Bulk Capacity Testing

DrkAngel

1 GW
Joined
Dec 15, 2010
Messages
5,300
Location
Upstate-Western-Southern Tier NY. USA
Note!
Read entire thread!
"Methods" improve-evolve as thread progresses.


Update ... See - Latest cell testing methodology

Large Quantity Comparative Capacity Test
...... Method 1 ......


Recommend performing Bleed Down Test 1st!(To eliminate most bad cells)

Premise 1 = Some 120V AC power tools-devices will run on 120V DC.
Premise 2 = Cells in series will discharge Amps equally.
Build a single 30s of li-ion cells = ~120V DC (60 cells per batch, if left in pairs! )
30s x 4.1V = 123V DC output
28s x 4.2V = 117.6V DC output
(Danger - 120V can be dangerous!) - Take care!
Discharge at moderate rate (120w light bulb (2x60w) 1A, recommend a .25 - .5C discharge rate)
Amps will be drawn from each cell equally!
Check voltage of each cell at intervals (every 10 minutes?)
Discharge until worst cells hit max DOD (if some cells reach DOD very quickly, remove and continue test at lower voltage)
Residual voltage will give an excellent comparative capacity rating!
 
Comparative IR Check

While cells discharging, immediately check for voltage sag!
This will give you a comparative IR (Internal Resistance)
Mark IR on each cell, or pair - .24V, .50V etc.
A very high IR usually means 1 cell of pair bad!

I used a 2.5A 12V fan to test 3s2p cells (pulled from Dell workstation - careful! can levitate! fast sharp blades can remove tips of fingers!!! )
 
arkmundi said:
Thanks for the suggestion. Is this something you've done successfully? :?:
Not yet.
Next logical step-application of method I used for Finding Weak Cells.

Moving into tests on recycled 3s2p recycled Lipo.
First 20 6packs estimated as all 3.000Ah+.(1Ah discharge test left plenty of voltage.)
Passed bleed down test - maintained above 4.10V after 3 months.
Cells in each 6pack still nicely equalized after 25% discharge.
Recharged to 4.10V.
Now will capacity test as 30s2p with 120w light bulbs. (123V DC output)
28s recommended if charged to 4.2V. (117.6V DC)
OEM rated at 4.320Ah per pair.

Test will confirm 1Ah capacity per hour of discharge.
Will stop test if any cell reaches DOD of 3.7V.
Otherwise I will stop test at 3 hours and classify 6packs by residual voltage.
Another 20+ 6packs of non-equal, less than great, will be a more "fun" test of the method.
 
Just ran my 1st trial of my new prototype Bulk Capacity Test.

Discharged 30s2p using a 120w incandescent light bulb.

Cells
50 3s2p (11.1V 4.320Ah) recycled Li-ion secondary bay batteries (CD bay)
These cells were pulled from the packs that failed to charge properly.
All packs were 1st tested to maintain good voltage equally among all (6) cells.
(bleed down test and all cells equal during IR test)
47of the 6packs performed well.
3 additional (slightly defective) packs were added to complete the 5th 30s2p test rig.


IR (Internal Resistance)
Packs were tested for a comparative IR.
Resting voltage was recorded.
A 2.5Ah drain was applied.
Voltage was recorded after voltage sag stabilized.
Resting voltage minus discharging voltage was noted as the (comparative) IR.
IR varied from .5V to 1V.
10 6pack testing packs were grouped by similar IR.

