Samsung INR18650-25R cycle test.

[DrkAngel]

Firstly, DOD has been thoroughly documented as a factor in cycle life. Yet findings cannot be summarized as easily as 100%, 200% 400% etc. A particular cell I know has no effect of lowering DOD below 4,1V, but experiences 400-600% increase in cyclelife compared to 4,2V. It seems to me you started studying batteries and invented a semi-scientific method to compare results.

"Firstly"? ... well firstly ... you have no understanding as to what DOD means!
DOD = Depth of Discharge = how deeply a cell is discharged example 2.5V vs 3.1V vs 3.2V.
(explained multiple times, so please learn how to read.)

Secondly, you need to stop with your capacity map jibberish. "A very slow discharge should produce a near static voltage capacity map allowing precise determination of optimal charge voltage" - I mean, WTF. When did you wake up and determine to discount the effect of internal resistance in a battery?

Secondly a.
Just because you are incapable of understanding something does not make it gibberish.
It might be gibberish to you but others with more discerning minds find it enlightening and very useful.

Secondly b.
Internal resistance is directly considered.
The purpose of a very slow (low rate) discharge is that it nearly eliminates internal resistance in a battery as a factor for mapping capacity!

 

[Teh Stork]

Sorry, I mixed DeltaSOC and DOD in my previous post.

On the topic of DOD and cycle life, it has been studied in detail with neural-network algorithms. Basically there is a difference. For this particular dataset - influence from temperature is 3x more critical than DOD. As with all li-ion technology, there are no rules to be made for "all li-ion tech".

As IR often increases on low SOC, the battery often heats up considerably if full DOD is used. Increased temperature is undeniably linked to capacity degradation.

A normal 0,2C discharge curve, with voltage as a function of mAh delivered - can easily be used to determine voltage at different DOD.

Please enlighten me, how is voltage capacity map very useful? Google doesn't find any user, but you, which claims to determine perfect charge voltage and DOD using this method.

 

[zener]

Maybe it would help to stop discharge at different stages and than after a little resting time measure the voltage under no load?

 

[madin88]

I can make a good guess on the calendar life of the cells at room temperature. This is because I've read books on li-ion batteries, electrochemistry and published papers - instead of acting like a battery expert on an internet forum.

Open your eyes.

if you have something useful to say, than say it instead of speaking bad about others here.
yes the calender life also should be considered, but it should be quite high with NCA chemistry.
i find the voltage capacity map very useful because it makes it more visible as a normal discharge curve does, what he already has explained

@ DrkAngel

thank you for putting the excel list into such nice graphs. its very interesting to see that.
do you think other cells have similar dischage / charge capacity map, or will it look rather different?

when considering all the test published here by now, i suggest to charge to a maximum voltage somewhere between 4,1 and 4,15V. maximum discharge depends on the current so i think i will set LVC to 3V per cell for having more amps at lower SOC levels and prevent early controller cut off, but we than should stop applying throttle if current drops below 2A per cell and voltage below 3,1-3,2V.
the "optimal" no-load rest voltage when discharged is 3,4V right? below 3,4V there is alomost no capacity left (big voltage sag even at low current) and the cell can be called empty, right?

looking forward to the cylce tests to see the effect of different charge and discharge rates.
what about comparing 5A discharge with 10A?
FWIK most who have a batterie built with 25R will use them at 10A or higher (for instance 10p and 100A or 12p and 120A is very common..) so i think the cylce life at this current would be nice to know.

good point about charging to higher volts and cut off early, or with other words shortening the CV stage. it would be great if this does not worsen cylce life compared to normal charge. should be added to the planned tests^^

 

[DrkAngel]

Every tested cell has definite bulges of capacity.
But differing formulations have various locations and differing optimal static DODs.
Some at 3.6V, others at 3.4V or 3.2V.

