EVE 304 Cells from 18650batterystore & zke-40 testing.

[Sam Cho TX]

Yes absolutely LOL ?I at least know this much

 

[Sam Cho TX]

I just finished testing, and well I'm pretty ticked off tbh.

victron shunt said 265 and sfk app said 267 ah

the sfk app recorder showed me some information:

Sun Fun Kits DIY LifePo4 Lithium Battery Kits & Ready Made AGM / Lead Acid Replacement Batteries for RV, Golf Carts, Solar Storage, Back Up Power, Hurricane Outage . Made in Baton Rouge, LA, USA Ready to Ship out no overseas delay! Sunfunkits.com

Sun Fun Kits DIY LifePo4 Lithium Battery Kits & Ready Made AGM / Lead Acid Replacement Batteries for RV, Golf Carts, Solar Storage, Back Up Power, Hurricane Outage . Made in Baton Rouge, LA, USA Ready to Ship out no overseas delay! Sunfunkits.com

www.sunfunkits.com

Looks like cell 2 is just loosing power during the test the other cells are fine. If you go down to the cell voltage area you can see. What the hell this was the cell that tested to 306!

Here is how they are on the BMS:

Cell 1 : Label 305.6 AH - testing to 292 on the zke
Cell 2 : Label 303 AH - testing to 306 on the zke
Cell 3 : Label 305 Ah - testing to 317 AH on the zke
Cell 4 : Label 302 ah - testing to 297 on he zke

I don't understand thought for sure I would get at least 288 ish but cell 2 the one that tested to 306 only provided 265 Ah under like a 750-850 watt load.

This sucks!

 

[Sam Cho TX]

Notice the end of discharge voltage decay drop off on cell 303.

It is quite different of the other cells and is indicative of a cell that has not had enough initial cycles where the cathode and anode capacity curves have not reached their final alignment. It may gain in AH's after a few cycles. It would surprise me they did not catch this and do more cycles to correct.

It is also possible the neg anode graphite print thickness was on the thin side so anode does not have the extra 15-20% graphite. This would be bad as not only will it produce less AH's of capacity but it will also age quicker than other cells. This would be a manufacturing process control screw up. (something that never happens in China )

These graphs are plotted with X axis in incremental AH's consumed, not the time increments that the zketech plots out.

View attachment 142705
View attachment 142713

OMG how did you know the 303 was going to be bad! It dragged the whole test down I was only able to get 267 Ah from the pack! Look at the report: https://www.sunfunkits.com/app/ViewBenchmarkDetails?id=a8d694597315

 

[sunshine_eggo]

Did the BMS report a similar capacity?

Did you go until 2.5V cell dropout?

 

[Alkaline]

Wow those results are bad! I say recharge and run it again just to be sure. Sunshine looks like the lowest cell went to 2.58 then it looked it like bounced back up on to down to 2.6 again.

Thats a neat graph, can you use the sfk app on a JBD 8s? I downloaded it but it did not detect it.

 

[sunshine_eggo]

Wow those results are bad! I say recharge and run it again just to be sure. Sunshine looks like the lowest cell went to 2.58 then it looked it like bounced back up on to down to 2.6 again.

Thats a neat graph, can you use the sfk app on a JBD 8s? I downloaded it but it did not detect it.

Thanks. I didn't scroll.

 

[Sam Cho TX]

I have been recharging to retry the test, but now the balance is completely off, its been running a charge for 18 hours now, I'm using a 50 amp lithium charger so I know it should have only taken max 6 hours to charge. 1 cycle is all it takes to throw off the balance on grade B?

 

[RCinFLA]

You will have a tough time keeping them in balance if you run a near full discharge, more than about 230 AH DoD. Absolute max discharge will be limited to ballpark of 280 AH, after which you will need to do some significant rebalancing. Only as strong as weakest link. Greater load current may make things worse.

You have two really poor cells from the data you attached, 304.1 and 303. Only 305 comes close to looking like a 304 AH cell. I would definitely leave 303 out of the series stack as it will cause greater misbalancing of cells during deep discharge.

 

[Alkaline]

Under that 62 amp test the 303 cell only tested to 267, yet was testing to 306 under a .1c load did you see the result from the bms, I can't say I have seen such a big drop going form .1c discharge to .2c discharge.

 

[Checkthisout]

My problem with your tests, how do you know your equipment is accurate to the margin you're trying to test to?

What makes you think your setup is capable of doing 100% accurate tests and that you are doing the tests properly?

 

[Alkaline]

OP did you contact 18650 what did they say?

 

[Alkaline]

My problem with your tests, how do you know your equipment is accurate to the margin you're trying to test to?

What makes you think your setup is capable of doing 100% accurate tests and that you are doing the tests properly?

Its the zke-40 tester many of use, I'm quite certain people have said its almost as good as high dollar ones. Still the ZKE tester produced much better results than the inverter because the 303 has some strange graph its not going down like usual cells instead it seems to be slanted.

 

[Checkthisout]

I dunno. It would seem like accurate testing would require overcharge and then going down to below comfort level.

And then look at the ending voltage.

How does the factory test them?

 

[Alkaline]

This voltage curve is very strange, the cell that is dropping; he tested to 306 AH on the zke at 30 amp discharge, but RCinFLA called it, said it was a bad cell before he tested with his inverter.

 

[Sam Cho TX]

You will have a tough time keeping them in balance if you run a near full discharge, more than about 230 AH DoD. Absolute max discharge will be limited to ballpark of 280 AH, after which you will need to do some significant rebalancing. Only as strong as weakest link. Greater load current may make things worse.

