LiFePO4 in 18650 format?

[flippy]

yes i do have the curve:
https://drive.google.com/open?id=1v8PGphx0BzGzPJsUpUzXVG9fTgijW61g

the sawtooth shape is due to the temperature changes in the room over time. once you average out those changes you are left with no measureable differences in usable capacity. even if you take in any percentages in measuring error you are still left with a cell that has basically the same usable capacity since day 1.

i have a much more "worn out" set of 29E's that have done about 1200 cycles in a actual vechicle with loads between 0.7A continous and up to 10A peaks but has been charged no more then 4.05V and discharge usually no more then 3.3V that has lost about 5% capacity wich is pretty great considering the abuse it got.

comparing those figures with the test bench numbers that are coming out of it keeps the notion standing that stopping the charging at 4.05V and charging at low currents is much more important then actual currents being drawn. if temperatures are kept within spec that is. i have no doubt that a pack with active heating to 30 degrees or so will preform even better. the old pack was used daily during winter and summer without any insulation and still preforms like new.

and the datasheet is running tests between 4.2 and 2.5V. my tests are between 4.05 and 3V so dont compare those.

 

[Honk]

Assuming I'm looking at the correct datasheet, the manufacturer claims 70% capacity after 500 cycles: https://eu.nkon.nl/sk/k/29E.pdf

Yes, that might be correct if discharged from fully charged 4.2V down to 2.5V.
But you have missed the point of Flippys 500 cycle test.
He discharged between 4.05V to 3V.
That is hell of a difference compared to a full 100% DoD.

 

[john61ct]

Sorry, confused.

These 29E appear to be of chemistry

NMC, LiNiMnCoO2

Is that correct?

 

[flippy]

Sorry, confused.
These 29E appear to be of chemistry
NMC, LiNiMnCoO2
Is that correct?

you are asking the wrong question: does it matter?

 

[docware]

the sawtooth shape is due to the temperature changes in the room over time. once you average out those changes you are left with no measureable differences in usable capacity. even if you take in any percentages in measuring error you are still left with a cell that has basically the same usable capacity since day 1.

i have a much more "worn out" set of 29E's that have done about 1200 cycles in a actual vechicle with loads between 0.7A continous and up to 10A peaks but has been charged no more then 4.05V and discharge usually no more then 3.3V that has lost about 5% capacity wich is pretty great considering the abuse it got.

comparing those figures with the test bench numbers that are coming out of it keeps the notion standing that stopping the charging at 4.05V and charging at low currents is much more important then actual currents being drawn. if temperatures are kept within spec that is. i have no doubt that a pack with active heating to 30 degrees or so will preform even better. the old pack was used daily during winter and summer without any insulation and still preforms like new.

Yes, perhaps everybody agree that keeping max voltage below 4,1 is positive for the cell lifetime.

You are twisting too much themes in your sentences. Leaving my questions about calibration without answer.

It is clear from your picture and you agree that temperature can have significant influence at capacity measurement.
Your 0,5 % capacity loss calculation is based on these measurements :
new: 9.534Wh
after 500 cycles:9.487Wh
What temperature was during measurement of the new cell and what after 500 cycles ?

When you say new, do you mean that it was new, never used cell ?

To Punx0r:
Hopefully Honk cleared up the dilemma to you enough.

 

[flippy]

current callibration was also done with the brymen as it is my most accurate meter. should not be much of a stretch to figure that one out.
and i never disagreed with you on temperature, and i also stated that it does not matter during the lifecycle testing in such a small temperature window. it will see much bigger temperature swings in actual use. and my test run with my oldes set slready showed that. you can also see that on average the capacity did not drop in any meaningful way during the 500 cycles.

first and final capacity test was done at 20C with the cell resting for about an hour or so after charging. it would not be a fair reading otherwise. the new cells did get 10 low current (500mA) cycles on them to break them in before starting the offical readings
and yes, cells are brand new, factory date puts them at 2,5~3 months old before testing started.

