Lithiums - mAh/100th V - Discharge Tests

[DrkAngel]

Apparently the cells, that I believed to be LiFePO4, are in actuality -
lithium iron magnesium phosphate LiFeMgPO4
I suspected something to be odd, when my iMax, set to LiFe - 3.6V charge, only charged the cell to 3.4V ...
Yes! I did test additional cells, from multiple packs, all similar!
Further inspection found "Valence Technology" on the battery's "host device".
Investigation suggested - LiFeMgPO4

Anyhow, this discharge curve is extremely different, from anything else, that I have charted!
Charge settled to 3.36V, overnight.

The "host device" case, lists voltage at 12.8V, (4 cells = 3.2V per cell), which is, right in the center of the mah peak.
The cells seem specifically designed to provide an extremely stable 12V power supply.

Lets add a "LiPo", for comparison!

.

Odd duck ... but presents a multitude of possibilities!

 

[Toorbough ULL-Zeveigh]

Drkangel- Do you just f- with us, or are you also in the race competeting to be the dumbest human on earth. You and Morph both, I can't help but think you guys are just F-ing with us.

it's all there in his sig, especially the last line.

A little learning is a dangerous thing;
Drink deep, or taste not the Pierian spring:
There shallow draughts intoxicate the brain,
And drinking largely, sobers us again.

I enjoy enlightening, and enlivening the spirit of the innovators.

he enjoys being deliberately provocative, i.e. stirring the shit.
just need to be aware that you're flagellating with a troll.

and as you can read for yourself he also likes to revise history after the fact.
for clarity it would help to embed his posts within a quote.

 

[DrkAngel]

Drkangel- Do you just f- with us, or are you also in the race competeting to be the dumbest human on earth. You and Morph both, I can't help but think you guys are just F-ing with us.

it's all there in his sig, especially the last line.

I enjoy enlightening, and enlivening the spirit of the innovators.

he enjoys being deliberately provocative, i.e. stirring the shit.

I love planting a seed and watching it grow in a fertile mind.

Sadly, the hardheaded seem to take no benefit.
The seed falls into a hard dry mind, seems to dry and wither leaving a bitter ... irritation?

Many seem to build their minds with "bricks of knowledge", strong inflexible walls of rigid ideas and beliefs.
This tends to create something similar to a "mental constipation".
It's difficult for any new-better ideas to make an impression.
And it is rare that anything better than putrid scoffings get out.

I, on the other hand, have a flexible mind.
Constantly stretching it with new ideas and projects, flexing my brain power.
Taking to heart that "certainty is a rare and precious thing", something to be admired ... not molded into dull bricks and stacked away in their dusty attics.

 

[DrkAngel]

Apparently the cells, that I believed to be LiFePO4, are in actuality -
lithium iron magnesium phosphate LiFeMgPO4

Odd duck ... but presents a multitude of possibilities!

Could replace the, less than, 6Ah usable, SLA, in the EZip pack, with 15Ah.
Almost 3X the capacity.
2500 cycles = 8 years - daily cycled.
Should outilve a few ebikes!

Enough cells for 5 EZip packs! But ... gotta think of some better use ...

Solar storage would be ideal, if you can find cells, cheap enough, in quantity.
Might last decades!

Valence Technology does not specifically rate these cells.
However, their current production of their LiFeMgPO4 cells, advertise:
2C continuous output.
3C surge output.
12.8V - 3.2V per cell
13.8V - 3.45V per cell float voltage
14.6V - 3.65 charge voltage
.5C - standard charge

Charging voltages look compatible with SLA chargers.
Charge rate is much higher than SLA recommended - a good thing.

The Valence LiFeMgPO4 technology - cells are Patent - branded Saphion.

The company - Valence Technology

OMG! They are a US based company!


"Valence manufacturers it’s cathode material and packs in two wholly owned subsidiaries Valence Energy Tech in Suzhou, China." - Oh! ... figures ...

Read more: http://www.answers.com/topic/valence-technology#ixzz1dmRAdeya

 

[DrkAngel]

I do have a few kWh of Sony cells.
Most impressive is that they still display 5 led, (on pack. power level, meter), after 6 months, from initial charge.
3.6V- 4.1V looks to be recommended "usable".

Compared to the "unbranded" cells, from same model pack.
3.7V ... maybe 3.6V, - 4.1V recommended "usable".

