How important is cell voltage during PCB population?

A

aventeren

New member
Joined
May 20, 2019
Messages
21
Howdy; I am currently in the process of testing (360) 18650 cells using the LiitoKala Lii-500, which can do (4) cells at a time--and I have two units. I also have another "dumb"charger only that does (4) cells at 1A per cell.

My capacity tests are taking 6-8 hrs each at 1A ea--and that is using cells that have been charged to 4.2V first using my "dumb" chargerso that the cells don't have to first be charged up to 4.2V by the Lii-500 units...plus, I am also pulling the cells out after the END of the capacity test while the cells are being charged, which means that my tested cells have Ah capacities and mR resistance figures--but they have not been charged back up to 4.2V, which means that the tested cells are in various voltage states (2.9-4.2V).

I am building (3) 7s16p banks into a 7s48p battery using Jehu's PCBs (plus a BMS per bank). I plan on using rePackr to determine where to put the cells to createbalanced 7s16p banks.

My question is how big of a deal is it that the tested cells be populated in the PCBs be voltage matched? For instance, is it a big deal if I load (7) cells onto a given board--and then add that 7s cell into a larger 16p bank--with all of the individual 18650 cells in various states of voltage levels?

The reason I ask is that to charge the tested cells back up to 4.2V so that all cells have a similar voltage level is taking 2-4 hrs using my "dumb" charger...and given that I have so many 18650 cells to test, re-charging all the cells back to 4.2V is adding quite a bit of charging time. So I'm just trying to figure out if I can let my BMS handle the voltage balancing once my 18650 cells are populated into my 7s16p banks. From what understand, the BMS should balance all of the cells within the 7s16p bank. Is there a reason that I can just let the BMS equalize all of the voltages during the first charge of the 7s16p banks?

Thanks for your help--I REALLY appreciate it.
 
If you connect cells together where the voltage is off by more than about 0.1V (100mV) large currents will flow while one "fast charges" the other.
This could easily blow fuses, melt contacts, melt PCB tracks & damage the cells.
It's pretty important to have them reasonably close - half a volt (empty to full) is going to cause trouble.
If you put say all cells at 3.6V in one bank & then say all cells at 4.0V in another & joined the two in parallel, you'd have the high current thing above.
If you put those two banks in series & tried to use the system, the lower pack would hit cell volt minimum or higher one to maximum way earlier & your system "should" know this & shutdown.
BMS's aren't designed to use currents large enough to equalize like this, it could take weeks.
You could charge the lower pack (separately, disconnected from system unless charger is isolated) up to closer to the higher pack....
Patience is needed :)
 
Agree with above and would add - one of the standard/recommended steps in the process is to charge up to 4.2 and let the cells sit for a while - to find/weed-out self discharging cells. This has the side affect of leaving good cells at very similar voltage (e.g. 4.13-4.17'ish) and you don't have to worry when building/assembling a pack.

If you want to lower the voltage for longer term storage - you can discharge the whole pack or whole batterydown to 3.6'ish (or whatever) in mass:)
 
aventeren said:
[...] but they have not been charged back up to 4.2V, which means that the tested cells are in various voltage states (2.9-4.2V)
Click to expand...

Is 4.2V a typo? (in "2-9-4.2V") After a full discharge they should be resting far below 4.2V. If they have all been fully discharged to very low SOC then it is generally safe to parallel them (they needn't be within 0.1V) because IR (internal resistance) is very high at low SOC - which inhibits large current when one charges the other(s).

Further, generally it is true that Li-ion batteriescan safely handlehigher currents (than rated) at low SOC (that's the property that isexploited to the hilt bymodern fast charging algorithms, such as Philips boostcharging algorithm, or multistage CC/CV algorithms, such asTesla's fast chargingalgorithms,andrecent fast-charging algorithms used in some phones).

Caution The above does notapply to cells that are over discharged. They aregenerally revived via a timedlow current "precharge" stage to bring them back up around 3.0V. Only after such a prechargeshould you apply your standard charging algorithm (be itfast or normal).

