recommends not welding within a centered 6mm circle on the bottom (essentially, keep away from the spiral) so that can be a consideration. Other manufacturers for conventional “tabbed“ cells recommend staying out of a 3mm circle to avoid the internal weld so the larger contact patch for the internal welds of these “tabless” cells would seem to require a larger ”keep out” area IMO.

Huh, this is new. Is there any literature on the subject on spot welding location affecting cycle life? I'm only seeing papers on the subject of better spot welding locations decreasing interconnect resistance:

Radware Bot Manager Captcha

I've never heard of such a thing being a problem and I wonder why it would be.

I can only infer that the lower melting temperature of the aluminium contacts may be causing issues regarding internal weld contact resistance, so that might be a factor.

 

[CamLight]

Huh, this is new. Is there any literature on the subject on spot welding location affecting cycle life? I'm only seeing papers on the subject of better spot welding locations decreasing interconnect resistance:
Radware Bot Manager Captcha
I've never heard of such a thing being a problem and I wonder why it would be.

I can only infer that the lower melting temperature of the aluminium contacts may be causing issues regarding internal weld contact resistance, so that might be a factor.

My apologies for the delayed response.

I’ve seen spot-welding location issues discussed in research papers (li-ion aging mechanisms) but I don’t remember which or have links. Beyond spot-welding’s degradation of the electrolyte (and deposition onto the SEI of the byproducts, increasing IR) the integrity of the internal weld itself can be affected by the heat of the external weld.

It could affect the thermal or shock/vibration fatigue life of the internal welds or directly impact their resistance. External welds could even completely separate internal welds? If it’s enough heat to weld onto the cell it could be enough heat to separate (or at least weaken) an internal weld.

Ampace has a 6mm diameter keep-out circle. Murata even has two keep-out zones for their VTC5D. The outside ring might be to prevent heating (damage) of an insulating disk?

 

[riba2233]

Wow, and some people still think that soldering is an acceptable practice...

 

[rafalg]

soldering is relatively low temperature, you can't 'unsolder' material that has been spot welded
soldering temp - 350 *C , and temperature needed for spot weld would be probably above 1000 *C (nickel melts at 1400 *C) - so you'd need much more heat to break internal spot welds

 

[BlueSwordM]

My apologies for the delayed response.

I’ve seen spot-welding location issues discussed in research papers (li-ion aging mechanisms) but I don’t remember which or have links. Beyond spot-welding’s degradation of the electrolyte (and deposition onto the SEI of the byproducts, increasing IR) the integrity of the internal weld itself can be affected by the heat of the external weld.

It could affect the thermal or shock/vibration fatigue life of the internal welds or directly impact their resistance. External welds could even completely separate internal welds? If it’s enough heat to weld onto the cell it could be enough heat to separate (or at least weaken) an internal weld.

Ampace has a 6mm diameter keep-out circle. Murata even has two keep-out zones for their VTC5D. The outside ring might be to prevent heating (damage) of an insulating disk?

F it, I'll ask directly at the source: a battery researcher
Hopefully I get a quick answer.

 

[Pajda]

small update in the first post Tabless design cylindrical cells tests

News:
- EVE 40PL (MP) cycled at nominal 0.5C-5C @ 4.20-2.50V was halted at 300 cycles due to poor cycle life.
- BAK 45D started 3C-3C HP cycling.
- EVE 40PL (MP) started at nominal 0.5C-5C @ 4.20-2.50V without welding (bare cell in standard 4W fixture)

Lets definitely check if spot-welding make any difference here. From the actual results after 100 cycles of EVE40PL with 2.8V LVC, I am more inclined to believe that the problem was caused by the LVC set at 2.5V. I think paradoxically the EVE40PL sample in @CamLight torture test could help the fact that it was cycled at higher C-rate than in my 5C-rate dchg test, because higher C-rates leads to the lower actual DoD due to the DCIR effect.

 

[CamLight]

small update in the first post Tabless design cylindrical cells tests

News:
- EVE 40PL (MP) cycled at nominal 0.5C-5C @ 4.20-2.50V was halted at 300 cycles due to poor cycle life.
- BAK 45D started 3C-3C HP cycling.
- EVE 40PL (MP) started at nominal 0.5C-5C @ 4.20-2.50V without welding (bare cell in standard 4W fixture)

Lets definitely check if spot-welding make any difference here. From the actual results after 100 cycles of EVE40PL with 2.8V LVC, I am more inclined to believe that the problem was caused by the LVC set at 2.5V. I think paradoxically the EVE40PL sample in @CamLight torture test could help the fact that it was cycled at higher C-rate than in my 5C-rate dchg test, because higher C-rates leads to the lower actual DoD due to the DCIR effect.

Hmm…interesting possibility. A battle between lithium plating and high temps (from high C rate) vs DoD damage. I’ll wait for your results of the 5C to 2.50V cycling and if it looks like it might be useful I can run 100 of the same cycles to compare results against.

Thank you for doing all this extra testing!

 

[CamLight]

F it, I'll ask directly at the source: a battery researcher
Hopefully I get a quick answer.

Excellent! Has to be some reason to avoid the weld areas though or otherwise the manufacturers wouldn’t have these warnings. Hopefully she/he can give you some great info!

Do you also have access to an applications or test engineer? They’d be the ones who test production cells and/or get the feedback from customers. The cell development labs are often dealing with just coin and pouch test cells

 

[thermal_runaway]

Ampace has a 6mm diameter keep-out circle. Murata even has two keep-out zones for their VTC5D. The outside ring might be to prevent heating (damage) of an insulating disk?

Very likely, same-side laser welds will melt through the insulating disk.