Speeding up measuring the capacity of cells .

A lengthy part of our process is measuring the AHr capacity of each cell ...It seems everybody does this , then writes the capacity to the nearest mAHr !! on the side of the cell , then puts them in bins for latter use .

We don't really need great accuracy in this, since some will divide all into 3 or 4 categories ... perhaps dumping any below 1AHr ... using 1 to1.5for one thing ....1.5 to 2 for another ... and over 2 for yet another ....

Perhaps most will just dump those below 1AHr and use those above 1 AHrin their firewall , so we don't need extreme accuracy in measuring the cell's capacity ....anyway capacitywill depend how low you discharge ...

I've been using this well knowndevice for discharge ....





I set it up with a resistor for 1A discharge

I put in a charged cell and it shows the open circuit voltage , press the left button and it starts discharge .. displaying dischargecurrent , voltage , and capacity , cycling through the three readings ..

I've noticed that if after 10 secs the voltage is below 3.65 then the cell will have a capacity at the end of the 1hr test , of below 1AHr .. a good cell will show 3.75 after 10 secs and an excellent cell 3.85 after 10 secs ...

So why go through the whole discharge cycle , and then have to charge the cell up again , when you can tell it's condition in 10 secs ??

I realise those who have been building powerwalls for some time may have spent a lot of effort , and in some cases $100's , in establishing capacity measuring set ups... they will be looking for reasons to justify what they have created....

But for beginners I think all that is needed is a 10 or 20 cell charging arrangement , using the charging modules (10 for $3) , and one discharger pictured above ...

There is great diversity among builders upon which capacity cells to put in packs , some only use above 2AHr ...others anything above 1AHr ... so this is not critical ....capacity testing really does not need to be accurate.

As I see it: it's the choice either to check the cells carefully before assembling some pack, or to be first to assemble and first to deal with the cells not noticed to be bad (heaters, low capacity, higher internal resistance).
Being lazy in repairs, I opted for the first.

Dont forget that it doesnt really take "you" any time to do full test since you dont have to sit and watch it. Ie yes it takes time but it doesnt take you time.

I switch cells on my board morning and evening. I spend like 5 min doing that.
Yes i only do 80 per day in this case but thats more than i have time to solder per day anyways

More arduous work up front, is less work in the end. If a cell goes bonkers after it's in the pack, it could be a nightmare.

Also, something to consider, is that a cell may appear to be fine up front, dropping a few tenths of a volt to start with. But, as it's being loaded down for a some time, say, 5 minutes even, it could crash well below tolerance and being in parallel with some others, could start shorting out, or start becoming heaters that eventually turn to fires.

Safety up front in my book when it comes to powerwalls.

Granted! Your idea has merit. If these cells are going to be used only as singles or pairs in some device, then that setup design testing is legit. Just not in powerwall construction.

Each idea has it's place in the world. Just not every idea is a great idea for every situation. At least you're thinking and posting here to get feedback. That's a huge plus in my book!

Capacity testing is an exercise in risk-aversion:
The more testing you do, the greater your confidence in the battery that you assemble.

It is also a compromise:
On one end of the spectrum, you could not test at all....just charge, let them rest for a few weeks, and keep everything at 4.1 volt and up.
At the other end of the spectrum, you could test every cell 25 times to get a true measure of capacity and battery behavior under load.

Most people seem to regard a single discharge test as a good compromise between the amount of work and certainty of the cell behavior.

I personally would rather put a little more work in up front, to reduce the risk of having to diagnose and repair later. The consequences of bad cell in the system and the amount of work to repair is greater to me than the inconvenience of a single capacity test for each cell.

Once put into service, a bad cell could mean removing the entire pack for repair, which could shut down your whole powerwall if you dont have replacement packs to insert during repair.

Your level of risk tolerance may be greater.
The choice is yours grasshopper.

JustJim said:

Conducting a full capacity test also helps to identify those cells that may have further issues.....Heaters

Click to expand...

How does it do that??? ... it gives the capacity number only.

Any heaters must be self shorting , forthem to get hot they are consuming electricity , and this will show up by them not holding voltage when charging is attempted..

First charge em up ...leave for a day ...any problem cells will show when put in the discharger by theopen circuit voltage being low/zero ...

ALL of thecells which hold voltage overnightcould be put in powerwalls ... even those with 100mAHrs will add capacity ... but it's up to the particular builder what he want's to use .... fuses should remove any cells which become heaters during the use of the powerwall.

As I thought lots of illogical reasons to stick with the system people already have ... my method gives an approximate figure for the capacity ... the old method give's an accurate one , but this accurate one doesn't help at all because people are simply discarding those below a chosen value , which the new system allows them to do also.

ozz93666 said:

JustJim said:

Conducting a full capacity test also helps to identify those cells that may have further issues.....Heaters

Click to expand...

How does it do that??? ... it gives the capacity number only.

Any heaters must be self shorting , forthem to get hot they are consuming electricity , and this will show up by them not holding voltage when charging is attempted..