Generally manufacturers do not recommend charging any cells that are below 1.5V (though some go to 1.0V) because this can lead to safety problems (copper can dissolve at low voltages and then during chargereplateintodendrites, possibly leading to dangerous internal shorts at some future time - without any warning).For some very old cells (lacking modern safety improvements) it may not even be safe to charge them when below 2.0V.
 
gauss163 said:
aventeren said:
[...] but they have not been charged back up to 4.2V, which means that the tested cells are in various voltage states (2.9-4.2V)
Click to expand...

Is 4.2V a typo? (in "2-9-4.2V") After a full discharge they should be resting far below 4.2V. If they have all been fully discharged to very low SOC then it is generally safe to parallel them (they needn't be within 0.1V) because IR (internal resistance) is very high at low SOC - which inhibits large current when one charges the other(s).
Click to expand...

The cells are various voltage states because I have charged some back up to 4.2V after the capacity test with the "dumb" charger, whereas with others, I just pull the cells out at the end of the capacity test (and those cells are somewhere between the min discharge voltage of 2.8-9V if I catch them right at the end of the capacity test...but more realistically they hit the test end and then start re-charging back up to 4.2V...but I usually catch them in the mid 3V range before I swap them out with new cells to test), and then load new cells into the Lii-500 units to start the testing process. I can charge (8) cells using my one "dumb" charger in the time it takes my (2) Lii-500 units to test (8) cells--so re-charging with my (1) dumb charger makes more sense. I even use my "dumb" charger to do the initial charge before I start my capacity tests so that the test cycles don't have to first charge the cells to 4.2V. Instead I start the test with the cells in the 4.15-4.2V range to speed up the tests.

I now see why people have a whole wall of Lii-500 (or similar) units. This takes forever!
 
>I now see why people have a whole wall of Lii-500 (or similar) units. This takes forever!
It sure does. I have 3 OPUSs... and I'm retired... and I hang out in my office a lot of the day... For me, I just sit with OPUSs on my desk and do 12 at a time inbetween youtubesfor the last 2.5yrs and I'm nearing 10,000 cells processed and 84 packs built to date (e.g. 80kwh battery bank).

I agree - if you want to make rapid progress you either need a lot of smaller chargers or go EV (larger batteries + maybe a larger charger) and/or focus on if the sheer quantity of work is OK for you're situation / time-frame :)
 
aventeren said:
I now see why people have a whole wall of Lii-500 (or similar) units. This takes forever!
Click to expand...
Yes we have all kinds of testers.

image_ausqcw.jpg


Several things to mention.The LiitoKala only charges at 1000mA and its max discharge rate is 500mA.
That is one reason it takes a long time. Especially if you have high capacity cells.
OffGridInTheCity mentioned this and you should really do it. That is to finish the C/D/C sequence so the cells you remove are at ~4.2 V.You can finish this in your "dumb" charger if you wish but you really should finish them so that after 30 day rest you can check for any substantial V drop. This is necessary to find any SD cells you will not want in your pack. After that if you want you can discharge to ~3.6 storage V if you need to store them for an extended period of time.
It wouldn't be a post by me if I didn't mention IR The LiitoKala is not your optimum IR tester by any stretch of the imagination consistency is very lacking.
There are many posts by others and me to discuss this so look for them. Here is a good starter.https://secondlifestorage.com/showthread.php?tid=6575&pid=42065#pid42065

Wolf
 
Crimp Daddy said:
It's very important... all cells should be at the same voltage when populating the boards.
Click to expand...

Why do you believe that? Did you read the (different) answers above?
 
gauss163 said:
CrimpDaddy said:
It's very important... all cells should be at the same voltage when populating the boards.
Click to expand...

Why do you believe that? Did you read the (different) answers above?
Click to expand...

I did not read the above answers... I am providing a general response to ato a general question "How important is cell voltage during PCB population?"

The larger the voltage delta, the larger the inrush current when being put in parallel with cells of different voltage. Ir may or may not exceed the current path.

Voltage delta and cell IR will play a roll in the actual current flow during equalization.
 