 

[docware]

Did you check the current during whole first and final measurement ? I mean checking the current during the whole discharge time to see the variations of the current.

 

[flippy]

there was no real difference between callibrations. i usually callibrate on the more extreme ends of the device rating, at 0.5v, 0.1a and 4.5v, 9A. biggest difference i have seen between 500 cycles was a few mA and single digit millivolts. hardly worth the effort.

also, upgrading the internal wires and solder joints help a lot. also cranking up the internal 5v power supply with a 0.5v higher output to account for voltage drop makes it a bit more stable.

you keep on trying to poke holes in the data, but please nock it off. i know how to take a proper measurement when it matters.

 

[docware]

Flippy, I consider cycle testing really important. So I am thankful to any information. Hope I´m not bothering you too much.

Probability that any cell have only 0,5 % capacity loss after 500 cycles between 4,05 and 3 V is very very low. Value 0,5 % seems to be rather accuracy of the equipment. Looks like measurement of one cell resting on the shelf in between.

By the way, the temperatures 15 – 20 °C are out of range of any spec. If you don´t carefully measure close to the cell during the whole procedure (rest before, charge, rest, discharge), than is temp measuring worthless. Can you please also show picture of the thermometer ?

Let´s make a summary. Cycle test consist of : charge 1,3 A to 4,05 V, discharge 2,75 A to 3 V. Cut off current ? Rest times between cycles ?
Was the first and final capacity measurement done at the same equipment ? What fixture was used for capacity measurement ? Any photo of the assembly ? Any curves of the first and final capacity measurement ?

Calibration is maybe fine. But if you don´t continuously measure and calibrate the current during the first and final measurement, you don´t know what is accuracy of the measurement. The real current can be quite different than you calibrated before the capacity measurement. And definitely will vary during the measurement.

10 low current (500mA) cycles before the cycling differing from normal procedure. Capacity loss is ussually biggest during the first 10 – 30 cycles. So you are in fact erasing this capacity loss at the beginning. Very different from normal testing.

 

[flippy]

i think you are not seeing the difference between capacity testing and lifecycle testing and treat them both as equals. they are not.

for you reference: during discharge as soon as the voltage hits 3v it cuts out. then it waits for 5 min to let the cell recover (it usually bounches back to 3.2v or so) and then starts charging again. top charge is 50mA cutoff, 5 min rest then discharge. pretty standard.

being off by a percent (or more) is NOT a problem in lifecycle testing. having sub 10mAh accuracy is compeltly usless. the goal of lifecycle testing is to wear out the cell, not have the cell in a perfect controlled enviroment. as shown it does not matter in the first and final proper test that is done according to industry standards. the performance might change a few percent but in the end it does not matter in the slightest.
having <1% accuracy during the test is simply not needed, or do you think having 2.9mAh would make much of a dent in the final conclusion? even if you take a much wider range and take 3% error, that is still 0.3Wh of difference at best. on a 10.44Ah cell that is not much to concern yourself about.
as with the temperature measuring it is bascially the same thing. temp changes (within reason) have no bearing on the actual cyles the cell needs to do. it might impact capacity during that cycle with a few percent but in the end it averages out over the course of the entire lifespan. and the actual capacity during the lifecycle testing is not meaningful. it does make a nice graph but it does not tell you much in the end.

you are focussing on lots of crap that does not matter during the test like the cell holders. those dont matter just like what equipment was used to do the different mesurements, if you get different results with the same test voltages and currents you need to callibrate the equipment. prehaps with double digit currents and very low voltages you do need to mind what you measure where, but with low currents and with proper holders this is just not a thing to worry about.

the low current cycles have no measureable effect on lifespan. and it is just to weed out any dodgy cells before slapping them on the tester for 3 months. i have done 50 cycle testing with fresh cells and dont see any difference between runs. "running in" cells is a old wives tale for normal cells.