First charge em up ...leave for a day ...any problem cells will show when put in the discharger by theopen circuit voltage being low/zero ...

Click to expand...

I have to completely disagree here. If you've watched my first video, I show a blue box containing cells that are >4V capacity; they were that way when I pulled them from the battery packs.
Now, after charging several of those cells, above 1/2 are getting hot; above 120F. I didn't start charging them right away. They had been sitting in that bin for several weeks and still had a high charge.
So, if a cell is to be shorting only when it's "not in use", then these cells should be dead. They only reared the ugly hot sided head when under a charge. And that's at low amps to boot! The TP4056 drops CC as it favors CV towards the end of the charge cycle. These cells "never" saw CC because they were already over the voltage required for the switch over.
Now granted, I haven't done a "load" test on these cell yet. But a charging cell will exhibitover heating even if it doesn't self discharge.

Korishan said:

ozz93666 said:

JustJim said:

Conducting a full capacity test also helps to identify those cells that may have further issues.....Heaters

Click to expand...

How does it do that??? ... it gives the capacity number only.

Any heaters must be self shorting , forthem to get hot they are consuming electricity , and this will show up by them not holding voltage when charging is attempted..

First charge em up ...leave for a day ...any problem cells will show when put in the discharger by theopen circuit voltage being low/zero ...

Click to expand...

I have to completely disagree here. If you've watched my first video, I show a blue box containing cells that are >4V capacity; they were that way when I pulled them from the battery packs.
Now, after charging several of those cells, above 1/2 are getting hot; above 120F. I didn't start charging them right away. They had been sitting in that bin for several weeks and still had a high charge.
So, if a cell is to be shorting only when it's "not in use", then these cells should be dead. They only reared the ugly hot sided head when under a charge. And that's at low amps to boot! The TP4056 drops CC as it favors CV towards the end of the charge cycle. These cells "never" saw CC because they were already over the voltage required for the switch over.
Now granted, I haven't done a "load" test on these cell yet. But a charging cell will exhibitover heating even if it doesn't self discharge.

Click to expand...

Can you give me a link to the video, Kor ... I'd like to see that , and perhaps comment (Am I the only one not on Youtube ???...lol... I'll have to make some videos when I've got a system worked out)

ozz93666 said:

ALL of thecells which hold voltage overnightcould be put in powerwalls ... even those with 100mAHrs will add capacity ... but it's up to the particular builder what he want's to use .... fuses should remove any cells which become heaters during the use of the powerwall.

Click to expand...

In general its "OK", but not really advised, to put lower capacity cells in parallel. But never in series. So unless you're building a 3.7 volt power wall, its really not a good idea to mix in cells of just any capacity.

The charge and discharge of a mixed battery pack will see the average of the voltage and AH. The low capacity cells will discharge first, and the high capacity cells will become overcharged because the low cap cell is keeping everyone else down.

The goal is to minimize the variation so the charge and discharge loads remain in balance.

If you want to use 100 mAh cells, the only way I would even consider it is to build an entire pack with just 100's, and add it in parallel. And to do that you need to test capacity...and if you're going to do that, I would first experiment with some small packs of maybe 6 or 8 cells. Make a pack or two with all high cap cells, and then other with 2 low cap mixed in. Do some charge & discharge test with the same load and see how much faster the mixed one dies.
Without some idea of capacity, you may be building a pack with 10 low cap cells, or 30, or none. You will never know.

BTW, a fuse will not always eliminate a bad cell. A bad cell can draw from all the others at a low rate which does nothing other than lower the capacity of the pack.

In Fact, the reason we have all these scrapped 18650's is because usually 1 or 2 cells in the pack has failed, causing the BMS to say "NOPE".

APD said:

ozz93666 said:

ALL of thecells which hold voltage overnightcould be put in powerwalls ... even those with 100mAHrs will add capacity ... but it's up to the particular builder what he want's to use .... fuses should remove any cells which become heaters during the use of the powerwall.

Click to expand...

BTW, a fuse will not always eliminate a bad cell. A bad cell can draw from all the others at a low rate which does nothing other than lower the capacity of the pack.

Click to expand...

This must be our greatest concern .. a cell in a pack which selfdischarges slowly, not enough to blow a fuse but it will empty the (3.7V)pack given enough time ...

If a cell is already bad in this way ,it cannot hold voltage ,so would be detected andwouldn't be put in a pack ... accurate capacity measurement cannot indicate which cells are about to go bad in this way ... so it seems there's nothing to be done ...

