Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Current Capacity Tester Meta (which to buy?)
#21
I’ll never discharge my pack with 1 amp per cell. So 250 or 500 gives me a more accurate ah rating. (I have a lot of crap cells)
Reply
#22
(11-21-2019, 11:38 PM)camthecam Wrote: I’ll never discharge my pack with 1 amp per cell. So 250 or 500 gives me a more accurate ah rating. (I have a lot of crap cells)

Yes but won't give you any indication if that one time you do go there, that cell won't overheat and at the very least damage it further at the worst fire. Finding all crap cells before build is easier than after PLUS your taking 4x longer to test cells at 250ma (thats madness)
100kwh-hunter likes this post
Save 10% at Battery Hookup with code 'Powerwalls' Just enter it in checkout www.batteryhookup.com
Reply
#23
(11-21-2019, 11:38 PM)camthecam Wrote: I’ll never discharge my pack with 1 amp per cell. So 250 or 500 gives me a more accurate ah rating. (I have a lot of crap cells)

Accuracy means nothing if your cell cannot sustain proper current.
If you want true capacity accuracy you will have to test the cell at the exact discharged current specified by the manufacturer. Only then you will have 100% accuracy. And for that you will probably need a very expensive tester. Testing at higher than spec current will show you less capacity and testing at lower current will show you more.

I test all cells at 1A discharge to actually filter quality cells from junk one. Cells passing the capacity and current mark are registered as premiums while the rest will be fit in lower categories.
100kwh-hunter likes this post
Reply
#24
Quote:Finding all crap cells before build is easier than after PLUS your taking 4x longer to test cells at 250ma (thats madness)

I'm onboard with you there.

Capacity testing cells, to my priority, is, oh, 70% filtering out bad cells and 30% determining the capacity of a cell. And with respect to determining the capacity of a cell, I only care about the precision of the test (consistency of results), not the accuracy (closeness to reality).

To me, there is almost zero benefit to testing at anything other than the maximum discharge rate the capacity tester can do.

Thinking about this has helped me define some basic principles, feedback welcomed:

1 - Figuring out bad cells is the main goal, figuring out capacity is a secondary goal.

2 - Measuring voltage when cells arrive is going to help you reject 90%+(?) of the bad cells outright.

3 - Charging cells and filtering out heaters or those that never reach high voltage is going to filter out another 5%(?) of bad cells. Without capacity- or discharge-testing, you've already determined 95%(?) of the cells to reject.

4 - Capacity-testing cells and regarding capacity alone (not heat), is going to determine another 4.5%(?) of cells that passed previous tests. So now 99.5% of bad cells have been found.

5 - Stress testing cells at higher discharge will help (by heat) figure out all of the bad cells, but this will almost always show up in the capacity results as well.

6 - On energy-prioritized cells (laptops), if there's a significant failure difference between 500mA and 1000mA discharge, then you should be using the 1000mA. The discharge rate is a significant portion of their max discharge and thus they're actually being stress tested a bit. Figuring out if the cell has gone bad is the biggest reason you're testing them.

7 - On power-prioritized cells (tools), there is unlikely to be a significant failure difference, because even 1000mA is minuscule compared to their ability (30,000mA) and not likely to rule anything out that would not be redundantly also be determined by lowered capacity.

8 - Presuming the general failure rate of your batch is relatively low, (say, ~5% as mine is), and higher stress test rejects more cells, it's not too impactful on your kWh accumulation to reject those that fail only at higher stress rates, so most of you might as well. If your failure rate was 60% at 500mA, and jumped to 90% if you tested at 1000mA, well then I can see frustration of your accumulation taking too long and wanting to squeeze more kWh by not rejecting marginal cells. But that's abnormally high for most people's batches.

9 - There may be a few people who's use case will allow them to keep cells that choke at 1000mA but can handle 500mA, but, I think that symbolizes cells that are on their way out, and that's flirting with diagnostics problems in the future (if not outright fire/failure). So even those that might be able to use those cells, might not end up being effective in the long term when those marginal cells accelerate their failure rate even under their non-challenging use case.