 

[docware]

No, I was nearly all the time speaking about capacity measurement. Unfortunately I didn´t differ enough the questions to cycle test. (" Let´s make a summary. Cycle test consist of : charge 1,3 A to 4,05 V, discharge 2,75 A to 3 V. Cut off current ? Rest times between cycles ? ")

flippy, flippy, from all your feedback is obvious that you live in your bubble convinced that your testing is flawless. So I will not be loosing more time and let you live in that. But please take into account that presenting such results as 0,5 % capacity loss after 500 cycles is .... nonsense ? crap ? disservice to ES members ?

Howgh

 

[Punx0r]

Simple solution: establish test variability by taking repeated measurements of the same cell in the same condition(e.g. test every day at different times for a week), assuming normal distribution of error, calculate standard deviation and multiply by 2 (standard coverage factor).

That gives a +/- figure that can be applied to capacity loss results that will be correct in 95% of cases. I.e. if your result is 0.5% capacity loss +/- 0.05% then you know you're good. If it's 0.5% +/- 5%, then not so much.

It won't mean the measured capacity in absolute terms is accurate (you have to take into account the accuracy of all your measuring equipment for that), only the relative capacity, which is fine as you're measuring capacity loss. So it doesn't matter if your instruments are inaccurate, as long as they're roughly linear in their error in the measurement range spanning first to last cycle.

 

[flippy]

No, I was nearly all the time speaking about capacity measurement. Unfortunately I didn´t differ enough the questions to cycle test. (" Let´s make a summary. Cycle test consist of : charge 1,3 A to 4,05 V, discharge 2,75 A to 3 V. Cut off current ? Rest times between cycles ? ")

flippy, flippy, from all your feedback is obvious that you live in your bubble convinced that your testing is flawless. So I will not be loosing more time and let you live in that. But please take into account that presenting such results as 0,5 % capacity loss after 500 cycles is .... nonsense ? crap ? disservice to ES members ?

Howgh

i am not living in a bubble, stop talking people down.
my testing is not " perfect" but it counts where it matters. that that is something you simply cant understand. there is no need to do lifecycle testing in perfect 20c enviroments with 0.1% resolution. the cell needs to wear out in the parameters that has been set. that is it, nothing more or less. you dont need constant daily callibrations or controlled enviroments for that.
if you think my testing is wrong then prove me wrong with hard data invalidating my testing parameters. otherwise bow out of this conversation.

 

[Honk]

I support Flippy 100% in this.
There is no need for ultra high precision in these tests.
What matters is consistency to see the degradation over time.

 

[john61ct]

Sorry, confused.
These 29E appear to be of chemistry
NMC, LiNiMnCoO2
Is that correct?

you are asking the wrong question: does it matter?

Of course cell chemistry matters.

Wildly different charge profiles, longevity, thermal behaviour.

For members interested in LFP cells, this thread now has dozens of posts full of irrelevant details about a battery type that has nothing to do with the OP.

 

[agniusm]

The key here i think is safety. Its for underground so to minimize fire risk. Lfp is second to titanite in safety.

 

[flippy]

The key here i think is safety. Its for underground so to minimize fire risk. Lfp is second to titanite in safety.

granted, but you also need much more of it due to the lower energy density. so 30% more of slightly less flammable stuff is probably equally as dangerous as having less of slighty more dangerous stuff.

the real question is what kind of danger are you specifically trying to remove in said use case?

if you have proper venting space for regular 18650 lipo containing 10Wh or so then a cell fail will be confined to that specific cell and that will be the end of it with a puff of offgassing.
having a single 100Ah lifepo4 cell going thermal means having a cell containing 320Wh go nuclear. that is WAY more difficult to contain and control then a puny 18650 and has much more risk of igniting the ajoining cells as well.