1C 2.5A discharge = in progress (500 cycles scheduled)
2C 5.0A discharge = scheduled
4C 10.0A discharge = Might make a good future test (Not sure of discharge device capability?)
8C 20.0A discharge = mapped to 250cycles

 

[DrkAngel]

Please enlighten me, how is voltage capacity map very useful? Google doesn't find any user, but you, which claims to determine perfect charge voltage and DOD using this method.

Wow!
Thank you.
I would not have presumed such a claim ...
But if you insist that I am the #1 innovative pioneer of battery capacity mapping ...
I would be proud to accept the mantle!

 

[electricbike]

440 cycles uppdated. I will finish this test first before i start with the rest
of the cycle tests.

Those charger i am using can do 10A discharge,no problem.

 

[electricbike]

2 more cycletest started today.

4.15V >>> 3.10V @ 2.5A Charge 1C \ 5.0A Discharge 2C (to map discharge rate deterioration - direct comparison to 4.15V >>> 3.10V @ 2.5A C\D)
A direct comparison of a 1C discharge vs a 2C discharge. Cycletest started today.

4.15V >>> 3.10V @ 4.0A Charge 1.6C \ 5.0A Discharge 2C (to directly compare charge rate deterioration - @ maximum recommended charge rate)
A direct comparison of 1C charge rate vs a 1.6C charge rate
Alternate? - 4.15V >>> 3.10V @ 5.0A Charge 2C \ 5.0A Discharge 2C.Cycletest started today.

 

[madin88]

if the charger cut off at 3,1V with 5A current, what voltage does the cell than have unloaded?
because i think if you increase the discharge current, you could lower cut off voltage..

such cycle test would be great:

4,15V>>>3.00V or 2,9V
2,5A charge, 10A discharge

and a comparison between charging with normal CV stage, and charging to a higher voltage but with cut off when charge current drops below 0,5C or something like that (no load voltage sould be similar).

 

[Hillhater]

if the charger cut off at 3,1V with 5A current, what voltage does the cell than have unloaded?
because i think if you increase the discharge current, you could lower cut off voltage..

such cycle test would be great ).

This is not devaluing the results you are getting here, just adding to the "thought bank".....
There was some info from the Tesla cell tests that suggested pulsed discharges gave significant improvement in cell life, and were more representative of normal use.
I guess ideally a discharge test would be "pulsed" ,..eg, discharge at 10 amps for say 20secs, then let the cell recover for 10 secs, ...recording loaded and resting voltages ,....then repeat until the loaded voltage hits whatever low limit you decide.
(Pick your own time periods :wink: )
Similar to this ..https://endless-sphere.com/forums/viewtopic.php?f=14&t=73701
And of course you get the DCIR data also.
Just armchair comments really !...

 

[Offroader]

if the charger cut off at 3,1V with 5A current, what voltage does the cell than have unloaded?
because i think if you increase the discharge current, you could lower cut off voltage..

such cycle test would be great ).

This is not devaluing the results you are getting here, just adding to the "thought bank".....
There was some info from the Tesla cell tests that suggested pulsed discharges gave significant improvement in cell life, and were more representative of normal use.
I guess ideally a discharge test would be "pulsed" ,..eg, discharge at 10 amps for say 20secs, then let the cell recover for 10 secs, ...recording loaded and resting voltages ,....then repeat until the loaded voltage hits whatever low limit you decide.
(Pick your own time periods :wink: )
Similar to this ..https://endless-sphere.com/forums/viewtopic.php?f=14&t=73701
And of course you get the DCIR data also.
Just armchair comments really !...

Could you say that an average of what you would use pulsed would give somewhat good results? Like if you use use between 0 and 10 amps per cell, wouldn't 5 amps be a good number to use that is an average of what you would use.

 

[DrkAngel]

Most all controllers use a pulse type, PWM (Pulse Width Modulation), output, but many times per second.
A prolonged full throttle for several seconds then off for several seconds is much worse on battery-motor-bike-wrist-nerves-traveling companions etc.