10 - Very few projects require ultra accurate capacity measurements. This would only be projects where you just barely build enough capacity for a task, within single digit percentage. Maybe some weight or bulk prioritized projects would be valid here (e-bike drag racing?), but that use case is going to murder cells by pushing them to 100% anyway. Also, those projects probably wouldn't be using recycled cells. Most projects only require moderately precise capacity measurements so that the rePackr will design a balanced pack. So testing at lower mA rates is not as necessary. You'll get the same result faster by testing at higher rates.

11 - Unless you lack desktop space, "$$ per cell slot per cycle time" is the key metric as to which charger you should buy. I.E. $20 for 4 cells at 4 hours/cycle is equivalent to $40 for 4 cells at 2 hours/cycle because you need 2 of the former to match 1 of the latter. Presuming you have room for twice as many chargers at your workstation. If you capacity test while you sleep, then fast cycling time matters less because the faster ones aren't getting filled during the night and any capacity tester is done by morning anyway.

...


Quote:Even though the "tester" shows 1000mA on the display, in the background it cycles the discharge between 500mA to some spikes as close to ~2.5A.

A few notes:

 - This variance only matters if heat accumulation is non-linear. I.E. 1 minute at 2A, 1 minute 0A, is not equivalent to the average of 2 minutes a 1A.
 - Even if non-linear, we're seeing variance flicker multiple cycles per second. Heat accumulation is probably no different than steady discharge at this quick of a rate.

Interesting stuff though.
Reply
#25
(11-22-2019, 07:41 AM)MattsAwesomeStuff Wrote: Thinking about this has helped me define some basic principles, feedback welcomed:

1 - Figuring out bad cells is the main goal, figuring out capacity is a secondary goal.

2 - Measuring voltage when cells arrive is going to help you reject 90%+(?) of the bad cells outright.

3 - Charging cells and filtering out heaters or those that never reach high voltage is going to filter out another 5%(?) of bad cells. Without capacity- or discharge-testing, you've already determined 95%(?) of the cells to reject.

4 - Capacity-testing cells and regarding capacity alone (not heat), is going to determine another 4.5%(?) of cells that passed previous tests. So now 99.5% of bad cells have been found.

5 - Stress testing cells at higher discharge will help (by heat) figure out all of the bad cells, but this will almost always show up in the capacity results as well.


@MattsAwesomeStuff,

Number 1- is right on the money,  "Figuring out bad cells is the main goal"

Number 2- yea measure voltage we all do that (at least I assume we all do) as a first step then I would just add an IR reading and now you have 2 parameters to judge the cell by. The YR1030 or equivalent does this in a matter of seconds and you can eliminate 99% of the bad cells.

If you follow best practices as far as IR (Charts are published here and here) and V is concerned, numbers 3,4, and 5 are practically eliminated because once you have proper IR and V on any cell you will not have any heaters/ SD or low capacity results.
Charging and testing these cells becomes a pleasure instead of an aggravation as you will have really good results.


Quote:
Quote:Even though the "tester" shows 1000mA on the display, in the background it cycles the discharge between 500mA to some spikes as close to ~2.5A.

A few notes:

 - This variance only matters if heat accumulation is non-linear. I.E. 1 minute at 2A, 1 minute 0A, is not equivalent to the average of 2 minutes a 1A.
 - Even if non-linear, we're seeing variance flicker multiple cycles per second. Heat accumulation is probably no different than steady discharge at this quick of a rate.

Interesting stuff though.

Yea I thought it would be interesting to run these testers through their self claimed 1000mA cycle and get a true reading.  As the cell is removed from the charger a more accurate temperature can be extrapolated during the charge discharge cycle.
After all the testers have been charted I will run a chart on a CC/CV charger and a ZH-YU ZB206+ CC discharger at 1A and see if there is a temp difference.