Results
1st 2 test rigs discharged 2.5Ah before the 1st 6pack reached 3.3V per cell (<10V per 6pack) under discharge (recovered to near 3.7V when discharge stopped - I used this as my extraction voltage)
Of the 48 "good" 6packs:
1 removed at 2000mAh
9 were removed at 2500mAh (rated against each other by residual voltage)
36 stopped discharge at 3000mAh

Resultant Voltage - Of the 36
3000mAh+
6 - >11.1V - 36 matched cells - fair
12 - >11.2V - 72 matched cells - good
12 - >11.3V - 72 matched cells - very good
6 - >11.4V - 36 matched cells - excellent
(12.27V beginning voltage)

Resultant Voltage - Of the 9
2500mAh+
1 - >11.1V - 6 matched cells
2 - >11.2V - 12 matched cells
6 - >11.3V - 36 matched cells
(12.27V beginning voltage)

Notes:
Removed 3 6packs from each test rig at the 2500mAh point so as to maintain similar "equal" discharge rates between "rigs".
(Lowering voltage also reduces amps of discharge?)

IR was fairly irrelevant in predicting cell condition-capacity!
The poorest residual capacity was attained (largely) by the .6 IR, while the best were predominantly the .8 IR.
There was a fair mix of various IRs throughout the entire range of capacities.

Older formulations?
These cells are stamped with date of manufacture.
The older cells seemed to have a better IR though displayed lower capacity.
Logically, this seems to indicate that the older cells used a higher percentage of Anode-Cathode vs Electrolyte.
As capacity (long run time) demands increased the need for high drain cells dwindled.
Perhaps the best of these older cells will be perfect for my power tool power pack rebuilds?
Cells tested in B&D Firestorm 14.4V and 18V packs. Work great in drills and weed eater.
14.4V 1.4A nicd < 14.8Ah 4.32Ah Lipo
18V 1.4Ah nicd < 18.5V 4.32Ah Lipo
(14.8V 5.2Ah 18650 Li-ion cells worked nicely!)
Special charger required!!!
 
I am happy with my preliminary results.

But am still looking to improve-perfect this method.

Large Quantity Capacity (mAh) Test
...... Method 2 ......


Next advancement
The use of a volt, watt, mAh meter.
This will allow an accurate mAh capacity rating of each cell.

Problems:
1. My present meters "reset" when I disconnect power to remove a cell.
2. My present meters are limited to 60V DC.

Solution! - My present budget ($13.69) Volt, watt, amp, mAh meter is also 3-wire capable.
This requires an additional battery to power meter functions but maintains all functions and settings while disconnected from metered cells.
Connection points for dedicated battery clearly marked under back panel.
4V is insufficient 8V works 12V recommended . Maximum 3wire voltage ... unknown ... 90V?
UPDATE - remove solder "jumper" on JP2 to isolate meter battery from test load voltage.
.... Obsolete but informative --- (((WARNING! I hooked up 7.2V Lion to the external power input.
Hooked to 11.1V power pack - to test.
DANGER! Got an 12V+ feedback into the external 7.2V power pack. DANGER!
A small diode, on the positive, between the battery and meter solved the problem.
(a diode works as a one way valve, allowing electrical discharge but preventing charging - there is a noticeable loss of voltage through a diode .2-1V ))) Obsolete ....

Listing states 80V limit but instructions and label indicate 90V. (VAM9020)
DC-DC 3.8V-60V (80V if 3wire) 20A Dual-display Digital LCD Power Current Voltage, AMP Meter, watt meter, Ah meter $13.69



This will allow up to 21 cells in series. (4.1V-4.2V)
Individual cells can be extracted from series and labeled with mAh of capacity.
Voltage reduction might vary drain amps but mAh meter function will provide an accurate reading.

Test can be accelerated by adding additional light bulb(s).
 
At 75V DC 100w bulb only drew about 50w, so I dug up an old track light and plugged in 2 100w and drain is now back to ~100w.
Draining 5200mAh cells, so might take 4-5 hours.
 
Just ran a 24cell capacity test.
Took about 4 hours with preliminary cell voltage monitoring - every 1/2 hour.
In the 3rd hour I began testing at 10 min intervals, then sooner.
The 36 2600mAh cells were left in 3s2p 6packs.
1st cell pair hit 3.3V discharging voltage at 3300mAh.
1 - 3300mAh
5 - 3700mAh
2 - 3800mAh
2 - 4100mAh
2- 4200mAh
2 - 4300mAh
4 - 4500mAh
18 pairs thoroughly tested and accurately rated for capacity in 4 hours with minimal attention till the last hour.

