Speeding up measuring the capacity of cells .

ozz93666 said:
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 .
Click to expand...

I have been trying to devise a way to process 320 new [old stock] batteries for any outliers or variation in properties. It might be simpler than your endeavor of sorting salvaged cells, but it's sort of similar.

How well would it workif you were to_

* sort your cells by voltage and divide them into groups;
* individually charge or discharge the cellsto the midpointvoltage of the group- e.g. 2.75V 3.0V, 3.25V, etc.;
* build a rack tofit an xsyp array that your charger or power supply can handle and of a convenient size to accommodate your voltage groupings;
* charge or discharge your groups byvoltage, while monitoring the cells for heat and for voltage;


After a sufficient duration or logical stopping point, sort the cells by voltage. If you were charging the array, the weakest cells should have the highest voltages, because their smaller capacities would fill faster. If you were discharging, the weakest should have the lowest voltages.

After finishing the first round of groups, now you have new groups of cells sorted by voltage. Repeat the entire process a few times. Your cells should be sorted by capacity.

The process needs some refinement, but the idea should be sound.

It's sort of like reverse balancing or intentional unbalancing.

It could be done like a bubble sort. Charge the lowest voltage group up to the voltage of the next group. Replace the weak cells with cells from the next group and so on. At the top, reverse direction, replacing the weak ones on the way down. At the bottom, reverse direction again.The ones that are left should be your strongest cells. Regroup the weak ones for the next round.

The bubble sort isn't the most efficient sorting algorithm. Maybe another one will come to mind
 
skyfridge said:
ozz93666 said:
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 .
Click to expand...


* sort your cells by voltage and divide them into groups;
Click to expand...

I'm not sure I understand you ... the first step of the process , which I think everyone is agreed on , is to charge all cellsup to about 4.1V , any that get hot when attempting to do this , and won't take a charge have to be discarded ...

Then when the remainder are charged , disconnect and leave for a day ... any that have lost too much voltage over that day must be discarded too ...
 
ozz93666 said:
skyfridge said:
ozz93666 said:
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 .
Click to expand...


* sort your cells by voltage and divide them into groups;
Click to expand...

I'm not sure I understand you ... the first step of the process , which I think everyone is agreed on , is to charge all cellsup to about 4.1V , any that get hot when attempting to do this , and won't take a charge have to be discarded ...

Then when the remainder are charged , disconnect and leave for a day ... any that have lost too much voltage over that day must be discarded too ...
Click to expand...
Ah, okay. I missed that part. If you've already charged them all to a set voltage (4.1V)to weed out the heaters and corpses...

Batch processing sounds easier than individual capacity tests.

Pick a number for your strings, e.g., 8s. What happens when you charge a string of unmatched cells without balanced charging? You either have to stop the charge when the first one reaches the topor that cell can pop. You either have to stop discharging when the first cell reaches the bottom or the cell can bottom out.

If yourcells capacities aren't well matched to start, the one that reaches the top first is the weakest. The one that reaches the bottom first is the weakest.

Take 8 8s strings. Put them in parallel. Discharge it. Stop when the first cell reaches the bottom. Each string might have a weakling. Pluck out the runts. Who will replace the runts? All of your other cells are hanging out up there at 4.1.

Grab another batch. Bring them down. Pluck out the runts. Etc.

Now everyone is at the bottom. So you'll have to balance the strings to put them in parallel or run the risk of a bunch of racing currents. ..

Like I said, there are refinements to be made.

In OP, you lamented individual capacity tests. I'm looking for a way around that, too. Batch processing would cut down on the work.
 
skyfridge said:
ozz93666 said:
skyfridge said:
ozz93666 said:
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 .
Click to expand...


* sort your cells by voltage and divide them into groups;
Click to expand...

I'm not sure I understand you ... the first step of the process , which I think everyone is agreed on , is to charge all cellsup to about 4.1V , any that get hot when attempting to do this , and won't take a charge have to be discarded ...

Then when the remainder are charged , disconnect and leave for a day ... any that have lost too much voltage over that day must be discarded too ...
Click to expand...
Ah, okay. I missed that part. If you've already charged them all to a set voltage (4.1V)to weed out the heaters and corpses...

Batch processing sounds easier than individual capacity tests.

Pick a number for your strings, e.g., 8s. What happens when you charge a string of unmatched cells without balanced charging? You either have to stop the charge when the first one reaches the topor that cell can pop. You either have to stop discharging when the first cell reaches the bottom or the cell can bottom out.

If yourcells capacities aren't well matched to start, the one that reaches the top first is the weakest. The one that reaches the bottom first is the weakest.

Take 8 8s strings. Put them in parallel. Discharge it. Stop when the first cell reaches the bottom. Each string might have a weakling. Pluck out the runts. Who will replace the runts? All of your other cells are hanging out up there at 4.1.

Grab another batch. Bring them down. Pluck out the runts. Etc.

Now everyone is at the bottom. So you'll have to balance the strings to put them in parallel or run the risk of a bunch of racing currents. ..

Like I said, there are refinements to be made.

In OP, you lamented individual capacity tests. I'm looking for a way around that, too. Batch processing would cut down on the work.
Click to expand...

Always good to brainstorm ....

so we've weeded out the corpses and the heaters ... what is it that we want to do next???????