Appropriately, both types of pack, are charged to only 4.10V by the, oem, bulk charger & notebook.
Tested at 4.25V, produced no noticeable additional capacity!

Larger versions "attached" below -

 

[DrkAngel]

My recycled 4320 Dbl. recycled cells.
The "unbranded" cells have a wider, "beefier", discharge curve.
(Yes, you may be hearing "Where's the beef?", from me ... later.)
The Sony cells have a much more "even" curve, with a better "bottom end".


I have 2 kWh of the Sony cells tagged for my "44 Magnum" project.
8kWh of the "unbranded", set aside for my eMotorcycle, just got a genuine ETEK, as motovation.

 

[DrkAngel]

All for now!
I ran out of battery types, for testing.
Would like to add LiFePO4, but haven't acquired any specimens.

Could try SLA ... I suppose ...

Oops, just remembered I have some Nimh.

I would like to invite others to use this method and post their results.
I would recommend the following standards:

List the Brand, mAh capacity, formulation - LiPo, LiCo, LiMn, NiMh, SLA etc.
Use a regulated discharge (digitally metered) I used an iMax B6 - about $25
Discharge at a low rate to reduce "voltage sag"(.2C = 5 hours)
Digitally meter voltage at the cell - much more accurate, than through the discharge device
Top and bottom voltages are the most volatile - need the closest monitoring.

A faster discharge, .5C?, should work OK, might even demonstrate a better "actual use" capacity-output? (2 hour test)

Graph program is a free download from http://www.padowan.dk/graph/.

 

[DrkAngel]

Warning! - Capacity maps vary greatly by formulation.

2 examples - compared ...

Yellow - I interpret LiPo 4320mAh mapped cell as being usable from 4.1V to 3.7V, with, to 3.6V being an emergency, "limp home".
Judging, by area(=mAh) of, my rated, usable:
4.1V = 100%
3.98V = 75%
3.9V = 50%
3.83V = 25%
3.7V = Empty - with emergency reserve

Green - On the Winforce 5700mAh rated usable would be, from 4.2V to 3.4V, but with an abrupt, empty!
There is an is a severe drop from greatest capacity to ... dead?
Extreme care recommended when fully using capacity!
4.2V = 100%
3.9V = 75%
3.75V = 50%
3.6V = 25%
3.4V = Empty - no usable reserve-safety

 

[DrkAngel]

Using a different measuring method I was able to confirm the validity of my results.
See - Determining Capacity Profile

 

[DrkAngel]

Optimal DOD (Depth Of Discharge) varies by battery formulation!
(These 3 discharge profiles are all Lithium Cobalt variations.)
Optimal discharge voltages vary between 3.5V - 3.65V - 3.75V!
Some cell types retain good energy density right to about 4.20V while another had poor energy density till below 4.10V

1st step is to determine the best high and low voltage for your battery pack.
I demonstrated a couple methods in earlier in thread.

Simplest?
1. Find a "track", circular path of moderate length. No stop signs or interruption is best.
(Around a block?, back and forth along an open road or trail?)
Charge battery pack to its maximum recommended voltage (4.2V per cell?).
Attach voltage meter to pack or to single bank. (100th V capability recommended.)
Object is - to determine the variance of voltage loss at various starting voltages, using identical amounts of power usage.
2. Run identical laps, same speed and throttle usage.
Release throttle, at the same point, several seconds before passing your start point. (should recover from voltage sag)
Call out voltage as you pass your start point, have a friend charting your voltage. (Graph paper is nice or chart into graph program to see results?)
Initial lap(s) should show a heavy drop.
Severity of drop will decrease and stabilize.
At some point the severity of drop will increase, this would be your maximum DOD.
(This method will give you a chart similar to the common discharge graph, without the drastic voltage sag.)

You should be able to determine your optimal charge voltage and discharge voltage.
You want to stay within the stable area of the graph, charge to below the sharp initial drop and do not discharge below the point where the drop again sharpens.
Charging to a voltage where little energy is retained is needlessly damaging.
Deeply discharging where minimal energy is available is needlessly damaging.

Staying within your batteries optimal range will greatly increase its lifespan and preserve its performance.

Note: LiFePO4 and similar formulations have the majority of there energy concentrated in a narrow voltage range!(Near the 3.2V region)

 

[DrkAngel]

While this method of mapping battery capacity is valuable and informative ... mapping capacity during charging looks to be a much better method of determining optimal DOD (Depth of Discharge)and HVC (High Voltage Cut-off).

