Panny or Sanyo?

[Taswegian]

Thanks Matador for taking the time to lay it out so fully and with precise writing, well done, and this does agree with other information on the topic.

There are so many variables, and of course shoddy quality, in materials purity or manufacturing QA will override these finer distinctions.

So, my conclusion, again, is **if** the use case does not involve high discharge rates, say under .2C

**nor** cycling down to very high DoD, say 70-80%

then better longevity will come from energy-dense / higher capacity cells, rather than those designed for power density / higher discharge rates.

I don't quite follow your reasoning here. Just because you don't abuse high capacity cells with excessive current doesn't mean better longevity than high power cells (which are inherently more resistant to such abuse due to lower IR).

Comparing high capacity with high power, with both types used conservatively within their respective max discharge rates; What is it about high capacity cells that would then yield better cycle life?

Is there something about the typical chemistry or construction of high capacity cells that lends itself to this? I don't see it.

[docware]

Taswegian,
there are numerous degradation factors affecting the Li-ion cell during its life. Read my answer to Matador at previous page. Try to read articles at Electropaedia pages.

The best way to investigate lifetime is long term cycling test, see Pajda´s testing. In real operation is cell burden much more complex (vibration, temperature cycling, …..), but this long term testing is currently best way to compare lifetime od various cells.

[Dak77]

I wish I could give more specific limits and targets. All in all, this will be ridden as a cruise around the town (flat ground to occasional 10% slopes) and occasional grocery getter bike. I realize due to mostly budget that it won't be a pack that will last for 5+ years . I doubt the bike will even be worth keeping much longer than that considering it's already over 12 years old , so if we can get 500+ cycles with decent performance, it will be a win. I just don't want something that will start sagging dramatically after 300 cycles . Without too much to go on , I feel pretty confident that the amp load will stay between 2 and 5 amps per cell , with 5 being short peaks . Also worth mentioning, 80% of the life of the pack will be spent between 4.1v- 3.5v and have good air circulation between each cell.

[Dak77]

Comparison applys to using cells a a givent load.

For a given load (in amps) power dense (High amps or high wattage) cells (aka lower internal resistance cells) generally have longer life because they heat less. But they pack less energy (Wh) density because much of the cell can volume is occupied by a bigger cathode (to allow for higher currents without turning iron red hot). Thus leaving less space for the lithium which give you the range.

For the same given load, more energy dense cells (with higher amp-hours) will have smaller cathode (suitable only for low currents), and more space for lithium within the cell. But that smaller cathode means higher resistance. So for the same given load, thr more energy dense cell will heat up more. Thus lithium chemistry will degrade much faster and cycle life will be LESS than for mor power dense cells.

You are right about the triangle between energy density (long range but limited power and if power rating exceeded, premature failure and lower cycle life), power density (high power cells are the standard for lightweight EVs,) and life cycle.

Also, it is utopic to believe you will ever get 3400 mAh out of a 3400 mAh low drain cell. Run it higher than 0.2C, and the real usable capacity will start to take a plunge. On the other hand, higher density cells will sag less and thus you will be able to extarct much more capacity until you reach the low voltage cutoff threashold of your controller and your BMS (around 3.0v per cell).

Matador

Matador, your explanation is over-simplistic as doesn´t involve numerous degradation modes of li-ion cells. Low temperature issues, parasitic chemical reactions, lithium loss, SEI layer build up , lithium plating, dendritic growth, …….
Here is one of numerous articles at highly recommended Electropaedia pages :

https://www.mpoweruk.com/life.htm

Warning : danger of the information overload !!!


And here another interesting file :

ALLPlus+ Technology.pdf

Regarding your concentration to internal resistance only, here is small comparison of DCIR at 3,8 V, 24 °C :
PF …. 34 miliohm
GA …. 36 – 37 miliohm
29E …. 31 miliohm
As you can see, the DCIR is similar.

Compare please GA versus PF voltage sag at the same load. Equipment : calibrated ZKETECH EBC A-10H, BF-2A fixture, 25 ±1 °C.

GA v PF 1 A b.jpg

GA v PF 2 A b.jpg


IMHO, your conclusions are erratic.

