Homemade Battery Packs

xmasta said:

Hi all,

5.2 would make them cells to be 2600mah, but 4.53 comes up with some funky 2265mah .. ?

I'm confuse! Not all of them are marked in this confusing way now i wonder, do i have a lot of 2200 in my hand or 2600 ..

SE US18650GR is written on them, thats the only part that gives results in google, suggesting sony 2200? (second line is "T 8A1230I24T")

thanks in advance for your input, also if anyone knows, are those cells good? in the world of 18650 cells kindof.

Click to expand...

Where did you get 4.53?

Even if the batteries are from the same brand the cells inside may be different. The number written on the outside of the pack should match che capacity of the cells inside the pack.

8)

spuzzete said:

Where did you get 4.53?

Even if the batteries are from the same brand the cells inside may be different. The number written on the outside of the pack should match che capacity of the cells inside the pack.

8)

Click to expand...

View attachment 1



three different type laptop batts for different laptops, all having double capacity numbers.

I think this should go out as a warning, because some ppl may not even notice the second value, I didnt at first.
So they may label their cells wrong and end up with very uneven and bad pack.
So I think good advice would be to check your lappy packs for two values and google your cells when markings are available to be sure.

Likely ...
The one "rating" is the rated capacity of the cells.
The second rating is packs designed "rating".

The difference being that the pack manufacturer restricts maximum and minimum voltage ...
So as to prolong durability and lifespan!

4.1V per cell might provide double the life cycles compared to 4.2V ...
at a 5% reduction(?) in output!
Another 5% by stopping at 3.6V, instead of draining below 3.5V.

Near "full", and "empty", Li-ion cells store a very low percentage of energy.

If 3.6V - 4.1V = 100% recommended capacity (20% capacity per 1/10th V - averaged) ...
Discharging-charging 3.5V - 4.2V (110% recommended capacity, 5% per added 1/10th V)...
Hardly seems worth it ... for the wear and tear damage it causes!
*Based on measured capacity of the LiCo lipo I've been using lately.

25.9V 25.92Ah pack, built by recycling laptop Lipo cells is nearing retirement.

After 1 1/2 years,
2 Summers,
about 4500 miles,
hundreds of recharges,
I'm about ready to retire my eZip Lipo rebuild.

Capacity has reduced noticeably!
Understandable.
Especially considering that all the cells were pulled from defective packs, the ones that would not charge properly!
Not that it is worthless!
It still gives me more than double the range of the oem SLA pack.

I think it is time that I re-build using my "good" cells.
Also I plan on adhering to a reduced charge-discharge voltage range.
These cells demonstrated a +90% capacity between 3.6V and 4.1V.
I'm certain that my initial 4.2V charging regime had a needlessly damaging affect.

Will try to add, and document, improvements.

DrkAngel said:

I think it is time that I re-build using my "good" cells.
Also I plan on adhering to a reduced charge-discharge voltage range.
These cells demonstrated a +90% capacity between 3.6V and 4.1V.
I'm certain that my initial 4.2V charging regime had a needlessly damaging affect.

Click to expand...

OEM, laptop & external chargers, only charged to 4.1V.


Discharge map constructed using - MAh per 100thV discharge test

I am tempted towards an even tighter voltage range!
3.7V - 4.05V takes advantage of the "beefiest" area of the discharge profile.
It also leaves a healthy "emergency reserve", for "limping" home.


Be aware! These results are for a specific cell type-formulation and might have no relation to other brands-types-formulations!

Also! I just posted my MeanWell latest success - Meanwell S-150-24 in Series - 37V 7A Li-ion charger.

Using some A123 1100mAh 18650 cells. It's a 4s7p configuration to replace my lead battery. Works great.




This is the battery while the motorcycle is at idle.

I finally acquired a 30V capable voltmeter with 100ths V capability.
Combined with a cheap Ah meter, I was able to measure AHs on my recycled Lipo pack build. (Lithium ion (cobalt))
7S, 25.9V, 25.92 (rated) Ah
18.4Ah - (25.9V-29.4V) = 3.7-4.2V = 70% of rated

I decided to determine additional capacity added by various charge voltages.
1st - charged pack to 28.7V = 4.1V per cell
2nd - charged pack to 29.05V = 4.15V per cell (+1.06 Ah) 5.76% of full charge
3rd - charged pack to 29.4V = 4.2V per cell (+.79 Ah) 4.3% of full charge
(4.1V to 4.2V = 1.85Ah, 10.16% capacity. )
Note - Ah recorded when charge rate diminished to .01C = .15A

Charging to 4.1V, instead of 4.2V, loses almost exactly 10% of full capacity.
Worth it for an estimated doubling + of life cycles?