See - The Test
And - Capacity mapping by charge mAh ... completed.

While my discharge capacity map seemed to indicate a 3.7V and 4.05V as optimal DOD and HVC, the charge capacity map clearly shows 3.75V and, most importantly, 4.15V as optimal.
4.05V charge limit wasted a substantial "bulge" of available capacity.

Important! - This "capacity map" is for the Lipo cells I pulled from Dell "secondary bay" battery packs only! All cells vary in formulation and technology, capacity map and optimal voltages will vary greatly.

 

[DrInnovation]

Thanks for these plots.. While I see value in the standard discharge curves, they don't answer some important questions for me, at least not easily. I'm looking at building my first pack and one of the questions I face is the charge level. Interesting to know how much I loose by charging to a lower voltage. I'm looking lifepo4, in particular headway but did not see them here did you ever acquire any?

I see a value in these type of plots in helping decide charge voltage and deciding on a BMS. Since headways are big and single cell i cannot easily balance them by hand if they need regular balanceing, so will likely go with a BMS which will also to protect them if people mess with/turn the bike on while I'm parked at school. Looking at different BMSs for lifepo4 I see many different balance voltages, some like Supower balance for 3.5 .. some go as high as 3.9 (which I won't do), SignalAB is at 3.8. Headways's BMS balances at 3.6 (their recommend setting) and I found a few others that do too. But the headway BMS is 70$ more than the supower one. so it would be interesting to know the capacity loss for 3.5 as the max. So it depends on how sharp is the peak (e.g. your valance cells are very sharp and missing their charge by .1v would make a huge difference). (For my ride capacity is not that important.. power is as I have 300+ ft vertical with 8% and 10% grades).

Anyhow I figured I'd ask DrAngle if you ever did get to lifepo4 testing.. if not maybe I'll have to generate one if go with headways and then decide which BMS.

I also to say thanks I've read many many of the thread with useful data and lively commentary. Was considering an 18650 build based on stated C specs but after reading I no realize that unless I used a123 cells, anything that could deliver my desired 30A at 24-29V would not fill well in my Lashout's existing frame-mounted battery (which holds 2 12v12AH SLAs).

 

[dnmun]

Thanks for these plots.. While I see value in the standard discharge curves, they don't answer some important questions for me, at least not easily. I'm looking at building my first pack and one of the questions I face is the charge level. Interesting to know how much I loose by charging to a lower voltage. I'm looking lifepo4, in particular headway but did not see them here did you ever acquire any?

I see a value in these type of plots in helping decide charge voltage and deciding on a BMS. Since headways are big and single cell i cannot easily balance them by hand if they need regular balanceing, so will likely go with a BMS which will also to protect them if people mess with/turn the bike on while I'm parked at school. Looking at different BMSs for lifepo4 I see many different balance voltages, some like Supower balance for 3.5 .. some go as high as 3.9 (which I won't do), SignalAB is at 3.8. Headways's BMS balances at 3.6 (their recommend setting) and I found a few others that do too. But the headway BMS is 70$ more than the supower one. so it would be interesting to know the capacity loss for 3.5 as the max. So it depends on how sharp is the peak (e.g. your valance cells are very sharp and missing their charge by .1v would make a huge difference). (For my ride capacity is not that important.. power is as I have 300+ ft vertical with 8% and 10% grades).

Anyhow I figured I'd ask DrAngle if you ever did get to lifepo4 testing.. if not maybe I'll have to generate one if go with headways and then decide which BMS.

I also to say thanks I've read many many of the thread with useful data and lively commentary. Was considering an 18650 build based on stated C specs but after reading I no realize that unless I used a123 cells, anything that could deliver my desired 30A at 24-29V would not fill well in my Lashout's existing frame-mounted battery (which holds 2 12v12AH SLAs).

asking advice here would make you even more confused. you already seem misinformed about charging lithium so you now will be deliberately misinformed.

 

[DrkAngel]

Capacity mapping LiFe is much less precise.
I did map LiFeMgPO4:

Apparently the cells, that I believed to be LiFePO4, are in actuality -
lithium iron magnesium phosphate LiFeMgPO4
I suspected something to be odd, when my iMax, set to LiFe - 3.6V charge, only charged the cell to 3.4V ...
Yes! I did test additional cells, from multiple packs, all similar!
Further inspection found "Valence Technology" on the battery's "host device".
Investigation suggested - LiFeMgPO4

Anyhow, this discharge curve is extremely different, from anything else, that I have charted!
Charge settled to 3.36V, overnight.

