Samsung 40T and other cells like it?

[Pajda]

I am not saying that there is no dependency, I am trying to say that there is no general rule to evaluate this dependency (or at least I did not find it). If I use example you can find cell which have after 10% capacity fade ca 30% rise of DCIR (good one) and you can have another cell (with similar datasheet specs) which have after the same measured 10% capacity fade ca 80% rise of DCIR (bad one).

Unfortunatelly this is problem of almost all parameters. You simply cannot generalize the measurement results from one cell to another . In the next thread, docware posted thesis where the author makes excellent scientific work about li-ion cell aging with the cell with one of the shittiest operational life on the market.

 

[rg12]

Thats very interesting!
Is that from the datasheet?
I would like to know the temps for about double these currents for about 1 minute constant

It´s my measurement.
1 minute is too short time to see the temperature change on the surface of the cell. Can you describe intention of such measurement more extensively ?

My most common use is high powered ebikes where nothing is constant current and it all works with bursts and even full throttle is not max current since it settles at max speed so about 1 minute at max current is the max my pack will see in addition to some cruising with full throttle bursts.

Where can I get that cell holder and what equipment do you use for testing to create the charts?

 

[docware]

Current is variable at all ebikes. Solution is estimation of the average current and perform the warming test at constant average current.

You can get the cell holder BF-2A from Aliexpress. My version is 10 A max.
Equipment : GW Instek GDM-8351, Maynuo M9712, ZKETECH EBC-X for charging, Omega HH520, K-thermocouple Omega wires diameter 0,127 mm.

Here is another warming test of Samsung 40T cell in subtle plastic holder and packed in the expanded polystyrene. Purport of the test was effort to simulate warming in the battery pack.

 

[rg12]

Current is variable at all ebikes. Solution is estimation of the average current and perform the warming test at constant average current.

You can get the cell holder BF-2A from Aliexpress. My version is 10 A max.
Equipment : GW Instek GDM-8351, Maynuo M9712, ZKETECH EBC-X for charging, Omega HH520, K-thermocouple Omega wires diameter 0,127 mm.

Here is another warming test of Samsung 40T cell in subtle plastic holder and packed in the expanded polystyrene. Purport of the test was effort to simulate warming in the battery pack.

1.jpg

7.jpg

9.jpg

Samsung 40T 10A disch warming b.jpg

I like the average over time idea

Thank you for listing all of your awesome testing gear.
I got all I need except for the cell holder which I will order and the thermal logger.
Do you know of any cheaper ones? yours is a 4 channel and very expensive...

Also, do you know how people test longevity with number of cycles?
Is it a device that charges and discharges the battery 24/7?
If so, will it really produce realistic results? because in real life the battery is not being charged and discharged non stop nor from full to empty every cycle...

EDIT: Found this device, do you think it can log to a computer and create a chart like yours?
https://www.aliexpress.com/item/327...chweb0_0,searchweb201602_6,searchweb201603_55

 

[docware]

ZKETECH EBC-X0510 is 8 – channel equipment for long term cycling obtainable at Aliexpress. Allows full control of testing parameters : charge and discharge currents, charge cut-off, charge and discharge voltage, waiting time between cycles. Accuracy better than 0,3 %.

 

[rg12]

ZKETECH EBC-X0510 is 8 – channel equipment for long term cycling obtainable at Aliexpress. Allows full control of testing parameters : charge and discharge currents, charge cut-off, charge and discharge voltage, waiting time between cycles. Accuracy better than 0,3 %.

Wow that is amazing!

Can you please check out the link I attached in the previous message and tell me if it's a device that enables to create charts?
btw, what is the difference between the temp probe that comes in the link I posted to the super thin one you use?

 

[flippy]

ZKETECH EBC-X0510 is 8 – channel equipment for long term cycling obtainable at Aliexpress. Allows full control of testing parameters : charge and discharge currents, charge cut-off, charge and discharge voltage, waiting time between cycles. Accuracy better than 0,3 %.

i use the same one. but callibrating is a b itch. i also got a single channel one that can do 50A for testing big cells or large parrallel cell blocks.

 

[docware]

Can you please check out the link I attached in the previous message and tell me if it's a device that enables to create charts?
btw, what is the difference between the temp probe that comes in the link I posted to the super thin one you use?

Thermometer MASTECH MS6514 has logging possibility and USB connector, so data acquisition and transfer to Excel should not be problem. It has standard thermocouple mini connectors, you can use any other thermocouple with this connector.
The main difference will be time response, my is much faster.