Well why not use all???? this would be an unacceptable idea for anything mobile , a car or electric bike , it would make the battery too heavy .... but this is for a powerwall , what does it matter if some cells only hold 500mAHr ?? they don't take up much space , they're already beenshown to hold voltage overnight so they won't discharge the pack... the only reason can be that cells with low capacity maybe near the end of their life , could turn bad .... but we're not sure about this ....
Also those with high internal resistance are not efficient , waste more of the power in heat , so they may need to be removed , these are often the cells with low capacity ....
 
Since we are not sure of the low end on how long they actually would last my suggestion is to keep them in their own pack. Then its easier to pull that pack out if... It not turns out they go dead directly. (Low cells with low iR that is)

Braiiin.... And of some reason this image is what i got in my head...



image_spvjam.jpg
 
ozz93666 said:
Always good to brainstorm ....

so we've weeded out the corpses and the heaters ... what is it that we want to do next???????

Well why not use all???? this would be an unacceptable idea for anything mobile , a car or electric bike , it would make the battery too heavy .... but this is for a powerwall , what does it matter if some cells only hold 500mAHr ?? they don't take up much space , they're already beenshown to hold voltage overnight so they won't discharge the pack... the only reason can be that cells with low capacity maybe near the end of their life , could turn bad .... but we're not sure about this ....
Also those with high internal resistance are not efficient , waste more of the power in heat , so they may need to be removed , these are often the cells with low capacity ....
Click to expand...
Now that the cells are all sorted and the duds have been pulled out, it's time to look at what you have and optimize your architecture. Dividing into modulesbycapacity sounds like the most optimal. You don't want the runits bringing down the capacity of an entire pack.

To clarify. As you bring each batch down or up, stopping each time that a cell reaches your low and high voltage limit, each string string is now unbalanced, with voltage indicating capacity. Sort and groupthe cells by voltage. If you were discharging, nowyou can balance them by bringing them all to the same low voltage and iteratively sort the ones of closer capacity after charging back up, where you stop when the first cell reaches the high voltage limit. Sort, groupand balanceagain at the top.

At the tops and bottoms, it is a shorter distance to balance them all for the next round.

After a few iterative sorts, then you can test an individual cell from a group to get an idea of the capacity of each cell in a group, since the cells in the group should have similar capacities
 
skyfridge said:
ozz93666 said:
Always good to brainstorm ....

so we've weeded out the corpses and the heaters ... what is it that we want to do next???????

Well why not use all???? this would be an unacceptable idea for anything mobile , a car or electric bike , it would make the battery too heavy .... but this is for a powerwall , what does it matter if some cells only hold 500mAHr ?? they don't take up much space , they're already beenshown to hold voltage overnight so they won't discharge the pack... the only reason can be that cells with low capacity maybe near the end of their life , could turn bad .... but we're not sure about this ....
Also those with high internal resistance are not efficient , waste more of the power in heat , so they may need to be removed , these are often the cells with low capacity ....
Click to expand...
Now that the cells are all sorted and the duds have been pulled out, it's time to look at what you have and optimize your architecture. Dividing into modulesbycapacity sounds like the most optimal. You don't want the runits bringing down the capacity of an entire pack.

To clarify. As you bring each batch down or up, stopping each time that a cell reaches your low and high voltage limit, each string string is now unbalanced, with voltage indicating capacity. Sort and groupthe cells by voltage. If you were discharging, nowyou can balance them by bringing them all to the same low voltage and iteratively sort the ones of closer capacity after charging back up, where you stop when the first cell reaches the high voltage limit. Sort, groupand balanceagain at the top.

At the tops and bottoms, it is a shorter distance to balance them all for the next round.

After a few iterative sorts, then you can test an individual cell from a group to get an idea of the capacity of each cell in a group, since the cells in the group should have similar capacities
Click to expand...


This describes what I was thinking as well.

The trick is to disable charging or discharging automatically so you don't have to watch it.
The first time it may be fun to watch, but after that it's like watching paint dry. It could be downright unsafe on the charge. It could be discharged using a timer. Discharge for one hour at one amp.
 
Drkangel over on endless sphere has the most sophisticated harvesting program I've seen, I took a lot of lessons from him when I was doing my harvesting, or I did after I screwed it up a few times anyway.

https://endless-sphere.com/forums/viewtopic.php?t=26383

His position seems to be that as long as the banks have the same capacity then the capacity of the individual cells isn't a major issue although it's been a while since I was following his thinking.
 
[quote pid='3579' dateline='1489520308']
This describes what I was thinking as well.

The trick is to disable charg8ing or discharging automatically so you don't have to watch it.
The first time it may be fun to watch, but after that it's like watching paint dry. It could be downright unsafe on the charge. It could be discharged using a timer. Discharge for one hour at one amp.


[/quote]
There's no need to take it all the way to the top or bottom. Unless the cells have a flat discharge curve, such as with LiFePO4 (which I'm using), isn't the curve steep enough to accomodate a strategythat would eliminate risks?

[quote pid='3579' dateline='1489520308']
This describes what I was thinking as well.

The trick is to disable charg8ing or discharging automatically so you don't have to watch it.
The first time it may be fun to watch, but after that it's like watching paint dry. It could be downright unsafe on the charge. It could be discharged using a timer. Discharge for one hour at one amp.


[/quote]
There's no need to take it all the way to the top or bottom. Unless the cells have a flat discharge curve, such as with LiFePO4 (which I'm using), isn't the curve steep enough to accomodate a strategythat would eliminate risks?
 
You must log in or register to reply here.
Back
Top