The "host device" case, lists voltage at 12.8V, (4 cells = 3.2V per cell), which is, right in the center of the mah peak.
The cells seem specifically designed to provide an extremely stable 12V power supply.

Lets add a "LiPo", for comparison!

.

Odd duck ... but presents a multitude of possibilities!

See - Capacity Mapping (Optimal Charge-Discharge Voltages) for updated mapping methods procedures.

 

[DrInnovation]

asking advice here would make you even more confused. you already seem misinformed about charging lithium so you now will be deliberately misinformed.

I'm curious what I've said that suggest I'm misinformed about charging lithium.. Are there 18650 cells, other than A123, that could provide 30A of power in a 100mmx100mmx300mm space?


And to those that have questioned the value of these curves I've just explain one place where I think they have value. The curves are not new information.. they can be computed from the standard discharge curves (they are the are effectively an estimate of capacity at voltage which can be obtained by integrating d(Capacity)/dV between two voltages. It can be computed as area in horizontal slices of the standard curve). But because the discharge curves is so flat and its hard to actually estimate from the curve. This experimental data makes it more explicit if one is interested in how much usable energy one can get between a pair of voltages.

 

[Hillhater]

.. Are there 18650 cells, other than A123, that could provide 30A of power in a 100mmx100mmx300mm space?

yes ..many.
That volume could hold over 100, 18650 cells.
As you are requiring 24-29 volts, that would imply a 8s string of lico cells.( 3.7v each)
So you could have 8s and 12 parallel, which would mean each cell only need supply 2.5 amps.
Almost any decent 18650 cell can do that comfortably, and with much more capacity than the A123 cells
....and cheaper !! :wink:

 

[DrInnovation]

.. Are there 18650 cells, other than A123, that could provide 30A of power in a 100mmx100mmx300mm space?

yes ..many.
That volume could hold over 100, 18650 cells.
As you are requiring 24-29 volts, that would imply a 8s string of lico cells.( 3.7v each)
So you could have 8s and 12 parallel, which would mean each cell only need supply 2.5 amps.
Almost any decent 18650 cell can do that comfortably, and with much more capacity than the A123 cells
....and cheaper !! :wink:

Thanks.. I've already tested and can only so 7s.. its a pretty strict 30v max and fully charged 8s was too high for the controller. I did consider going 8s but charging to a lower voltage (hence my looking at things like this thread), but I'd have to limit to about 3.75v to stay <30 which looses a good bit of power unless I do 8s in lifepo.
I do have a lot of 18650 cells around, but getting to large numbers will mean a build will a mixed set of brands and a good bit more time than just buying some 8 headways.. Buying a lot of new cells for 18650 seems like asking for more work with little gain.. that won't save me much over the headways which will be less work.

I've tried some packing with the cells I have and if I wanted to make it easy to run parallel strings I could get 7s8p in a regular layout flat layer -- which I did not think was enough as that would need 3.75A per cell. Rethinking it last night I realized that I left a lot of volume unused with that layout If I stack them in 5x5block (making about 100mmx70mm) and then I can get 4 such stacks in lengthwise I think I can get 7x14 (98) 18650s in the case (well actually 100. as blocks of 25) which does take it back down to a more reasonable 2.15A/cell. I'll have to see if they actually fit as that is really tight on the spacing.. It will also be moderately complex wiring as the parallel groups will not all be in the same block (which is why I did not "see" that layout before).

 

[mushymelon]

I need to check the dates on these threads before i read it all. I wan't that 20 min back.

 

[Hillhater]

I do have a lot of 18650 cells around, but getting to large numbers will mean a build will a mixed set of brands and a good bit more time than just buying some 8 headways.. Buying a lot of new cells for 18650 seems like asking for more work with little gain.. that won't save me much over the headways which will be less work. .

Sure , the 8s Headway option may be simpler but you will have a lot less capacity, most likely more voltage "sag", a heavier pack, and many $$$s worse off.
Have you read through some of the 18650 pack build threads and seen the assembly kits that are available ?
There are other options also, but lets not confuse the situation more.