I have chosen Omega thermometer because I knew that company and it´s wide spectrum of products from my former job.
The thermocouple I made from fine thermocouple wires I purchased together with HH 520 from local distributor. Wires are only twisted together and fasten to the connector. Thanks to small diameter is thermocouple very fast to catch temperature changes on the Li-ion cell surface.

https://www.omegaeng.cz/pptst_eng/TFIR_CH_CI_CC_CY_AL.html

https://www.newark.com/omega/smpw-k-m/thermocouple-connector-k-type/dp/30AC8456

 

[docware]

i use the same one. but callibrating is a b itch. i also got a single channel one that can do 50A for testing big cells or large parrallel cell blocks.

I don´t like ZKETECH calibration, in fact I hate ZKETECH calibration
I killed my EBC-A10H tester by incessant calibration, I killed also one channel at EBC-X0510.
Support from ZKETECH close to zero.

That was reason I purchased Maynuo M9712. This instrument has accuracy better than 0,05 %, usually better than 0,02 %, pure joy for my accuracy obsession . After some current checking with precise shunt resistor 0,01 Ohm (tolerance 0,1 %, RIEDON RSN – 10) and DMM GW Instek GDM-8351 I am sure that Maynuo is that accurate and the accuracy is stable for the whole measurement.
Only disadvantage is that Maynuo is electronic load only and charging has to be done at different equipment (EBC-X0510 in my case). But has more programming options.

Also checked three channels of EBC-X0510, current accuracy is better than 0,3 %, often better than 0,2 %.

 

[rg12]

Can you please check out the link I attached in the previous message and tell me if it's a device that enables to create charts?
btw, what is the difference between the temp probe that comes in the link I posted to the super thin one you use?

Thermometer MASTECH MS6514 has logging possibility and USB connector, so data acquisition and transfer to Excel should not be problem. It has standard thermocouple mini connectors, you can use any other thermocouple with this connector.
The main difference will be time response, my is much faster.

I have chosen Omega thermometer because I knew that company and it´s wide spectrum of products from my former job.
The thermocouple I made from fine thermocouple wires I purchased together with HH 520 from local distributor. Wires are only twisted together and fasten to the connector. Thanks to small diameter is thermocouple very fast to catch temperature changes on the Li-ion cell surface.

https://www.omegaeng.cz/pptst_eng/TFIR_CH_CI_CC_CY_AL.html

https://www.newark.com/omega/smpw-k-m/thermocouple-connector-k-type/dp/30AC8456

I want to log it with a line chart, not excel data, how did you do that? is that a software that comes with your termal device?

Why would I need it to be so fast?
I don't think a cell would ever flactuate in temp very fast...

 

[flippy]

i use the same one. but callibrating is a b itch. i also got a single channel one that can do 50A for testing big cells or large parrallel cell blocks.

I don´t like ZKETECH calibration, in fact I hate ZKETECH calibration
I killed my EBC-A10H tester by incessant calibration, I killed also one channel at EBC-X0510.
Support from ZKETECH close to zero.

That was reason I purchased Maynuo M9712. This instrument has accuracy better than 0,05 %, usually better than 0,02 %, pure joy for my accuracy obsession . After some current checking with precise shunt resistor 0,01 Ohm (tolerance 0,1 %, RIEDON RSN – 10) and DMM GW Instek GDM-8351 I am sure that Maynuo is that accurate and the accuracy is stable for the whole measurement.
Only disadvantage is that Maynuo is electronic load only and charging has to be done at different equipment (EBC-X0510 in my case). But has more programming options.

Also checked three channels of EBC-X0510, current accuracy is better than 0,3 %, often better than 0,2 %.

i reconize your frustrations.

how did you kill the channel?

 

[john61ct]

Why would I need it to be so fast?
I don't think a cell would ever flactuate in temp very fast...

Not fluctuating up and down, key is catching **rising** temp, and exactly most needed when it does climb steeply.

The point is it starts running away internally long before any change makes it to the outside.

So you want sensitivity to be able to stop input as quickly as possible once a fast rise is detected.

Boom bad

 

[Hillhater]

I want to log it with a line chart, not excel data, how did you do that? .....

Excel charts, ?

 

[rg12]

I want to log it with a line chart, not excel data, how did you do that? .....

Excel charts, ?

Hmm right

 

[docware]

how did you kill the channel?

I killed EBC-A10H tester by incessant calibrating only high current during capacity measurements. Proper procedure is calibrate both, high and low current. Very probably that is also reason for issue with one channel of EBC-X0510. So it was my mistake.

 

[docware]

Why would I need it to be so fast?
I don't think a cell would ever flactuate in temp very fast...

I have chosen such subtle, thin wires because I wanted to have very fast thermocouple. For cells warming measurement is my thermocouple probably unnecessarily fast, but on the other hand I am sure that I am able to catch any temperature changes on the surface.