^^^ This was already explained in prior answers, where it is was also explained that being "the same voltage" is a needless restriction at low SOC (unless you have a very general definition of "same voltage" - which is not at all clear).
 
OffGridInTheCity said:
>I now see why people have a whole wall of Lii-500 (or similar) units. This takes forever!
It sure does. I have 3 OPUSs... and I'm retired... and I hang out in my office a lot of the day... For me, I just sit with OPUSs on my desk and do 12 at a time inbetween youtubesfor the last 2.5yrs and I'm nearing 10,000 cells processed and 84 packs built to date (e.g. 80kwh battery bank).

I agree - if you want to make rapid progress you either need a lot of smaller chargers or go EV (larger batteries + maybe a larger charger) and/or focus on if the sheer quantity of work is OK for you're situation / time-frame :)
Click to expand...
Yeah that was kinda my question. Could I just drop all of these tested cells into a 7s16p battery with a BMS, and then just hook up my 29.2V charger, and let the charger bulk charge all of the cells for a week. I would just do this one time--but that would then allow me to re-charge all of the cells back to 4.2V using the bulk charger in the 7s16p battery rather than having to re-charge them all (4) at a time in my "dumb" charger or with my Lii-500 chargers after all of the cells have been tested. I suppose I could then let the charged 7s16p battery rest for 30 days, and then test all of the 7s packs using my BattGo Meter to find any self discharge cells within the 7s packs.

Do you think that the above would be an acceptable way to do a one-timebulk re-charge these various cells--instead of having to re-charge them all after my testing using my one "dumb" charger and my (2) Lii-500 units after the testing is complete?

Wolf said:
aventeren said:
I now see why people have a whole wall of Lii-500 (or similar) units. This takes forever!
Click to expand...
Yes we have all kinds of testers.

image_ausqcw.jpg


Several things to mention.The LiitoKala only charges at 1000mA and its max discharge rate is 500mA.
That is one reason it takes a long time. Especially if you have high capacity cells.
OffGridInTheCity mentioned this and you should really do it. That is to finish the C/D/C sequence so the cells you remove are at ~4.2 V.You can finish this in your "dumb" charger if you wish but you really should finish them so that after 30 day rest you can check for any substantial V drop. This is necessary to find any SD cells you will not want in your pack. After that if you want you can discharge to ~3.6 storage V if you need to store them for an extended period of time.
It wouldn't be a post by me if I didn't mention IR The LiitoKala is not your optimum IR tester by any stretch of the imagination consistency is very lacking.
There are many posts by others and me to discuss this so look for them. Here is a good starter.https://secondlifestorage.com/showthread.php?tid=6575&pid=42065#pid42065

Wolf
Click to expand...
Fantastic link. Thank you! I'm seeing my mR readings using my Lii-500 tests somewhere in the high 20s to 50s. I'm don't think I've ran across one yet over 50 mR. FYI, these are all new 18650 Panasonic NCR18650BD cells.
 
aventeren said:
[...] Could I just drop all of these tested cells into a 7s16p battery with a BMS, and then just hook up my 29.2V charger, and let the charger bulk charge all of the cells for a week [...]
Click to expand...

You could do something like that to bulk-charge all the discharged cells. Then add the other already-charged cells to the newly charged (sub)pack and do a final charge (with low termination) to help top-balance them. Or, you could first discharge the high ones, then put all the discharged cells in a pack (generally it is safer to work with lower voltages), but then you might need some extra work to ensure they get top (vs. bottom) balanced (esp. necessary of the capacities are not closely matched).
 
aventeren said:
FYI, these are all new 18650 Panasonic NCR18650BD cells.
Click to expand...
Ok if they are brand new cells is there a reason you are testing them?
If I had a lotof new cells I would check them for storage/shipping V when they arrive and if that is OK build the pack.
Charge the pack test and end of story.

That way you are building the pack with a lower V usually ~3.6V don't have to worry about SD's charge up the parallel pack to 4.2 run a capacity test and your done. I wouldn't even worry about IR (well I would measure it cause I am anal).

Build your series pack and enjoy.

