bad Dewalt QC or what is normal and what isn't

[curious]

I've finally started working on battery packs (2x 6P A123) for my bike. I've bought only *factory sealed* packs from ebay - no signs of tampering was evident on the package. Today I've opened first 7 Dewalt packs. Out of these one pack had two (!) failed cells (0.45V) so I discarded it (remaining cells may find some use in RC), another pack had a cell at 2.5V and two more packs had one cell each at 3.2V. The rest of the cells were at 3.3V +-0.005V. Considering a very tight distribution of voltage levels on healthy cells should I be worried about these variations ? I did balance all three cells including one at 2.5V using CVCC PSU and they seem to take charge normally (slow and steady). I am going to wire 6P pack in parallel for each cell so I needed balancing to avoid sparks soldering cells together. But I am sort of worried about (i) strange outliers (it is not a normal distribution of voltages) (ii) remaining unopened packs (no time to do that today). What is your experience with these ?

 

[steveo]

If i were you; i would of brought them to a service center and get replacements as per they are still have warranty .. unless you butchered the shell to get to the cells..

even cells found below 2v could be revived by parallel another a123 cell to bring the voltage up...

sounds like your using a psu of some sort & some balancers to straighten everything out.

what i would do is charge all the packs; balance everything.. don't take the packs apart .. just mark the low cells you found ..
and then do a load test to see the results of all cells then go from there ..

if you have any spare packs that are untouched send this way ..

-steveo

 

[Doctorbass]

By my experience with problem regarding dewalt pack and bad A123 cells i can say that normally, it's often cell no 1, 2, 4 and 6 taht have problem... other may have too but the cells i mentioned are the most common low cell that i measure with voltage from 2.3 to 0.000000V

Hazard? I dont know, the placement of these cells in the plastic case seems to be ok in all condition including thermal management.

maybe it's the BMS circuit.. but... each cell voltage are buffered at the input of the 2 multipin connectors..

Usually cell that have low voltage may have been reversed and stored in defect state too much time.. most of them have 0.0000 volts... the "LESS" bad cells higher than .5V may be recovered. my experience shown me that used/deffect/long stored time cells will recover to around 1600 to 2000mAh.

The cells that have been stored for a short period can get from 2000 to 2200mAh. Only the good one will reach higher capacity than that.

Also, if you try to recover a bad cell, CHARGE THEM AT MAX 0.2C and best 0.1C to avoid them to heat and be damadged.. they are sensitive at low voltage if you apply too much current!!

Good luck!

Doc .. that have seen dozens of bad cells on dozens of dewalt pack.... ALL THAT COME FROM THE SERVICE CENTER FOR SURE!.. not brand new one!

 

[curious]

My plan for now is to use all packs that have cells above 2V. I definitely will not use the pack with a cell stored at 0.45V. It may be revived for a while but it will eventually kill 5 other cells wired in parallel and reduce the pack lifetime by a great deal. My experience with lipo cells is that reviving it from below 2V is a temporary measure - it will die much sooner than a healthy cell. But I am totally new to a123 chemistry so my extrapolations may be wrong.

 

[curious]

BTW being new from the pack I do not think the 0.45V cells were reversed, they are probably either defective from the factory or stored in wrong conditions for too long. I am not as much worried of these cells being a hazard (like lipo) as i am concerned about killing other cells in parallel to the suspicious one with time.

 

[Doctorbass]

With lower cell, you should absolutly match the capacity of each 1p pack. that mean lower cell need to have their capacity to be measured and matched correctly. A cheap RC charger can do that.

 

[curious]

I have a wattmeter, astro charger and CVCC PSU so I can do that no prob. But what would you suggest if the cell has a lower capacity of say 1800mah ? And other 5 in parallel are assumed to have normal 2200-2300mah capacity. Replace the cell from the middle of the pack? It is possible but requires cutting the welded tabs etc.
Mind you I am not talking about cells that were at 0.45V, these will not get into the bike pack (I'll try to revive them but only for RC use), I am more worried about one at 2.5V and less so about those at 3.2V.

 

[Doctorbass]

I have a wattmeter, astro charger and CVCC PSU so I can do that no prob. But what would you suggest if the cell has a lower capacity of say 1800mah ? And other 5 in parallel are assumed to have normal 2200-2300mah capacity. Replace the cell from the middle of the pack? It is possible but requires cutting the welded tabs etc.
Mind you I am not talking about cells that were at 0.45V, these will not get into the bike pack (I'll try to revive them but only for RC use), I am more worried about one at 2.5V and less so about those at 3.2V.

What you need to do is to arrange your cell to have the same total capacity per parallel pack.

ex: with a 1s 6p pack, cell with: 2200 2250 2300 2300 2100 2250 will give you 13400mAh

If the next pack have 2300 2050 2250 2250 2300 2250 =13400mAh that will be very similar and should be ok to match in serie with the previous pack.

so if you have a lower cell in a parallel pack, something like 1900 to 2100mAh, that will be ok, but you'll need that this pack have the same capacity then the other in serie.