I know that my pack has deteriorated over the past few hundred cycles, but I was surprised to see that it still provided almost 300% the energy of the oem eZip pack (SLA).
(SLA - 10Ah discharged in 1 hour outputs a pitiful 6.14 Ah)

Rebuild will take advantage of a 4.1V maximum charge.
I also will test capacity and condition of individual cells ... stringently.
My confidence in "bulk charging" has grown to the degree that I might forgo the use of an external balance connector.
The 4.1V average charge leaves a large safety margin!
However, a schedule of internal checks will be maintained!

I guess ... now, I must measure the Ah between 3.5V - 3.6V - 3.7V.
First I must recheck low voltage levels and possibly "bottom balance"...
Due to it's deterioration, I have been limiting discharge to an averaged 3.7V.

DrkAngel said:

I finally acquired a 30V capable voltmeter with 100ths V capability.
Combined with a cheap Ah meter, I was able to measure AHs on my recycled Lipo pack build. (Lithium ion (cobalt))
7S, 25.9V, 25.92 (rated) Ah
18.4Ah - (25.9V-29.4V) = 3.7-4.2V = 70% of rated

I decided to determine additional capacity added by various charge voltages.
1st - charged pack to 28.7V = 4.1V per cell
2nd - charged pack to 29.05V = 4.15V per cell (+1.06 Ah) 5.76% of full charge
3rd - charged pack to 29.4V = 4.2V per cell (+.79 Ah) 4.3% of full charge
(4.1V to 4.2V = 1.85Ah, 10.16% capacity. )
Note - Ah recorded when charge rate diminished to .01C = .15A
...

Click to expand...

I thought it might be illuminating to gauge the capacity of my lithium-ion (Cobalt) Lipo pack.
Method -
Begin with cells at 3.5V (my recommended minimum voltage)
Precisely charge to 3.6V - measure Ah needed
Repeat at .1V intervals until 4.2V (my recommended maximum voltage)

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

I was both disturbed and inspired at the lack of capacity below 3.7V

With these cells ... discharging below 3.7V is needlessly damaging.
Almost 90% of the usable Ahs is between 3.7V and 4.1V.

Of course, more precise measurement should be made near 3.7V & 4.1V!

Remember! Every formula of battery has a different discharge profile.
Other lipo have displayed excellent capacity even to 3.4V.

PS Re-balancing-equalizing, with cells at 4.1V

Better chart?
24.50V (3.50V)
charge .27Ah =
25.20V (3.60V) -
charge .53Ah =
25.90V (3.70V) -
charge 3.87Ah =
26.60V (3.80V) -
charge 3.15Ah =
27.30V (3.90V) -
charge 4.60Ah =
28.00V (4.00V) -
charge 4.10Ah =
28.70V (4.10V) -
charge 1.85Ah =
29.40V (4.20V) - f
----------------------------------------
Total. .................... 18.37Ah

DrkAngel said:

DrkAngel said:

I thought it might be illuminating to gauge the capacity of my lithium-ion (Cobalt) Lipo pack....

Click to expand...

Click to expand...

Thanks for that! Question: for ah capacity, you find that during charge and/or discharge and how exactly?? Also, has anyone does this kind of reference capacity test for an A123 AMP20 li-ion? Regards.

arkmundi said:

DrkAngel said:

DrkAngel said:

I thought it might be illuminating to gauge the capacity of my lithium-ion (Cobalt) Lipo pack....

Click to expand...

Click to expand...

Thanks for that! Question: for ah capacity, you find that during charge and/or discharge and how exactly?? .

Click to expand...

Measuring Ahs charging is much easier and more precise.
Measuring Ahs during discharge might be more useful ... easier to apply towards usage pattern.

Li-ions (cobalt type specifically) charge at a near 100% efficiency, (requires moderate C charging rate, <1C)

Required:
Ah meter, or watt meter
Voltage meter 100ths capable

Full to empty
Fully charge battery
Attach Ah meter
Run till empty
Ah displayed on meter

Empty to full
Empty-discharge battery to minimum usable
Attach Ah meter
Charge fully
Ah displayed on meter

I, further, determined capacity (Ah) stored at each 1/10th V, so as to determine my optimal "Full" and "Empty" voltages.