By the way "I don´t think" doesn´t mean I know the reality. Only way to know it is measurement done properly with proper instruments. Then you start to uncover that reality is more complicated than you ever anticipated. 8)

 

[flippy]

why do you stop recording after stopping the discharge? in my experience the temperaure continues to rise in high amp testing as the heat takes time to travel to the outside of the cell.

 

[docware]

why do you stop recording after stopping the discharge? in my experience the temperaure continues to rise in high amp testing as the heat takes time to travel to the outside of the cell.

Yes, you are right, there is always some thermal momentum adequate to amperage. I don´t stop temp recording with discharge, I continue for some time. In all graphs the highest temp point is displayed. Typical course of temperature :

 

[Hillhater]

The key here is not just heat generation, but also dissipation rates .
Have you considered running a variable/intermittent load profile test to simulate the typical use profile of a battery load. ?
IE, max discharge for a few seconds, (accelleration) ..reduced low load for a few more secs ( cruise),.. zero load for a period ( decell’/stopped), ..rinse & repeat in various combinations etc .
I realise there are many possible combinations, but maybe you can pick a “standard” load cycle to run comparisons .

 

[Dak77]

The key here is not just heat generation, but also dissipation rates .
Have you considered running a variable/intermittent load profile test to simulate the typical use profile of a battery load. ?
IE, max discharge for a few seconds, (accelleration) ..reduced low load for a few more secs ( cruise),.. zero load for a period ( decell’/stopped), ..rinse & repeat in various combinations etc .
I realise there are many possible combinations, but maybe you can pick a “standard” load cycle to run comparisons .

For suburban commuting , that sounds like a very viable set of conditions. Hard acceleration for 15 to 20 secs, hold at lower to mid amp draw for 30 secs, then 0 amp draw for 5 secs and repeat .

 

[rg12]

why do you stop recording after stopping the discharge? in my experience the temperaure continues to rise in high amp testing as the heat takes time to travel to the outside of the cell.

Yes, you are right, there is always some thermal momentum adequate to amperage. I don´t stop temp recording with discharge, I continue for some time. In all graphs the highest temp point is displayed. Typical course of temperature :

Samsung 40T 10A discharge warming test.jpg

I have a question since you seem to have done alot of heat tests of cells...
I'm using a brand new well built pack of 25R cells 20S 7P and pull way over the ratings I should (about 120A I'm guessing).
I've calculated my average amp draw with the time it took me to deplete 50% of my battery which was about 15 minute ride (about 10km riding pretty hard).
So in 30 minutes I will use all 17.5Ah capacity meaning that I was pulling 35A average current.
Now, 35A divided by 7P means exactly 5A for each cell in the pack which means that I was working an 8C rated cell at 2C for constant 30 minutes.
Now#2, pulling about 120A from a 17.5Ah 25R pack is abusing it as I've seen 25R packs being totaled after 10 month of use (not even daily) at 110A 20Ah which is bursts of 5.5C.
So I'm abusing my poor 7P 25R pack.
BUT, my average current draw calculation means 2C constant from an 8C cell so how can that be abusing?
Whats even more concerning is that the pack I talked about that was totaled after 10 month of use at 5.5C wasn't going beyond slightly warm when in that abusive use which seems that even when it's not getting hot I'm still killing it.
How can it be?

 

[Hillhater]

Its because you do not know what actual current you draw ar peak , and for how long...you are “guessing”
And averages mean nothing when you are trying to find out what causes damage.
Think.. you could swing a hammer at your head and “on average “ miss ...but one of those blows may hit and kill you !
Your 120A “Guess” would mean bursts of 17+ amps (or more ?) per cell.
How long and how often you do that is more relevant than your 5A / 2C average estimate.

 

[rg12]

Its because you do not know what actual current you draw ar peak , and for how long...you are “guessing”
And averages mean nothing when you are trying to find out what causes damage.
Think.. you could swing a hammer at your head and “on average “ miss ...but one of those blows may hit and kill you !
Your 120A “Guess” would mean bursts of 17+ amps (or more ?) per cell.
How long and how often you do that is more relevant than your 5A / 2C average estimate.

It's pretty hard to calculate as it's so random, for me it's about 5 second long bursts that come let's say every 20-30 seconds but still super hard to estimate...

 

[john61ct]

Use an ammeter / coulomb counter to give you actual data

 

[Hillhater]

I am trying to imagine in rg12’s “real world” pack of many 18650s, how some of those “internal” cells with no exposure to external cooling, but infact surrounded by other “Hot” cells ..what their temperature build profile is like when the pack is subjected to bursts of 120+ amps (8C) ?
Do those internal cells have any chance of shedding heat between bursts ? Or does each burst just jack the temp up further ?
A temp probe inside the pack on those inner cells would be a useful data source.