Improvements?
While the Lipo in preliminary test showed a good matched capacity among all cells in every "6pack", the 18650 Li-ions 6packs showed a great variance in capacity.
This might indicate a higher degree of precision-quality control in Lipo manufacture?
Future 18650 packs will likely be separated into pairs for testing.

Will find a lighting fixture with more bulb capacities, for additional-faster drain.
This will allow the test to take less time, especially during the preliminary discharge till 1st pair removal.
.5C maximum discharge rate recommended (2.6A for`5200mAh cell pairs)

Since the C rate of the discharge varies as more cells are removed, discharge should be disengaged before cell voltage is measured. This will eliminate the cell voltage sag variance and insure that all cells will be removed at the same DOD.

Use of a timer as a reminder of when to check cells again.

Will test the standard 120V AC light dimmer for ~120V DC capability. This will allow "tuning" DC amp draw to maintain same drain rate as additional cells are removed from series.
 
120V mAh meter available.
Allows s27-s30 (4.1-4.2V) 54-60 cell (in pairs) bulk capacity testing.

120V 25A volt, watt, Ah + battery %
- $24.50
 
Silly me!
Thinking that a 90V limited mAh meter could only test a 21s Li-ion string.(21 x 4.2V = 88.2V)

18s Li-ion as originally configured, through mAh meter draining through 100w light bulb.
Disconnect neg from light bulb, attach to pos of additional 12s Li-ion, run wire from neg of 12s to neg of light bulb.

This will supply ~120-125V to bulb.
MAh drained will still be equal throughout all cells.

1A (100w bulb) drain through 4.32Ah and 5.2Ah paired cells. = .25C - .2C discharge rate.

60 paired cells capacity rated per batch per ~5 hours.
Add 2nd 100w bulb for 2-3 hours per batch.

Cautions:
120V DC can be dangerous!
As cells near low voltage point, voltage drop is rapid, check cells often. Comparatively high amp drain on low voltage cells will cause a rapid heat rise. I pulled cells when 1 amp drain demonstrated as 3.3V(nearly 3.7V when discharge removed).
 
What is your method for putting the cells in parallel during testing? (to charge to equal voltage). Series would be easy with alligator clips.
 
veloman said:
What is your method for putting the cells in parallel during testing? (to charge to equal voltage). Series would be easy with alligator clips.

Modded a S-150-5 MeanWell to 4.2 30A.

Usually I solder tabbed pairs of similar voltage together using light gauge wire.
file.php


These packs had wires connecting to a pack BMS, I attached these wires to heavier gauge charging "rails". This allowed me to easily cut out the few cells that produced heat. (For thin light gauge wire I used to cut with toe nail clippers, now I use dog or cat nail clippers-shears at $1 each from the local Dollar Tree dollar store.) I moved one rail to the next "p" (parallel) after charge attained.

file.php


Added Note:
- As pictured, with pos & neg connected near center of rails, center cells charged faster than cells on the ends. I "equalized" charging method by moving pos to far left of "rail" and neg to far right of "rail". This seemed even up the charge process.

<$15 IR meter kept handy for quickly checking for any heated cells.


Due to the low energy content below 3.6V, I gang together all cells of 3V to 3.6V state of charge, energy transfer is of low amperage.
At above 3.6V a I gang together cells within a couple .1V with no concern of excessive charge or discharge ... the advantage of the laptop 18650 modest C rate capability.
 
just a note to say i'm keeping an eye on your thread, I have a sh** load of cells that have been sitting here for many many years ( 7 or more years), that I need to test, your method might be very useful. I know its generally considered a cell that has gone to zero voltage is useless, but surprisingly some of the cells that have lost all voltage ( at a very slow rate of many years), have come back up to good voltage, to my surprise, haven't tested the ahr capacity of those cells but they took a long time to charge so they might be ok. type of cell is lithium cobalt, thanks for puting up all the details
 
Thanks for the info.