 

[john61ct]

it's not a slight difference

you'd need to **create** on purpose the extreme conditions required for LFP to "go thermal"

just doesn't happen in normal usage, at least as safe as lead

To the point that they can be charged and used unattended, including inside a small living space.

Of course new cells not scavenged, properly engineered, protective electronics, etc

 

[flippy]

the same goes for lipo, especially for 18650 cells. and if a 18650 goes (even if it does) its very contained and the cell is quite small so any thermal issues are nothing as spectaculair as people make out. i tried in the past years to make panasonic PF's to do anything more then make a "puff" and it simply didnt happen. only once i overcharged to 4.4v for an hour and directly dead shorted the cell it -eventually- caused the offgassing to catch fire for a minute or so.
once you go with pouches then its a whole different story those are bloody dangerous and should never be used in large stationary setups.
i never have seen a quality made pack go thermal because a single cell popped it's top. its only cheap ass chinese cells or pouches that go nuclear.
especially cells like the 29E and PF or GA's simply pop their PTC and that is the end of that cell without any drama. especially the new 2018 models are almost impossible to make them do anything more then vent. the PTC goes long before anything else dramatic happens.

 

[john61ct]

That's fine, of course some people worry "too much" about safety, and there are dozens of variables to factor in.

But within that context, for those who don't need to care much about energy density and prefer to prioritize safety, it is not accurate to say LFP is only a little safer than the other LI chemistries usually used for propulsion use cases.

Put a family in a small boat out at sea or living in a camper with an LI bank, that would be IMO the only way to go.

 

[flippy]

in boats i generally use LFP simply because you have massive cells and only need a few of them and a hand of bolted interconnects and a very simple balance board to make a solid working pack without a need for maintenance. the weight and size is usually also not a problem in a boat.
making a 18650 battery is something i do use for camper conversions for weight and power density. by generally i use 14S packs with 48V Ac inverters and a 12-60v charger to charge the battery. a quarter of the amps and moch less losses make for a safer pack. battery is usually ouside the passenger compartiment so no issue there. i also build a heavy duty alu case with a vent so nothing happens if something happens.
last one i did was a conversion where i removed the generator and filled almost the whole space with 18650 cells for a 35kWh battery. never a need for a generator ever again.

 

[agniusm]

the same goes for lipo, especially for 18650 cells. and if a 18650 goes (even if it does) its very contained and the cell is quite small so any thermal issues are nothing as spectaculair as people make out. i tried in the past years to make panasonic PF's to do anything more then make a "puff" and it simply didnt happen. only once i overcharged to 4.4v for an hour and directly dead shorted the cell it -eventually- caused the offgassing to catch fire for a minute or so.
once you go with pouches then its a whole different story those are bloody dangerous and should never be used in large stationary setups.
i never have seen a quality made pack go thermal because a single cell popped it's top. its only cheap ass chinese cells or pouches that go nuclear.
especially cells like the 29E and PF or GA's simply pop their PTC and that is the end of that cell without any drama. especially the new 2018 models are almost impossible to make them do anything more then vent. the PTC goes long before anything else dramatic happens.

I dont even need to put a theory here. How many lifepo4 fires have you seen on as? Probably none. And how many Lipo fires? A lot. 18650 probably less but there was molten bikes and burned sheds and rooms.

 

[flippy]

Direct comparison is stupid. There are a million times more lipos in cirxulation for a lot higher demanding applications then lifepo4 can dream of.

 

[e-beach]

Direct comparison is stupid. There are a million times more lipos in cirxulation for a lot higher demanding applications then lifepo4 can dream of.

Unless one takes it to a percentage like, what percentage of lipo's sold catch on fire vs. what percentage of LiFePO4's sold catch on fire.

 

[john61ct]

No the greater relative safety of LFP over other LI chemistries is major, in fact extreme, and that is objective fact, incontrovertible and not subject to debate any more than "the earth is round" or "vaccines cause autism".