Dont forget!: ALWAYS measure cell in this conditions:

-in stable room temperature
-wait few hour after the last charge before to do the capacity test to let them to stabilyze
-Do a capacity test at around 0.2 to 1C (that will give you a better accuracy)
-Ensure that the cell have been fully charged few hours ago.

Doc

 

[curious]

Thanks for clarification. I agree with you 100% that this is a proper way to build the pack given you completely disassemble it into individual cells. I am trying to go less labor-intensive way and use 10s blocks as is (factory welded) except soldering buses to connect cells in parallel across 6 packs. I'll measure the more suspicious 2.5V cell in the evening and if it is over 2000mah I will use the 10s pack as is. I understand the implications of having non-equal capacity 1s6p units but breaking into individual cells and especially measuring each cell individual capacity will take much more time. Besides due to central limit theorem the variance of the capacity distribution of a N x p subblock should go to zero as N increases to infinity .

 

[Ypedal]

Hmm.. this is something i put alot of thought into a while back.

Paralell blocks of cells make charging and ballancing easier, but what do you do when a single cell fails or has a problem.. diagnosing problems can be difficult in those situations.

With series strings the voltage of each cell can be verified while the pack is assembled..

In a perfect world where cells that test " Good " and remain that way parallel groups work well.. but unfortunately this is not the case.

 

[curious]

I think continuously managing packs at the individual cell level is outside practical limits. In a more reasonable parallel setup a catastrophic cell failure (either short or very fast self-discharge) will indeed kill the parallel group. However capacity loss (even nearly complete) can be managed by adding cells in parallel to the lowest capacity parallel sub-blocks when the block becomes a clear outlier (let's say block capacity deviation from average becomes > 0.5 individual cell capacity). This requires a small maintenance pack that can be hooked up via balancing leads if they are beefy enough. And subblock capacity can be measured periodically. IMHO it is a bit of an overkill for what we are doing (or may be not).

 

[Doctorbass]

Parallel cell will never have a cell that fail itself. The voltage of the rest of each cells would hold the voltage to avoid a weak cell to short or go lower. In parallel block it is impossible that a cell become with lower voltage than other.. THEY ARE PARALLEL and the voltage of each cell follow the SOC.. what it will happen if a cell have 2000mAh and the rest of the paralle cell have 2300 in the same parallel pack is that when this pack will reach a discharge of 2000mAh, the rest of the parallel cells will continu to give current until the pack go to 2300... AND the 2000mAh cell will no more give current.. but will not drop voltage.


That is the right way to understant this situation.

Doc

 

[kbarrett]

Parallel cell will never have a cell that fail itself. The voltage of the rest of each cells would hold the voltage to avoid a weak cell to short or go lower.

Exactly.

Dewalt's decision to put 10 single cells in series to get 33V almost guaranteed individual cell failure. Add a wonky BMS and watch the battery packs get returned continually under warranty ...

If you use A123s, you really do need to break up the DeWalt packs.

 

[Ypedal]

Well.. that's the thing.... is voltage Always an indication of state of charge ?

I'm totally not an engineer, so my logic is most likely flawed .. so i'm asking the question to better understand..

Consider the discharge curve of a single cell.. 3.3v to 2.7v aprox.

When 2 cells of different voltage are connected in parallel, you get an equilization of voltage. that i understand.. but will the amount of energy contained in each cell equalize perfectly by voltage. ( Does internal resistance become a factor in this ? )

I"m likely not explaining this correctly, but i remember a post on Power-Assist from bocabikeguy that explained this in a way that having serial strings of cells vs groups of parallel cells in series was oposite of this discussion.

 

[curious]

arallel cell will never have a cell that fail itself. The voltage of the rest of each cells would hold the voltage to avoid a weak cell to short or go lower.

I disagree. Imagine a cell that has abnormally high self discharge (due to some chemistry degradation or manufacturing defect). It will present a resistive load to entire parallel block and will discharge other cells and therefore kill them if you do not constantly maintain the charge on that block.
BTW I do not argue the fact that voltage will be the same on all parallel cells

 

[magudaman]

I think it is rare to get a high self discharging cell and have not heard of this happening in a123's chemistry. I think the safest way to run lithium chemistry is in parallel. This ensures never go below set limit. Instead of needing to monitor 50 individual cells you can monitor 5. Paralleling cells insures that no two cells will ever go below set limits unless the whole strand down. When paralleling you don't have to really worry about cell Milliamp hours because the cell around it will help support it. This is of course unless you need all the current from all cells.

 

[Doctorbass]

arallel cell will never have a cell that fail itself. The voltage of the rest of each cells would hold the voltage to avoid a weak cell to short or go lower.