Stay tuned
Wolf
If 18 X 650 = 2200+mAh then we have power! 
May all your Cells have an IR of 75mΩ or less Smile
Last count as of 8/7/2019
Total Number of Cells Recorded and processed                 6149
Total Cells required for PowIRwall                                   2856
Total Cells ≥2200mAh, ≥80%, ≥35mΩ, ≤75mΩ, ≥4.12V   2760
For Info Google Drive
Not your average Wolf       
            Cool
Reply
#26
(11-22-2019, 07:40 AM)Overmind Wrote: Accuracy means nothing if your cell cannot sustain proper current.
If you want true capacity accuracy you will have to test the cell at the exact discharged current specified by the manufacturer. Only then you will have 100% accuracy. And for that you will probably need a very expensive tester. Testing at higher than spec current will show you less capacity and testing at lower current will show you more.

I test all cells at 1A discharge to actually filter quality cells from junk one. Cells passing the capacity and current mark are registered as premiums while the rest will be fit in lower categories.
@Overmind,

I suppose a SkyRC MC3000 at ~$110.00 see link that can be programed to your hearts content is a little more expensive but not very expensive.
I must respectfully disagree with the discharge rate making a major difference in capacity though.
I know you did not say major  but you said less and more.

Looking at a standard CGR18650CG the recommended charge is 4.2V max 1500mA  with a 110mA cut-off. The "standard discharge" current is 430mA with a 3.0V cut-off. So discharging this particular battery at 1A is not how the manufacturer rated the battery.

That being said the manufacturer does provide additional discharge scenarios with their typical discharge characteristics chart showing a discharge rate of 2.0C 1.0C and 0.2C.  The difference between 2C and 0.2C is 150mAh at best. So a deviance of ~5%. Now temperature does have an effect that is more noticeable but still at only about ~10%.

A 1A charge/discharge rate will most certainly weed out a bunch of underperforming cells that is granted.
A proper IR test will guarantee that the cell will perform well at that rate and it will like it by giving up a good mAh result and a limited amount of heat.

Wolf
If 18 X 650 = 2200+mAh then we have power! 
May all your Cells have an IR of 75mΩ or less Smile
Last count as of 8/7/2019
Total Number of Cells Recorded and processed                 6149
Total Cells required for PowIRwall                                   2856
Total Cells ≥2200mAh, ≥80%, ≥35mΩ, ≤75mΩ, ≥4.12V   2760
For Info Google Drive
Not your average Wolf       
            Cool
Reply
#27
As promised the full chart of the Foxnovo top up charge, discharge and recharge.
Settings on the Foxnovo are 1A cap test
The Foxnovo does a continuous 500mA discharge to 2.86V then a ~1A charge to a cutoff of 50mA at ~4.2V
Next the LiitoKala

Wolf

Chart Foxnovo C/D/C chart
If 18 X 650 = 2200+mAh then we have power! 
May all your Cells have an IR of 75mΩ or less Smile
Last count as of 8/7/2019
Total Number of Cells Recorded and processed                 6149
Total Cells required for PowIRwall                                   2856
Total Cells ≥2200mAh, ≥80%, ≥35mΩ, ≤75mΩ, ≥4.12V   2760
For Info Google Drive
Not your average Wolf       
            Cool
Reply
#28
@Wolf in all high current cells, any low current test is nearly irrelevant because it can show you good capacity but not sustain the discharge current.
For normal current cells, yes, the difference is small.

Thank you for the tester link.
100kwh-hunter likes this post
Reply
#29
(11-25-2019, 08:12 AM)Overmind Wrote: @Wolf in all high current cells, any low current test is nearly irrelevant because it can show you good capacity but not sustain the discharge current.
For normal current cells, yes, the difference is small.

Thank you for the tester link.

@Overmind
You are spot on which again brings up the subject of IR  High current cells have a different chemistry which reduces the internal IR of the cell to <35mΩ. Most new "high drain" cells run in the < 20mΩ range.
So if I understand you correctly................................. if you have a high drain cell with higher than normal IR it may produce a reasonable mAh result at low drain currents (0.2C) and fall flat on its face at a 2C discharge. Interesting. I will have to chart that on a graph.