Wolf
 
Wolf said:
aventeren said:
FYI, these are all new 18650 Panasonic NCR18650BD cells.
Click to expand...
Ok if they are brand new cells is there a reason you are testing them?
If I had a lotof new cells I would check them for storage/shipping V when they arrive and if that is OK build the pack.
Charge the pack test and end of story.

That way you are building the pack with a lower V usually ~3.6V don't have to worry about SD's charge up the parallel pack to 4.2 run a capacity test and your done. I wouldn't even worry about IR (well I would measure it cause I am anal).

Build your series pack and enjoy.

Wolf
Click to expand...

I guess I'm testing them because everything that I've watched or read says that I need to be populating each PCB with cells that are within 100 Ah of each other--and I'm seeing some that are outside of 100 Ah in my testing. I guess I don't understand the impacts well enough to know what populating one 7s PCB with a bad cell will do to a 7s or 7s16p or 7s48p battery. It's that uncertainty with not wanting to burn down my house that is leading me to test all of these.

But then that has created this voltage differences in the cells, as the ones that I test are somewhere between 3-4.2-ish volts...and this has created a balancing issue where I am now being told not to populate a PCB with cells that aren't within 0.1V of each other because I'll damage the PCB. So candidly, I'm getting myself wrapped around the axel here a bit.

For the most part, IR seems within 30-60 mR, so that's probably good.

I just remembered that I bought a BattGo BS-8G battery tester & balancer. It maxes out at 2A. So maybe I just populate these PCBs one at a time, and then set the BattGo unit to balance mode, which will then balance the voltage out....and then once that is done, add that 7s unit in parallel to the other balanced 7s units to populate the 7s16p banks...and then use the BMS and a 29.2V charger to charge up that bank to full capacity (the charger only delivers 10A at 29.2V, so on a per 18650 cell basis, the charing current will be a trickle). But it's the uncertainty of this whole thing that I'm trying to understand so that I don't mess something up.

I'll look back at this thread in 6 months and laugh--but I'm definitely on the steep part of the learning curve now.

Thanks a ton for your help! This is fun!
 
Re assembling low SoC cells, then charging a whole pack, sounds OK, just don't include any "full" cells.
Also somewhere in your testing you need find SD cells & they have to be separate for this.
Like Wolf suggests, testing IR is a key tool to weed out old/damaged cells too, so keep doing that.
 
aventeren said:
I guess I'm testing them because everything that I've watched or read says that I need to be populating each PCB with cells that are within 100 Ah of each other--and I'm seeing some that are outside of 100 Ah in my testing [...]
Click to expand...

Presumably you mean 100mAh. Don't expect much better than around 5% capacity accuracy with low-end analyzing chargers like the Lii-500. In particular, for your 3Ah cells the reported capacity could be off by 150mAh, so the 100mAh differences you are seeing are well within the noise. Keep in mind there will also be wide variances between slots, and between different units. You can't expectbetter at such a bargain-basementprice point.

A google search will turn up much evidence to support such, including a thread on this site, and alsosome sources of dubious competence,e.g. on the page of the featured google snippet is the following nonsensical remark, which raises serious doubtsthat the author has the required knowledge to properly perform Lii-500 accuracy tests

theBatterydoctor said:
Notice that NO capacity tester is 100% accurate that is because the battery cells "suck up"different amount of ions every time! compare this to a machine filling up a glass of water to exactly 100.00% of every time, no matter how good a machine is it will never be the exact same amount if microlitres in the glass [,,,]
Click to expand...

In fact, with higher end equipment one can get accuracy very close to 100% (and charging alone sufficessince Li-ion batteries have close to 100% Coulombicefficiency). That's how Tesla's guru Jeff Dahn manages to efficiently empirically optimize Li-ion chemistry- by making various chemistrytweaks and observing infinitesimalcapacity changes during cycling. But this is of little importfor the matter at hand.
 
aventeren said:
I'm curious to know how a 0.1 V difference will damge the pcb ?
CG.
Click to expand...
 
You must log in or register to reply here.
Back
Top