I am trying to find a discharge graph for NiCoMn 18650 cells. I mostly just want to know how much energy is below 3.6 or 3.7v resting. I've read that with lipo, there is very little below 3.7 and you shouldn't drop below that (resting).
 
veloman said:
Thanks for the info.

I am trying to find a discharge graph for NiCoMn 18650 cells. I mostly just want to know how much energy is below 3.6 or 3.7v resting. I've read that with lipo, there is very little below 3.7 and you shouldn't drop below that (resting).
Every cell type and variation has a differing discharge-capacity profile.
Each brand, manufacturer, formulation ... even batch ... varies

I have used several methods to determine the capacity profile of various cell types.

Easiest Method
The easiest requires a voltage adjustable power supply, an Ah meter and a precise voltmeter.
It clearly demonstrates the capacities available from discharging deeper, as well as the additional capacity added by charging to higher voltages.
Test was performed using 1/10th (.1V) volt increments. I would recommend using a 1/20th V (.05V) increments (or 1/50th - .02V).
Finer, more precise, measurements take longer (more steps) but test can run unattended, noting mAh then adjusting next voltage step.
See -Determining Capacity Profile - Simple Method
DrkAngel said:
Tools -
MeanWell 24V bulk charger (19.8 - 29.8V adjustable)
30V 4 digit volt meter, 100ths capable
Ah meter

I discharged my 25.9V recycled Lipo pack to 24.5V.
Then precisely equalized the cells at 3.50V.
I applied charge with MW (MeanWell) set to:
25.20V (3.60V) - full charge required .27Ah
25.90V (3.70V) - full charge required .53Ah
26.60V (3.80V) - full charge required 3.87Ah
27.30V (3.90V) - full charge required 3.15Ah
28.00V (4.00V) - full charge required 4.60Ah
28.70V (4.10V) - full charge required 4.10Ah
29.40V (4.20V) - full charge required 1.85Ah
(Ah is capacity between each .10V)
----------------------------------------
Total. .................... 18.37Ah
Ah capacity previously confirmed with -
iMax B8 full charge ..... 18.4Ah

My 1st method - Capacity Graph
This requires a volt meter and an Ah meter.
Using the, linked to , graph program plugging your acquired data in will produce an impressive graphical representation.

See - Lithiums - mAh/100th V - Discharge Tests
file.php


Low-Tech Method
The low-tech method requires only a pencil and paper.
If you plug the data into the graphing program you can produce a graph similar to the familiar "discharge graph" typically provided by manufacturers.
See - Low-Tech (No Tool) Method
 
another method is using icharger and the logging software ( i forget the name of it), you can use globes to discharge and gets graphs
 
Correction ... errr ... refinement!
120DC 100w (100w light bulb) discharge was previously approximated as 1000mA (1A).
120V DC 120w (2-60w light bulbs) discharge, is an accurate 1000mA discharge rate.
1000mAh hour of capacity per each 1 hour discharge during bulk discharge capacity test.


Next, I intend on bulk capacity discharge testing, and then capacity mapping, my 4 year old "new" Lipo 3s2p 4320mAh 6 packs.
15-6 packs per eZip build, I have 30-6 packs and will bulk test in 3-30s2p "packs".
More to confirm of equal capacity than to rate capacity, 7 to 9 year old used cells of this type, perform excellently!
See - Capacity mapping by charge mAh ... completed.
2004 & 2006 Lithium Cobalt formulations seem identical or very similar.
It will be interesting to note any change in the 2009 manufacture.
I also have many similar 6 packs of Lipo, with cells of "genuine" SONY manufacture, that I will "capacity map".
 
Just received a "DC 120V 25A Amp Volt Combo Meter Battery Monitor Capacity Power Charge Discharge" - $24.50 shipped.