 

[DrInnovation]

I do have a lot of 18650 cells around, but getting to large numbers will mean a build will a mixed set of brands and a good bit more time than just buying some 8 headways.. Buying a lot of new cells for 18650 seems like asking for more work with little gain.. that won't save me much over the headways which will be less work. .

Sure , the 8s Headway option may be simpler but you will have a lot less capacity, most likely more voltage "sag", a heavier pack, and many $$$s worse off.
Have you read through some of the 18650 pack build threads and seen the assembly kits that are available ?
There are other options also, but lets not confuse the situation more.

Thanks..Capacity is far less importance to me than the ability to provide 30A current as I have +250ft to work and more if I take some side trips first, and I don't want to show up sweaty.
Its all up hill with only a few flat spots and some rather steep (8% and 10% grade) sections so little time for me, or the battery, to recover. I don't have a watt meter or cell logger (yet), but know I'm making out my poor little 600w motor. While the headway would have less capacity from the reading I've done so far, they should be able to provide the 30A .. did I miss some threads showing excessive sag? I do remember one but in the end that was a loose nut.. If you think they will sag a lot I'll keep digging about them.

I have read through lots of 18650 pack build threads and the homepack threads, the no-solder thread and a few kit threads including one kit for sale (http://endless-sphere.com/forums/viewtopic.php?f=31&t=62344) but that not in a configuration that would fit in my existing bike either voltage wise or space wise. Are there other kits I missed for sale?


But now that I think I might be able to pack in the 7s14p, then I'd be at 21Ah, if the 2200mah cells in capacity and then my 30A is only 1.5C so within reason. The 18650 build is also interesting to me because the sustainability people on campus though if I do it and then run a DIY seminar there would be a good bit of student interest. Might even get them to spring for parts for a tab welder and maybe some testing/charging hardware. I'll need to pull open the existing battery pack and see if I can get 5x5 in there.. if I can only get 64 cells it won't have enough juice.

Today since my SLAs no longer get me to work and even the last 50'ft vertical on a bike with dead SLA is painful I started thinking about a short-term hybrid until I can collect enough cells/time for the full build... Long term I want to use the bikes existing battery housing, but for a short term boost I could add a small soft pack. Yup I've been reading those threads too. Any comments on that idea?

 

[Hillhater]

Find the discharge test data/graph for those headways at 30amps and see what the volt drop is.
Compare to similar situation for your 18650s at whatever load you calculate they will see (~2.5A ?)
Volt drop/sag under load can be a real spoiler in use.

 

[DrInnovation]

Find the discharge test data/graph for those headways at 30amps and see what the volt drop is.
Compare to similar situation for your 18650s at whatever load you calculate they will see (~2.5A ?)
Volt drop/sag under load can be a real spoiler in use.

Did that already, question is if I should believe the charts/specs.

At 30A, the 10ah headways will sag to about 3.1v.. and will stay above 2.75 (22v pack) until about 8ah (so 80% capacity). 30A (3C) is the max sustained discharge (100A/10C is max)
The 8ah HP headways are rated for 10C continus (25 peak) and would make it out to 7ah even at 5C...

For the 18650 anasonic CGR18650DA (my most common) datasheets lists 4900 (2C) ass the "max" for these cells (unclear if its burst bax or sustained. If I can get the 7s14p, I'd estimates at 2.5C it will sag a lot but will stay above 3.15 (22 pack voltage) until about 1700mah (about 70% of capacity) but could not find a definitive chart.

Both of those are for "new".. at ideal temp. I expect pushing the 18650 cells that hard will shorten their life.

 

[Hillhater]

With 14P of those 18650's you wont need 2.5C ,..more like 1C would give you the 30A
At 1C , most of the Panasonic cells hold over 3.4v until they are mostly empty.
http://www.dampfakkus.de/liste_akkus.php
From this test data, it looks like the Headways will struggle to hold 3v for discharge at 3C ..( interpreting between the 2 and 4C traces.)

 

[Matador]

Is there any available graph like that for the Sony US18650 VTC4 ??
Can't seem to find it even though the VTC4 is a popular one I think.

I tried here and google but no luck

Looking for a graph similar to this ; with d(mAh) in function of Voltage.
I'm trying to find the best compromise between capacity and long cycle life... finding where most of the capacity is located in terms of voltage range.



Matador.

 

[DrkAngel]

I never mapped that particular cell.
See - Capacity Mapping for methods I used.