I disagree. Imagine a cell that has abnormally high self discharge (due to some chemistry degradation or manufacturing defect). It will present a resistive load to entire parallel block and will discharge other cells and therefore kill them if you do not constantly maintain the charge on that block.
BTW I do not argue the fact that voltage will be the same on all parallel cells

Refering to the Battery university site ( THE REFERENCE ),

In the middle of this page: http://batteryuniversity.com/partone-24.htm

A weak cell will not affect the voltage but provide a low runtime. A shorted cell could cause excessive heat and create a fire hazard.----------An electrical short would be more devastating because the faulty cell would drain the energy from the other cells, causing a fire hazard

In the last phrase they say : A SHORTED CELL... not a lower capacity cell!

I've built a boost pack of 4s4p with all my lower capacity cell (1600 to 2000mAh) cell and i matched the pack to have the same total capacity in each paralle group... and in each group, some cell gave 2000 and 1600mAh... and it still work fine after hundreds of cycles...

I agree that a shorted cell or a cell that show 0.2V would definitively not allowable in a good pack. But I would for a cell with 20% less capacity.


Every battery anthousiast should read the entire battery university lesson... they are my guideline! and are made of many years of experimented people!

Doc

 

[Doctorbass]

Well.. that's the thing.... is voltage Always an indication of state of charge ?

I'm totally not an engineer, so my logic is most likely flawed .. so i'm asking the question to better understand..

Consider the discharge curve of a single cell.. 3.3v to 2.7v aprox.

When 2 cells of different voltage are connected in parallel, you get an equilization of voltage. that i understand.. but will the amount of energy contained in each cell equalize perfectly by voltage. ( Does internal resistance become a factor in this ? )

I"m likely not explaining this correctly, but i remember a post on Power-Assist from bocabikeguy that explained this in a way that having serial strings of cells vs groups of parallel cells in series was oposite of this discussion.

Ypedal, just be carefull when paralleling cells, they absolutly need to have less than 100-200mV MAX difference. If they don't, the lower cell will recharge to higher than 10A and will lower its cycle life!

Always manually equalize cells before to paralle them! Or use a resistor to limit the current (like 1ohm 10W or a 12V bulb) when wiring them in parallel and wait until they have less than 100mV before to solter them together

If they are close to 3.2V, use the 100mV rules
If they are close to 3.0V, use the 200mV rules
If they are close to 2.8V, use the 50mV rules


The internal resistance is a bit more complicated to measure correctly dur to sensitive temperarure and SOC factor.
BUT it is always better to match the string of serie cell with a very close internal resistance difference between the string.

ex: a 10s 1p pack should have cells that have a very close internal resistor

a 10s 4p pack will need that the equivalent internal resistor of each 1s4p pack to be very close.

to calculate the equivalent internal resistor of a paralle pack, it's the same as normal resistor formula.

here is the formula and a calculator: :wink: http://www.1728.com/resistrs.htm

enjoy!

Doc


Doc

 

[Doctorbass]

Ypedal, it's my turn! [attachment=0]P1030685_800x600.jpg[/attachment]

Doc

 

[curious]

Mmm... have to grab something to eat. Anyway the 2.5V cell tested at 2100mah give or take 50mah because I used lipo astro charger and manually tapered the current as the cell approached 3.6V. So I give it a go and started wiring the pack. Perhaps I should start a build thread in ebike general section.

 

[curious]

Ok, I've opened the rest of Dewalt packs - I've got another bad pack with a cell at 0.6V. I start to wonder if there is some sort of factory sealed B-stock or they were stored for too long. This is a bit frustrating, I need to order two more packs (or one if I mess with the welded tabs and try to replace a single cell).

 

[Dee Jay]

I've always been leery of Ebay and all this auction ballyhoo. Factory sealed B-stock are probably what they are and these sellers are operating under the excuse, "They're sealed, so it can't possibly be MY fault! It's on you buddy boy, grow up and just suck-it-up "~ Beware ~

Then you got these shady duct tape jobs. Like drug dealers, they'll sell a few good packs at a good price to drum up some hype, then sell bad packs and fly-by-night. cya suckuz! ~ Do Not Feed the ANIMALS ~

just my two ohms . . . :|

 

[curious]

Well, I stopped short of blaming the sellers. I am just guessing here and even if it is true I am not sure the sellers are aware. Also I had good experience buying from ebay in general (and I bough quite a bit).

Also there is no realistic alternative to getting a123 cells from ebay in dewalt packs. I am closely watching LiFeBatt thread and will probably use these in the future if I need big packs.

 

[Doctorbass]

(or one if I mess with the welded tabs and try to replace a single cell).

Easy :wink: .. just do not heat too much time the tab.

Cut the tab link with a dremel or some good cutter, replace the cell qith the good one and solder the tab half cutted link together.. you may use some cooper wire to renforce the contact. Nothing more simple!

Doc

 

[Doctorbass]

From my side i offer at 1/3 price slightly used pack and cells that YOU KNOW THAT 2 or rarely 3 are low... i offer to test those too so you are ready to match your cells 8)

Dewalt get alot more popular with their nano series these months (2008' +). I get more than i need.... ... did you heard correctly?... ....

info in my post in "selling surplus" section and PM me if you guys need some!!

Doc