Watching the voltage of each 3.7V pack in your chain (of say7) will show the problem by a packs low voltage ... but then what can you do? If the self discharge is not enough to heat up the offending cell , and itcan't be detected by thermal imaging , there's no way to find the bad cell in the pack without checking every cell!! ...perhaps take out the pack (of 80p cells) from the chain...put 4.3V across it , and at this higher voltage the bad cell may get hot, and can be detected. (cells can be charged upto 4.3 and higher ....you get more capacity but drastically shorter life)

If you have a resistive balancing BMS a leaker is a nightmare of energy loss, every other cell in the rest of the entire battery is drawn down to match the pack with the leaker.onstantly unless you turn off the balancing, then you lose capacity because of mismatched packs so your normal voltage swing becomes extreme on the low end because either your bms turns off the battery or the low pack gets extraordinarily low.

Once a pack has gone below shutoff voltage a resistive bms can do nothing to help it recover, an active bms on the other hand can use the energy of the high packs in the battery to help keep the low one from damage as long as possible.

Yes... I was just about to make a post on just that ...If you have an automatic balancer it could be hiding a partly bad cell in one pack , you would be none the wiser , because the balancer would pump more into this pack to keep it up with the others ...all wasted in heat!!

One youtuber has put a voltmeter on each of his 7 packs ...looks very cool ... I don't know how he balances . But even with voltmeters (no balancer) if one pack was continually dropping behind in normal cycling , this could just mean that pack had a lower capacity , and no self discharge problem .

You would have to disconnect the packs from any inputs and outputs and watch if the voltage drops too much on a pack over 24hrs ... if it does charge that pack up to 4.2/4.3 to see if any cell gets hot.

You can find these low volt meters on eBay which give voltage to 2 decimal places (3.78) ,not just one decimal place (3.7)as most do http://www.ebay.co.uk/itm/322372516...49&var=511312860417&ssPageName=STRK:MEBIDX:IT .....$1.4 each.

ozz93666 said:

APD said:

ozz93666 said:

ALL of thecells which hold voltage overnightcould be put in powerwalls ... even those with 100mAHrs will add capacity ... but it's up to the particular builder what he want's to use .... fuses should remove any cells which become heaters during the use of the powerwall.

Click to expand...

BTW, a fuse will not always eliminate a bad cell. A bad cell can draw from all the others at a low rate which does nothing other than lower the capacity of the pack.

Click to expand...

This must be our greatest concern .. a cell in a pack which selfdischarges slowly, not enough to blow a fuse but it will empty the (3.7V)pack given enough time ...

If a cell is already bad in this way ,it cannot hold voltage ,so would be detected andwouldn't be put in a pack ... accurate capacity measurement cannot indicate which cells are about to go bad in this way ... so it seems there's nothing to be done ...

Watching the voltage of each 3.7V pack in your chain (of say7) will show the problem by a packs low voltage ... but then what can you do? If the self discharge is not enough to heat up the offending cell , and itcan't be detected by thermal imaging , there's no way to find the bad cell in the pack without checking every cell!! ...perhaps take out the pack (of 80p cells) from the chain...put 4.3V across it , and at this higher voltage the bad cell may get hot, and can be detected. (cells can be charged upto 4.3 and higher ....you get more capacity but drastically shorter life)

Click to expand...

Ever record voltages under a known load with cells in series? Weed out the weak cells?
Seems it could save on equipment. Voltage measurement is pretty cheap compared to dischargers.

egam said:

Ever record voltages under a known load with cells in series? Weed out the weak cells?
Seems it could save on equipment. Voltage measurement is pretty cheap compared to dischargers.

Click to expand...

An 1156 12V tail light bulb is an almost perfect one amp current draw across a lithium cell. Hard to get much more simple than that.

Proper bms systems do not hide a pack of cells that have higher self discharge. it will actually show you what pack that either is balanced more than others or the other way around depending on if you rung passive or active balance.

I plan to have such recording so you can see what cell is being topped up more than others.

My point of view is safety first!
If you want to risk your house/shed/whatever up to you but if the media caught up your story.. it's gonna beat us hard, every one of us!

But if you really one to skip the testing procedure you can buy factory tested new cells on higher price.

One more thing, monitoring. Volt meters is a good idea on packs but if you can't check it on regular basis it's worth to get an automatic system, like a uC based voltage checker and it can send you a email/sms if something wrong.

FERCSA said:

My point of view is safety first!
If you want to risk your house/shed/whatever up to you ....

Click to expand...

I think everyone here fuses each cell.

With these fuses it's 100% guaranteed certain no fire can in anyway start!

Measuring each cells capacity to the nearest mA has no effecton improvingsafety.

The only real way you can isolate faulty cells early would be with a charger that had some smarts behind it (because you dont want to be there all day,

Charging at 1A, how much does the voltage spike and sag as we switch on charging, idea of ESR,
Temperature Delta Vs ESR at charge rate, if shoots up, its a leaker, and is dissipating the energy rather than storing it

For everything else, and the slow leakers, they need the full charge, but you can atleast get the lost causes out of the way at 5% Charge,

Rerouter, if you put the "smarts" in Excel or something I have already designed, programmed and built a multi channel discharger that could be reprogrammed to do what you are talking about, it can generate a CSV file that could be sent to a spreadsheet for analysis.