You have two really poor cells from the data you attached, 304.1 and 303. Only 305 comes close to looking like a 304 AH cell.

View attachment 143243

can you explain a bit about the 303 cell you said has some graphite problem, this was the cell that tested poorly when I tested it with a BMS and inverter. I'm still puzzle how you were able to figure this out before I did my test.


It tested to 306 AH when I tested at 30 amps with the zke, but very poorly when with the jbd 150 bms and 62 amps.


I think I'm done with grade B, this is just not worth the savings, you basically have to buy 1 extra cell as a stand by, at that point you are the same price as the EV grade.

 

[Sam Cho TX]

OP did you contact 18650 what did they say?

I sent them an email, will try to call tomorrow, but if I have to pay to ship back and get a 20% restocking fee I'm just not going to bother and just take the loss and move.

I'm planning on buying 64 cells for a large 48v bank, I also have luyuan 304 matched and batched genuine on order 4 to try out.

 

[Bud Martin]

I sent them an email, will try to call tomorrow, but if I have to pay to ship back and get a 20% restocking fee I'm just not going to bother and just take the loss and move.

I'm planning on buying 64 cells for a large 48v bank, I also have luyuan 304 matched and batched genuine on order 4 to try out.

Also sent them the link to this thread.

 

[upnorthandpersonal]

I dunno. It would seem like accurate testing would require overcharge and then going down to below comfort level.

And then look at the ending voltage.

How does the factory test them?

No, as per standard testing for LiFePO4: charge to 3.65V and saturate (tail current), discharge until the voltage drops to 2.5V. Discharge rate: 0.2C.

 

[RCinFLA]

When you use a sinewave inverter for a test load you have to be sure the voltage and current measurement instruments are okay with some 2x AC line frequency ripple due to the DC input current profile from the sinewave inverter.

The 2x AC line frequency peak to peak current from a sinewave inverter DC input will be up to nearly twice the average DC current which can cause some battery performance variance compared to a pure constant DC current.

Battery monitors have a low pass filter in the range of 1 to 2 Hz to average out the 120 Hz ripple from 60 Hz sinewave inverter.

The input capacitors on inverter DC input have little impact on filtering the 2x line freq ripple current. They are for filtering the high frequency PWM switching currents. Only at very low inverter load power with low DC average current do they reduce the 2x line frequency ripple current.

You would need several Farads with very low ESR resistance supercaps on inverter DC input to have an impact on inverter DC input 2x AC line freq ripple current.

 

[RCinFLA]

I sent them an email, will try to call tomorrow, but if I have to pay to ship back and get a 20% restocking fee I'm just not going to bother and just take the loss and move.

I'm planning on buying 64 cells for a large 48v bank, I also have luyuan 304 matched and batched genuine on order 4 to try out.

Hard lesson on the crap shoot of grade B cells. As Forest Gump said, 'Like a box of chocolates.'

 

[Kornbread]

Greater than 80% discharged.,

You mentioned earlier about one of the cells not being cycled enough and that capacity might increase with the number of cycles; can these be easy cycles or does the cell/cells need to sustain a certain charge/discharge rate for a certain length of time and number of cycles?

 

[RCinFLA]

You mentioned earlier about one of the cells not being cycled enough and that capacity might increase with the number of cycles; can these be easy cycles or does the cell/cells need to sustain a certain charge/discharge rate for a certain length of time and number of cycles?

I gave a couple of possible reasons for cell 303 discharge curve, one of which was possible improvement with cycling. More likely is the anode was manufactured with thinner graphite paste printing yielding lower anode capacity.

The graphite paste printing of the anode on copper foil is only 6-8 mils thick so keeping tight distribution on print thickness is not easy.

There is a normal manufacturing tolerance to electrode printing, and it can be compensated with a few more or less cell lamination wraps. Negative anode graphite thickness has to be ratioed to match positive cathode LFP thickness, however. A thin print anode has to be matched up with thin print cathode for the layer wrap compensation to be effective.

All requires attention to details on process control during manufacturing.

 

[Checkthisout]

No, as per standard testing for LiFePO4: charge to 3.65V and saturate (tail current), discharge until the voltage drops to 2.5V. Discharge rate:

What if my load tester is off by .1 amps or there is outside interference or only discharge to 2.51 volts etc.

 

[RCinFLA]

What if my load tester is off by .1 amps or there is outside interference or only discharge to 2.51 volts etc.

AH calculations accuracy relies on accurate current measurements and accurate sample time increments. WH accuracy also needs accurate voltage increments measurements.

Every measurement has a tolerance variability on accuracy. When cumulating small time increments, any resolution rounding error can add up.

For testing cell capacity, you also have to be aware of temperature during testing. Below 10 degs C (50 degs F), the cell will deliver less capacity with greater cell impedance that increases in degradation the colder the cell is. You should record the ambient temperature during testing to ensure apples to apples comparison to any test results taken at another time.

 

[Alkaline]

Well seems like the old adage of coin toss with grade B is still at play here, too bad I thought 18650 had cracked the code, but looks like they are subject to the same realities as the rest.

Good for them for disclosing these are grade B upfront, however, guaranteeing pulling rated capacity is going to probably cause them more problems. Maybe they were good when tested in China, but by the time they arrived here not so good.

Either way, OP I see you are in Austin, I'm about 500 miles from you, I have the zke-40 as well so If you want me to test these for you and can meet half way or something I will be glad to compare.

As for me, grade A Rupi 280s it will be when I'm ready to build a pack for a 48v system.