Due to full throttle inefficiencies, a 100% throttle 50% of the time would be worse in every way than 50% throttle 100% of the time.

 

[electricbike]

460 cycles uppdated.

 

[Hillhater]

....Due to full throttle inefficiencies, a 100% throttle 50% of the time would be worse in every way than 50% throttle 100% of the time.

Agree, but how many of us ride with a constant steady throttle ?
But my point was that there was some tests that showed that a battery with a pulsed discharge held a higher voltage profile than the constant discharge.....it tended to "recover" voltage more than expected between pulses.
..I will try to find that test thread...

 

[Teh Stork]

....Due to full throttle inefficiencies, a 100% throttle 50% of the time would be worse in every way than 50% throttle 100% of the time.

Agree, but how many of us ride with a constant steady throttle ?
But my point was that there was some tests that showed that a battery with a pulsed discharge held a higher voltage profile than the constant discharge.....it tended to "recover" voltage more than expected between pulses.
..I will try to find that test thread...

Here is a test from Okashira that shows this phenomenon.Thread here.

To simplify, the battery electrodes both act as sponges - just with ions instead of water. If you rest the cell, the electrodes has a chance to "equalize internally" and increase the number of ions avatible on the surface of the electrodes. I think Okashira called this surface charge, but that term is often used while discussing other aspects of li-ion batteries.

On the topic of throttle and efficiency. If you go WOT on a trap-controller (cheap, often found in ebikes), there will often be efficiency gains if the speed of the motor is high enough. This is basically anytime rated battery current is lower than rated controller battery current, while WOT is held. My old commuter bike would climb hills with 65 degC controller at WOT, but at less throttle (and less speed) - the controller would spike to 120-130 degC. Ofc it should not be this way, but the way those cheap ass controllers are built - that is the result.

 

[DrkAngel]

On the topic of throttle and efficiency. If you go WOT on a trap-controller (cheap, often found in ebikes), there will often be efficiency gains if the speed of the motor is high enough. This is basically anytime rated battery current is lower than rated controller battery current, while WOT is held. My old commuter bike would climb hills with 65 degC controller at WOT, but at less throttle (and less speed) - the controller would spike to 120-130 degC. Ofc it should not be this way, but the way those cheap ass controllers are built - that is the result.

That is normal and reasonable.
See - Attack them Hills
Gearing, for necessary terrain-gradient, is essential for keeping efficiency up and heat down!

 

[markz]

Firstly, DOD has been thoroughly documented as a factor in cycle life. Yet findings cannot be summarized as easily as 100%, 200% 400% etc. A particular cell I know has no effect of lowering DOD below 4,1V, but experiences 400-600% increase in cyclelife compared to 4,2V. It seems to me you started studying batteries and invented a semi-scientific method to compare results.

"Firstly"? ... well firstly ... you have no understanding as to what DOD means!
DOD = Depth of Discharge = how deeply a cell is discharged example 2.5V vs 3.1V vs 3.2V.
(repeated multiple times, so please learn how to read.)

Secondly, you need to stop with your capacity map jibberish. "A very slow discharge should produce a near static voltage capacity map allowing precise determination of optimal charge voltage" - I mean, WTF. When did you wake up and determine to discount the effect of internal resistance in a battery?

Secondly a.
Just because you are incapable of understanding something does not make it gibberish.
It might be gibberish to you but others with more discerning minds find it enlightening and very useful.

Secondly b.
Internal resistance is directly considered.
The purpose of a very slow (low rate) discharge is that it nearly eliminates internal resistance in a battery as a factor for mapping capacity!

I have been trying to understand this for the past hour, as I want to test the IR of my 80 individual a123 26650 M1A's

Found a youtube video, which I cant find now, gave me this document;
http://data.energizer.com/PDFs/BatteryIR.pdf

Same video showing the long form to get to this,
IR = [V(high) - V(Low)] divided by [I(low) - I(high)]

Two seperate calculations with two seperate resistors to get a High Current Draw and a Low Current Draw with different resistors for corresponding draws.