Better chart? - Simplified SOC map
24.50V (3.50V)
charge .27Ah =
25.20V (3.60V) -
charge .53Ah =
25.90V (3.70V) -
charge 3.87Ah =
26.60V (3.80V) -
charge 3.15Ah =
27.30V (3.90V) -
charge 4.60Ah =
28.00V (4.00V) -
charge 4.10Ah =
28.70V (4.10V) -
charge 1.85Ah =
29.40V (4.20V) -
-------------------
Total.... 18.37Ah

Clearly shows minimum and maximum "effective" charges as 3.7V - 4.1V
Recycled Li-ion (Cobalt) Lipo

All types-formulation of battery are different!

I previously determined SOC at various voltages using a mAh per 100th V method.

As demonstrated by this graph, different Li-ion formulations demonstrate wide variances in optimal voltages.

The yellow line is the recycled Lipo, tested above.



This graph (Yellow line - Recycled laptop Lipo) confirms the validity of a 3.7 - 4.1V optimal voltage "map". (.4V range.)

The green line (Winforce 5700mAh RC Lipo) has a much wider range.
Optimal discharge-charge range appears to be 3.5V - 4.2V. (.7V range.)
Due to the large energy capacity concentrated at low voltage and the steep "drop off", I would highly recommend "bottom balancing" at ~3.5V? To safely take best advantage of available capacity.

Necessary for Li-ion capacity testing and specifically "capacity mapping", is an Ah meter or watt meter.
I found a very versatile item ... for less than $15!
Click on Picture


Features
Dual display - assignable functions
Volts 3.8V - 60V
Amps - 20 max
Ah
Watts
Hours

No instructions for center button "mystery" functions

I also needed a 4 digit display 100th V capable - $3.88
Range: DC 4.50-30.00V ( Can Measure Voltage Directly )
Click on Picture

DrkAngel, thanks for the clear explaination - most informative. I'm going to give it a go on my 12S A123 AMP20 pack. Question - you mention discharging till "empty" but I've heard taking a cell to zero is really bad for the longevity of the cell????

arkmundi said:

DrkAngel, thanks for the clear explaination - most informative. I'm going to give it a go on my 12S A123 AMP20 pack. Question - you mention discharging till "empty" but I've heard taking a cell to zero is really bad for the longevity of the cell????

Click to expand...

"Empty" as in minimum usable!
With Li-ion, 4.2V max - per cell, lowest usable, that I have seen, is 3.4V.
I would recommend 3.5V as minimum voltage, for test.
If 3.5V to 3.6V demonstrates substantial capacity then retest at 3.4V ... etc.

A123 cells, if LiFePO4, have a nominal voltage of 3.2V.
The minimum resting voltage might be above 3.00V
Minimum voltage safe discharge is, one source quoted at, 2.5V, but needlessly damaging?
LiFePO4 discharge curve is much shallower than Li-ion (cobalt).
It will be easy to determine capacity - Ah, and minimum usable voltage, but LiFePO4 are energy dense in a very narrow voltage band.

Getting ready for Winter.
That means breaking out my 37V Li-ion packs for my Winter eBike.
Now, 4+ years old, they have noticeably "unbalanced" since the sparce use last "Winter".

I intend on a remanufacture of my entire Li-ion collection.
3 - 37V 20.8Ah
2 - 25.9V 31.2Ah

I used 2 variations of 2600mAh cells and will try to keep both types segregated.
Tho ... I might "re-produce" as 10.4Ah gangable modules. Considering the possibility of "tacking on" a high surge Lipo module.

I intend on rating and sorting cells based on "bleed down" from full charge.
Previous capacity test show these cells to have good energy storage capacity right to 4.20V.
I will charge all to 4.20V, then separate and let set for 24 hours.
Voltage "bleed down" seems to be an excellent indicator of cell condition-capacity!
Cells will be withdrawn from consideration at 1 day intervals, possibly 1 full week of eliminations.
Then each pair of cells will be metered while a measured drain is applied, (1C=5A-10min?), voltage sag and residual V will determine final evaluation.



I'll still have my Lipo rebuilds to get me around.

The "bad" (lesser capable) cells?
They might work just fine if charged to a lower voltage ... 4.10V?
I have a battery pack with 300w 110V AC inverter, GFCI outlets. Just pulled 34Ah of bad SLA, should be able to pack in ~100+Ah of Li-ion!
Bleed down, typically occurs to a specific lower voltage, keeping pack balanced at this lower voltage, then fully charged, just before use should work nicely! ... ?
or
A cheap 3S 800mAh balance charger might maintain satisfactorily? (<$10)

So here's my testing rig for them lappy cells, bought 25w lamp for it, thats what was written on it anyway, as for reality, well, ampermeter is showing 1.3A pull as you can see from the picture, so it's more like 15w ...
Anyway, given that most cells should be lets say 2200mAh, this 1.3A draw should apply 0.6C load and be something they should live through well enough, right?



way i do it is i write up stuff into excel file and not onto the cells, cells i just mark with a number.
initial test fails dont even get a number.
writing up hookup voltage after few moments, then 10min snap then 30min snap then 1h.
Initial sorting by 10min or 30mins.
Most weak ones die in a minute or two, if it's not dead (3.4 volt alarm) by 10mins or 30mins (if i have the time), i mark them for phase two.

xmasta said:

So here's my testing rig for them lappy cells,
... should apply 0.6C load and be something they should live through well enough, right?