There is also problematic use of the battery in Dak77 case. Mode of intended use will be probably much more as electric scooter than ebike. That means enormous battery load, high DOD, ….. Therefore 13s8p battery cca 1,36 kWh (GA) or 1,1 kWh (PF) may not be enough for good lifetime. Moreover, we haven´t accurate information about real max and average currents.
So 13s10p or more could be more appropriate solution.

Doc, I think your assumption that this will be used more like a scooter than an Ebike is pretty accurate. Just to get her out on a bike required me selling the point that she can mostly cruise on battery power . What can I say , she needs a little motivation to exercise . I have been looking at those same test charts on the GA vs PF and there's one aspect I'm a bit unsure of. That sag it shows in the 2amp discharge graph ; is that due to PF being lower capacity, thus using a higher C rate through it's discharge path? If you made a parallel group that had the same capacity as the GA group, would it still have more sag?

[docware]

10 % slopes is quite steep hill. Hopefully not too long. What do you mean when you say "good air circulation" ?
I understand that you want to motivate your wife to excercise . Maybe after some time you can make throttle „out of order“ to get her more pedalling. If the bike will have PAS sensor, pedalling is nearly without any effort.

I can recommend GA only if peak current per cell is 3 A and average 2 A or less. This is not your case, so your choice is then PF which have good references here at ES.
Yes, PF has worse sag thanks to its lower capacity. Paralel group with the same capacity would probably have the same or similar sag.

[Dak77]

10 % slopes is quite steep hill. Hopefully not too long. What do you mean when you say "good air circulation" ?
I understand that you want to motivate your wife to excercise . Maybe after some time you can make throttle „out of order“ to get her more pedalling. If the bike will have PAS sensor, pedalling is nearly without any effort.

I can recommend GA only if peak current per cell is 3 A and average 2 A or less. This is not your case, so your choice is then PF which have good references here at ES.
Yes, PF has worse sag thanks to its lower capacity. Paralel group with the same capacity would probably have the same or similar sag.

Haha a remote controlled "secret switch" sounds fun . I think peak 3amp would be a bit too optimistic. Probably more like 5 . Obviously, she's going to have to get a little involved on hills. It does have pedal assist. She's no hippo by any means, in fact she's built like a brick...nevermind , but I just want her to get outside and get me off my butt too . I've seen enough from various posters' testing and feedback to decide on PFs . The air circulation is next on my list, but my intent is to have about 2mm spacing between each cell, contained in a hardshell pvc or abs box with a push -> pull fan on opposite ends with their own external lithium pack to power them . Probably 80mm ATX case fans with a speed controller; the push fan on the front and the pull in the rear with the case mounted horizontally under the top tube . I might have to make another hard case under it to take up the space left over and provide some support and maybe light storage . Thanks for all the input. It's been a HUGE help.

[docware]

Do you really want to force all moisture and dust from surroundings to the cells ? Doesn´t sound well !

[Dak77]

Do you really want to force all moisture and dust from surroundings to the cells ? Doesn´t sound well !

Yeh I thought about that. Maybe a shroud with a small crankcase ventilation filter for the dust issue and just pop a baggy with a rubber band around it if you get caught in the rain ? I was thinking something similar to this https://www.monsterscooterparts.com/air ... agQAvD_BwE in principle, but handmade to the needed dimensions with epoxy to seal the joints and rubber seals for the access panel. It's a little more fluff and clutter than necessary , but it will be easily serviceable and it won't get hot inside. It provides a surface to secure the cells
to so they aren't bouncing like a Boa in a wheelbarrow and also a reusable enclosure that can be transferred to another bike .

[amberwolf]

Yes, PF has worse sag thanks to its lower capacity.

Capacity does not determine voltage sag.

Internal resistance determines voltage sag.

[Dak77]

Yes, PF has worse sag thanks to its lower capacity.

Capacity does not determine voltage sag.

Internal resistance determines voltage sag.

I'm glad you said that. So reading those discharge charts; the voltage values shown aren't actually illustrating "sag" correct? I noticed that the lower capacity batteries always droop more on those graphs than the high capacity cells.

[amberwolf]

If the lower capacity batteries are also higher internal resistance, then you'd see more sag, but it's not because of the capacity, it's because of the resistance.