OK I get that, do I seriously have to test each one of 80 batteries individually?
Is there a way to do lets say 8 in series or more?

 

[DrkAngel]

OK I get that, do I seriously have to test each one of 80 batteries individually?
Is there a way to do lets say 8 in series or more?

I run a comparative IR test.
Since tested cells will be used in same pack, they will have same draw. Any discharge device should indicate a reasonably accurate IR comparison.

But 1st!
Charge all cells to identical voltage (4.20V?)
Connect cells in parallel and bulk charge for precise equality (I leave laptop cells as 2p ... typically.)

2nd
Let set for several days then measure for self discharge >>> eliminate self discharging cells (more than a couple hundredths)

3rd
Now ... test IR (Internal resistance)
I used to test 3s2p laptop packs with a 3A+ HD server fan
With fan running, I metered each cell pair and labeled with voltage drop
EG 4.20V pair under 3A discharge sagged to 4.02V, so I labeled with 18 as a comparative IR
(replaced with 3D printer 12V 3.44A heating element in bottle of water - fan moved and had extremely sharp blades!)
(Excessively bad IR cell pairs were separated and individually tested)
As long as same discharge device is used on all cells comparative IR should be accurate
6V headlight(?), heating element(?), for individual cells

finally
With cells of similar IR
I rigged as high as 30s 2p (126V) and discharged with 2 x 60w 120V light bulbs (= 1A discharge rate = 1AH per hour discharged)
Precise meter of each cell or cell pair under discharge confirms of similar IR

• Good test for "new" cells confirms all cells of
• no self discharge
• of comparable IR
• of similar capacity
  Better test process than most manufactures or sellers provide!

This is my capacity test method for used cells - 3000mAh + residual V (3k - 3.76V)

Full capacity test can be run with cheap ($12) 90V self powered V-A-Ah meter
Rig as ~20P (84V) and discharge @ .5, 1A, + rate

remove cells as they hit your determined discharged voltage
mark each with metered mAh as removed
reconnect series and continue
I connect cells with small neodymium magnets

So, yes, accurate IR can be batched determined.

 

[electricbike]

480 cycles uppdated.

 

[TheBeastie]

Gearbest is having a sale on 25Rs at the moment $15 for 4 with free shipping. A lot of folks like to compare without shipping, also gearbest is paypal friendly so your free to test the crap out of your cells and try and prove their fake and get a full refund.
http://www.gearbest.com/batteries/pp_241350.html?wid=1

 

[Hillhater]

But my point was that there was some tests that showed that a battery with a pulsed discharge held a higher voltage profile than the constant discharge.....it tended to "recover" voltage more than expected between pulses.
..I will try to find that test thread...

Here is a test from Okashira that shows this phenomenon.Thread here..

.. Thanks TS, that was the thread I was trying to find. :wink:

 

[electricbike]

My first post uppdated with the graf of the new cycletest (different charge current) 40 cycles.

 

[DrkAngel]

My first post uppdated with the graf of the new cycletest (different charge current) 40 cycles.

Do you have "20" or starting "0" values for "difference"

 

[electricbike]

My first post uppdated with the graf of the new cycletest (different charge current) 40 cycles.

Do you have "20" or starting "0" values for "difference"

I actually forgott to take any measurment from the start so i had to put the start from 40 cycles.
I also put the charger to do 40 cycles insteed of 20, then i don't have to change the graf so often

 

[DrkAngel]

Starting points should be nearly identical.
After a few more cycle runs I should be able to extrapolate reasonable capacity paths.

A single quick test of charge 4.15 @ 2.5A / discharge 3.10V @ 5A (with cell of similar start condition) should give an accurate start point for both ... after #1 cycle test complete?
Start point is necessary to map capacity loss.
Capacity loss is most extreme in the beginning cycles.