Now, 30s2p bulk capacity testing is extremely simple. - Method 2
Best of all this meter came with documentation.
Circuit board is identical to my 90v previous model.
Turns out that when a dedicated meter battery is used, the solder connecting JP2 points must be removed.
(I placed a diode in-line to prevent the voltage backwash on the 90V ... no documentation, so I didn't know JP2 had to be disabled.)
 
Just found the picture of my 125V DC "stack".
10 - 6 packs of 3s2p recycled Lipo, configured as 30s2p.
30 x 4.15V = 124.5V
2 - 60w (120w @ ~125V,) light bulbs equals 1000mA discharge rate

file.php
 
Bulk Discharge - Capacity Test
Updated methodology ...
for testing large quantities of cells - more precise.

Step 1
Charge all cells to identical voltage - 4.10V? - 4.20V? (Yes, 100th's is important)

Step 2
Allow all cell pairs to set for several days.
Remove any with a noticeable voltage loss - bleed down - self-discharge, more than ~2/100ths

Step 3
Test each cell pair with a trial discharge.
1 minute through any device with a substantial discharge amperage ~1000mA or higher. iMax B6 time discharge @ 1A, a 1000 lumen rated Cree XML flashlight, or ... .
Measure discharging cell voltage at 1 - 5 min. Mark cell as comparative IR rating, (4.20V-4.14V = .06V). Also note resultant voltage - 2 minutes(?) after discharge stopped
5 minute test could process a dozen pairs, 24cells, per hour and might indicate all cells liable to fail 2 hour discharge test!. (6 minute 1A discharge = 100mAh)
Important to measure voltages at cell with 100th V capable voltage meter, iMax voltage readings are skewed by multiple factors!
Remove any with excessive sag or lower resultant voltage.
This will confirm at least fair capacity and both cells functional.

Step 4
Prepare soldered 28s2p "strings" of cells. 28 x 4.2V = 117.6V DC
Begin the 1000mAh hour discharge using 2 x 60w light bulbs (120w).

Occasionally check voltage on each pair.
If any pairs getting low at 1 hour, (<3.8V? under discharge might not go 2 hours), remove and put in "1000mAh+" pile, save for continued testing with like cells.
Replace with spare pair, or pairs, make sure to tag them -1000mAh?

Step 5
Resume discharging.
Monitor each pair voltages and repeat replacements, (<3.7V under discharge), at 2nd hour put cells in "2000mAh+" pile, (40%+ capacity).

Monitor cells, you might want to schedule 1/2 hour, 500mAh replacements?

3 hours = 60% of original capacity for ~2500mAh cells.
I think, 60% + remaining voltage, to be a satisfactory capacity rating for recycled cells.
Mark pairs as 3000mAh + ( remaining static voltage eg 3.84v).

Pairs are effectively empty if they read 3.3V under a .2-.25C discharge, (= ~3.7V "static" - not discharging).
If unsure of cell rest voltage, disconnect 120w discharge temporarily.

Rebuild and continue testing strings of previously eliminated pairs, if any, test 2000mAh+ first.
Carefully monitor each pair, If they perform well for next 1/2 hour you can label all as 2500mAh+ their remaining static voltages. Every 6 minutes = 100mAh

Preliminary eliminations, 1st hour removals, if any, might not be worth the bother?

Even a 2 hour discharge, + the remaining voltage, is sufficient to build a nicely balanced pack.
Line up all cell pairs, best to worst.
Starting with best, lay them into banks ...
12s = ...
1 2 3 4 5 6 7 8 9 10 11 12 12 11 10 9 8 7 6 5 4 3 2 1 1 2 3 4 5 6 7 8 9 10 11 12 etc
9s = ...
1 2 3 4 5 6 7 8 9 9 8 7 6 5 4 3 2 1 1 2 3 4 5 6 7 8 9 9 8 7 6 5 4 3 2 1 etc
etc.

Unless you have a few exceptionally good or bad cells, this should build banks of nearly identical capacity.