Click to expand...

"Lappy" cells are designed for 2-3 hour runtime.
Which equals .33 - .5C.
Maximum, efficient, drain seems to be in the 1C range.
Above 1C there is increasingly noticeable "voltage sag" and heat production.
Heat production is a clear indicator of wasted energy - inefficient use!
I try to aim for a pack capable of surges restricted below 1C = 30Ah pack for 30A controller etc.
And continuous use below .5C = 30Ah pack for 15A cruising drain.
The bigger pack the better.

While I might test my batteries at 4.2V, I have begun restricting peak charge voltage to 4.1V.
This seems to coincide with observed oem charges as observed from multiple laptops.
Reasonably, this should help extend capacity-lifespan.
Unfortunetely, most BMS\PCB\PCM's are not reprogramable for balancing at 4.1V, leading me to rely on inexpensive adjustable bulk chargers - with a schedule of voltage monitoring.
I've been producing-using MeanWell's rigged in parallel and series.
see - MeanWell Mods S-150-24


Whenever possible, I now test cells in a 3S2P format.
That is the oem configuration for my lappy Lipo and most of my 18650 lappy cells.
Charging is fast and easy with an iMax B6 - I gang together many 3S2p and balance charge, then seperate as 3S.
After charge, I do my bleed-down elimination trials.
Next I do my "draining" using a 12V - 110V AC power inverter through the volt-A-Ah-watt meter.
Click on Picture

my LOL battery

iperov said:

my LOL battery

Click to expand...

Good job!

I like the heatshrink tubing on the cells!

Where did you get it? I would like to do the same with my next battery pack!

8)

spuzzete said:

iperov said:

my LOL battery

Click to expand...

Good job!

I like the heatshrink tubing on the cells!

Where did you get it? I would like to do the same with my next battery pack!

8)

Click to expand...

Battery Space carries several sizes.... I've bought from them before and shipping is fast and cheap.

from taobao
various sizes
http://www.yoybuy.com/en/Show/5810575693
http://www.yoybuy.com/en/Show/14126279685
http://www.yoybuy.com/en/Show/7765231831
http://www.yoybuy.com/en/Show/13638562494
http://www.yoybuy.com/en/Show/6757591623 and other
and heater http://www.yoybuy.com/en/Show/10377133583

Thanks guys for the info!

@ iperov

which size did you use for your battery?

spuzzete said:

@ iperov

which size did you use for your battery?

Click to expand...

150mm

I decided to do an "in-depth" comparison of my Li-ion replacement vs the eZip oem SLA pack.

..... 25.9V 31.2Ah (9lb 12oz) = 808Wh ..... vs ..... 24V 10Ah** (15lb 2oz) = 240Wh (120-140Wh usable)


SLA is a major heat producer - 40-50% wasted as heat!
Li-ion and lipo, at modest rates, near 100% efficient. No noticeable heat production!

SLA (10Ah) - expended in 1 hour = 6.14Ah *
24V x 10Ah = 240wh
.614C drain = 147w output to motor for 1 hour but 93wh wasted heat ~60% efficient
1C drain (10A) = 240w output to motor for 1/2 hour but 120wh wasted heat ~50% efficient

Li-ion
25.9V x 31.2Ah = 808wh (Smaller than and weighs about 1/2 the 10Ah SLA pack!)
.2C drain = 160w output to motor for 5 hours but 2-3wh wasted heat per hour ~98% efficient
.333C drain = 260w output to motor for 3 hours but 10-15wh wasted heat per hour ~95% efficient

The recycled Li-ions I use are designed for a .25-.33-.5C discharge rate. 2-4 hour runtime.
Surge output, IMIO, should be limited to 1C. 31.2Ah pack for ~30A controller.

*SLA batteries are typically rated at a 20hr discharge.
A 10Ah discharged in 1hr, typically, outputs a meager 6.14Ah.
**Pictured SLA pack, rebuilt with 11Ah "rated".