So for the next experiment I will take an ICR18650-22P (basicly a hybrid of a LD and HD cell)  with a max discharge current of 10A and a measured IR of ~22mΩ and run a .5 A, 1A, and 3A discharge and see what happens as far as capacity is concerned. I am going to have to make some modifications to my testing environment as the max discharge current for my system is 3.2A per cell right now if you feel we need to go to 10A.

Then I will try to find a high drain cell with higher than normal IR. That's going to be a trick but I think I got some INR18650-13Qs and a couple of them have a higher than normal IR. More testing and graphing I love it.
If you have a cell preference for me to test give me the part number and if I have it I will test it and if not I will buy one.

Now back to the OP "which tester to buy."

OK here is a running chart of the most popular testers running a test on the same battery.
All tester where set to the highest C/D/C setting.
Results are:

The Foxnovo set to 1000mA capacity test does a charge at 1A but a discharge at only .5A Temperature was quite steady and the discharge and charge lines where very clean. Results 2633mAh

The OPUS set to 1000mA capacity test does a charge at ~1A and a discharge at ~1A(more like ~1.5A) I ran this test twice on 2 different chargers on different slots as the results vere very odd. But as you can see the curves are the same on both chargers in different slots so I guess that is the algorithm that that charger/tester uses. The temperature as expected was slightly elevated (about 5°C above ambient) during the discharge cycle.
Results 2723mAh

The LittoKala set to 1000mA does a discharge at 500mA and a charge at 1000mA The discharge trace looked to be a bit noisy but charge was relatively clean.
Not much as far as temperature variation. maybe 1°C to 2°C from ambient.
Results 2598 mAh

The Zanflair set to max which is only 500mA for charge and discharge. Very clean discharge trace but charging is very noisy.
Also temperature is negligible
Results 2830 mAh

The SKYRC is infinitely adjustable so I set it at 1A charge and 1A discharge. Very clean trace on both charge and discharge Temperature rise was expected at a 1A discharge and the cell did not disappoint ~3°C above ambient.
Results 2561mAh

The iCharger x6 was set to 1A charge and 1A discharge kind of a noisy discharge and charge trace. I may do this one again.
Temperature as expected at a 1A discharge
Results 2277mAh

If I had the money SKYRC all the way.
IF I want to stress the cells OPUS does a good job hence all the complaints about the heat.
I do like the Foxnovo as it is very clean but does not stress the cells at 1A so I count on IR of the cell to guide me in this scenario.
LittoKala and Zanflair are on the bottom of my list not worth the money even at half the price. IMHO.

There you go and all the graphs to back up the conclusion.

Wolf
If 18 X 650 = 2200+mAh then we have power! 
May all your Cells have an IR of 75mΩ or less Smile
Last count as of 8/7/2019
Total Number of Cells Recorded and processed                 6149
Total Cells required for PowIRwall                                   2856
Total Cells ≥2200mAh, ≥80%, ≥35mΩ, ≤75mΩ, ≥4.12V   2760
For Info Google Drive
Not your average Wolf       
            Cool
Reply
#30
I have to go with hb powerwall and overmind, plus one and plus one.
For the ir tester, its a GREAT device, it will take a lot of headache away.
Aldo i used it slite different(10% max deviation of cell manufactory optimum, those are my keepers, instead of a fixed maximum number, no offence wolf), i would also go for that one.
I wish wolf had done his magic half a year earlier, it would have saved me a lot of time, but better late than never.
While we are still in kindergarten with testing, wolf is writing essays for the universities.

Anyhow after a ir check you really want to charge and discharge with 1a or what your battery is going to have.
If a 10kwh battery would be fed with 50kwh solar panels or would be drained when making coffee with the oven on and the electric heater jumps on, ect and you tested them on 250ma, and now they must give you 10a a cell, or take 10a when charging.
If you test at those low numbers make sure your pw will never see numbers higher than that, no matter if it is in charge or discharge.
Or make a really big pw minimum 60-70kwh, then i think you can test and get away with 500mah.

My 2 cents, sorry
Oz18650 likes this post
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)