Regarding the discharge curve, that's just what voltage you'll see at a specific state of charge under a specific load. Technically it is also "sag" in that it is voltage drop under load, whcih is what sag is....

[Dak77]

Ok this is the PF . It's supposed to have similar resistance to the 35E .
https://lygte-info.dk/pic/Batteries2012 ... )-info.png

This is the 35E
https://lygte-info.dk/pic/Batteries2012 ... )-info.png

On this chart at 5A it looks like the PF "sags" way more .

[amberwolf]

Ok this is the PF . It's supposed to have similar resistance to the 35E .

I don't see a chart showing the panasonic PF.

I see a chart showing something called "orbitronic PD".

That could be anything, but I doubt it is actually a panasonic PF (even if "orbitronic" says it is, I wouldn't recommend accepting that, and would recommend locating actual known panasonic PF info). It's not unknown for "rebranded" cells to be recycled from other sources, and could even be used (in which case the actual results of any specific cell could


But comparing the two cells they do show there, the one with lower capacity drops in voltage faster over time partly because voltage does drop with SoC, and it has less charge so it has to drop faster.

[Dak77]

It was either Thunderheart or Mooch from the flashlight forum saying it's a PF rewrap.

Edit: This is where I read that. https://lygte-info.dk/review/batteries2012/Orbtronic 18650PD 2900mAh (Black) UK.html

[john61ct]

Capacity does not determine voltage sag.

Sure it does!

Not if the load's C-rate is held constant, proportional to the capacity.

But given the same chemistry and resistance,

put the same 3A load on a 3Ah cell, then put it on a 20A cell, of course the latter will sag less.

Not saying it's that relevant to this scenario where cell capacity differences are slight

nor am I saying chemistry & resistance aren't more important factors.

But say a A. 2200mAh cell and B. a 3400mAh cell are being tested with a 3A load for 40min.

Likely less sag, certainly less DoD wear & tear, if the range is fixed not proportional

90% DoD vs 60% DoD

that factor favours greater longevity.

Even with lower difference, needs to be taken into account.

But if range kept proportional, take advantage of B's greater range and sacrifice potential longevity gain, actually draw both down to 90% every cycle.

[john61ct]

Technically it is also "sag" in that it is voltage drop under load, whcih is what sag is....

To me sag is temporary, the delta between under a certain C-rate load and recovered V once resting.

As opposed to both the constant-loaded curve, and the resting V vs SoC curve, neither of which show sag.

[john61ct]

Maybe a shroud with a small crankcase ventilation filter for the dust issue and just pop a baggy with a rubber band around it if you get caught in the rain ?

Such a complex failure prone "solution", should only be implemented if there actually is a problem.

The ideal is 100% fully sealed, if not potted, without internal cell temps rising too much.

[docware]

Yes, PF has worse sag thanks to its lower capacity.

Capacity does not determine voltage sag.

Internal resistance determines voltage sag.

Both capacity and resistance determine voltage sag. Compare GA (3,4 Ah 36 miliohm) with VTC5A ( 2,5 Ah 15 miliohm) and VTC6 (3 Ah 18 miliohm).

To Dak77 :
Best solution for cells "cooling" is to overdesign battery, keeping low current per cell.
2 A could be low enough for PF. You definitely want to have cells perfectly sealed against any moisture, dust and dirt.

M36 polystyren 4.jpg (64.33 KiB) Viewed 688 times

[Dak77]

Yeh I see your points . Having an opening of any kind is a bad idea , I admit.

[docware]

Yeh I see your points . Having an opening of any kind is a bad idea , I admit.

Exactly.

In the meantime I made small re-evaluation. In fact, we can expect that warming of the cells in real battery pack will be lesser than in my test. Moreover, both Pajda and flippy assess Panasonic PF and Samsung 29E as quite abuse resistant cells. So load between 2 and 5 amp per cell should not be problem. On the graph bellow we can see that 29E voltage sag is a little bit better then PF thanks to lower DCIR (PF 34 miliohm, 29E 31 miliohm). Flippy´s favorite 29E has probably one of the best price/value ratio here in Europe (nkon).
Conclusion : you can have good 13s8p battery with either PF or 29E.