Smaller than 28s will require a watt meter to maintain accuracy

Of course if you add a watt meter to your string you can pull cells at any time.
Make sure you use a 3-wire meter, = self battery powered, or you will lose discharged amount when disconnected to remove cells.

14s2p ~60v or fewer would work nicely, but 1000mAh discharge might require ~300w halogen floodlight.
If you want to maintain a .2-.25 discharge rate.

DC-DC 3.8V-60V (80V if 3wire) 20A Dual-display Digital LCD Power Current Voltage, AMP Meter, watt meter, Ah meter $13.69
I've been using one for several months ... love it!
Many more functions but sketchy instructions are in Chinglish





Still believe this bulk method of capacity testing is not accurate enough?

Well ... then ...
Use bulk discharge method to discharge 28 paired cells for 3 hours (= 3000mAh ~60% of new capacity for pair of 2500-2600maAh cells).
Separate cells and use watt meter to discharge to desired "empty", possibly 1 hour more? (4000mAh = 80% of new rated capacity).

Bulk method would take 2-3 hours for 28 pairs but shows cell pairs of 2000-3000mAh (~40-60%) + resulting voltage, which is great for a comparative capacity between like tested cell pairs. (Discharging a typical sample further, in 100mAh steps and "mapping"resultant voltages, should effectively allow a reasonably accurate mAh rating for all cell pairs in the series!)

Bulk method + metered per each pair would take 3 hours + 28 hours = 31 hours for 28 pairs

Metered discharging each pair from full would take 4 hours x 28 = 112 hours for 28 pairs
 
DrkAngel said:
Step 4
Prepare soldered 28s2p "strings" of cells. 28 x 4.2V = 117.6V DC
Begin the 1000mAh hour discharge using 2 x 60w light bulbs (120w).

Occasionally check voltage on each pair.
If any pairs getting low at 1 hour, (<3.8V? under discharge might not go 2 hours), remove and put in "1000mAh+" pile, save for continued testing with like cells.
Replace with spare pair, or pairs, make sure to tag them -1000mAh?

Sounds like a great way to bulk capacity test cells.
Is it possible to do this test with 230V light bulbs we have here in Norway?
 
kje said:
Is it possible to do this test with 230V light bulbs we have here in Norway?
Yeah ... if you don't think 56s isn't too awkward.
 
DrkAngel said:
Step 4
Prepare soldered 28s2p "strings" of cells. 28 x 4.2V = 117.6V DC
Begin the 1000mAh hour discharge using 2 x 60w light bulbs (120w).

Occasionally check voltage on each pair.
If any pairs getting low at 1 hour, (<3.8V? under discharge might not go 2 hours), remove and put in "1000mAh+" pile, save for continued testing with like cells.
Replace with spare pair, or pairs, make sure to tag them -1000mAh?

The replacement spare pairs won`t be fully tested if they last to the end?

DrkAngel said:
kje said:
Is it possible to do this test with 230V light bulbs we have here in Norway?
Yeah ... if you don't think 56s isn't too awkward.

Yes, maybe it is too awkward. Is it possible to do this substituting the light bulbs with the Imax B6 charger? Prepare soldered 6s2p "strings" of cells. 6 x 3.7V = 22.2V Set it to discharge 1A and follow your description?
 
kje said:
Yes, maybe it is too awkward. Is it possible to do this substituting the light bulbs with the Imax B6 charger? Prepare soldered 6s2p "strings" of cells. 6 x 3.7V = 22.2V Set it to discharge 1A and follow your description?

I tried this and found out that the Imax B6 can`t discharge 1A but max 5W.
5W : 22.2V = 225mA Takes too long
However, if you like me don`t have 120V bulbs but 12V bulbs (12W) this could be done with 3s2p?

Just a little tip: I found an quick and easy way of connecting cells together with some leftovers tabs. Stick it under the tab on each cell pair. Just be careful with the positive end cause this is a place when you can have a short very easily. Don`t touch the negative rim of the case. I used some duck tape on the positive end. You can also use gasket.

20140413_195504.jpg
 
Back
Top