I think I found the issue... HEATERS!

hbpowerwall

Administrator
Joined
Oct 7, 2016
Messages
2,218
Video says it all - but short story - I think two cells were beside each other keeping each other warm

 
Based on seeing the videos, pictures, and chatter over in the other thread (https://secondlifestorage.com/index.php?threads/when-things-go-wrong-it-can-go-really-wrong.10317/), I will have to agree.

There should of been warning signs evident that were ignored, or misunderstood. A rise in balancing of the other packs should have triggered an alert to check the "non-balancing" pack(s) as to why it didn't need balancing as much as the others, or why the others have increased so much.
 
I think the takeaway from this is that 1Amp balancing current is great at masking problems.
For now, it should be significantly dialed back (50mA?) for regular use, so that faulty cells cause an imbalance in short order and trip the low voltage triggers.
Long term, agree with Korishan re need for the BMS to monitor abnormal balancing and sound alerts, and eventual shutdown.
 
I think the takeaway from this is that 1Amp balancing current is great at masking problems.
For now, it should be significantly dialed back (50mA?) for regular use, so that faulty cells cause an imbalance in short order and trip the low voltage triggers.
Long term, agree with Korishan re need for the BMS to monitor abnormal balancing and sound alerts, and eventual shutdown.
Yeah, kind of my position on all the balancing issue. A BMS which cannot tell you how much energy was transferred from where to where is pretty much useless in terms of balancing. Spending a lot of money in active balancers which at the end only light up some LEDs are just dangerous.
Let cells go out of balance! It is just indicating that there is a problem or a design flaw. Of course that is not a convenient situation. We tend to throw electronic solutions on every problem just so it manages itself away. But in DIY that might be the wrong approach. We need to stay in touch with what we have built. Don't get lazy or complacent...
 
Taking this a bit further... I don't need/have-balance turned off for regular operation. My 84 (healthy) packs slowly drift from a daily 40mv-60mv max difference to a 40mv-90mv max difference after 5 months. At that point, I do some touch-up balancing for a few days to bring it back to 40mv-60mv and they are good again for another several months. So far, my oldest 14packs are nearing 900 cycles but show no trouble staying in balance with the newest packs. I take all this as a criteria of health packs.

So healthy packs shouldn't need much balance. My plan is to aggressively fix (retest/rebuild/discard) packs that will no longer run for ... at least a month? ... without balance. By keeping on it I hope to avoid serious problems. It will be interesting to diagnose a pack that won't run without balance when I get one.
 
Last edited:
Based on seeing the videos, pictures, and chatter over in the other thread (https://secondlifestorage.com/index.php?threads/when-things-go-wrong-it-can-go-really-wrong.10317/), I will have to agree.

There should of been warning signs evident that were ignored, or misunderstood. A rise in balancing of the other packs should have triggered an alert to check the "non-balancing" pack(s) as to why it didn't need balancing as much as the others, or why the others have increased so much.
Yes, I would say the warning signs were ignored.. perhaps the train of thought was its just the bms doing what it should do.
I think the takeaway from this is that 1Amp balancing current is great at masking problems.
For now, it should be significantly dialed back (50mA?) for regular use, so that faulty cells cause an imbalance in short order and trip the low voltage triggers.
Long term, agree with Korishan re need for the BMS to monitor abnormal balancing and sound alerts, and eventual shutdown.
That's an interesting take on it..that said the longmons are easily turned down or up as needed with in the software.
Yeah, kind of my position on all the balancing issue. A BMS which cannot tell you how much energy was transferred from where to where is pretty much useless in terms of balancing. Spending a lot of money in active balancers which at the end only light up some LEDs are just dangerous.
Let cells go out of balance! It is just indicating that there is a problem or a design flaw. Of course that is not a convenient situation. We tend to throw electronic solutions on every problem just so it manages itself away. But in DIY that might be the wrong approach. We need to stay in touch with what we have built. Don't get lazy or complacent...
BMS well at least the WM4 and later all have years and years of loging also shows how much each cell has bypassed.
Taking this a bit further... I don't need/have-balance turned off for regular operation. My 84 (healthy) packs slowly drift from a daily 40mv-60mv max difference to a 40mv-90mv max difference after 5 months. At that point, I do some touch-up balancing for a few days to bring it back to 40mv-60mv and they are good again for another several months. So far, my oldest 14packs are nearing 900 cycles but show no trouble staying in balance with the newest packs. I take all this as a criteria of health packs.

So healthy packs shouldn't need much balance. My plan is to aggressively fix (retest/rebuild/discard) packs that will no longer run for ... at least a month? ... without balance. By keeping on it I hope to avoid serious problems. It will be interesting to diagnose a pack that won't run without balance when I get one.
A well build battery Tesla for example what do they have 0.01amp balance current.. Also home energy storage are often smaller batteries, and cycle more hence drift further.. the bigger the battery the less drift.
 
Just read these posts after posting on another thread...
The problem is the BMS does not measure enough to determine the real Wh dissipated (accurately) and the voltage measurement accuracy is not self calibrated over an extended period of time to allow a real pattern to be detected very early on. Korishan has it spot on and ajw22 comment of "balancing current is great at masking problem" sums it up.
 
Batrium will happily and easily show me my balance current and at a glance you can see the good, bad and evil cells in the bunch...
 

Attachments

  • Balance Current.JPG
    Balance Current.JPG
    680.4 KB · Views: 158
Wow are all these packs bee online the same amount of time? Some of those have massive bypass numbers
 
[...] Some of those have massive bypass numbers
Just FYI, the ones with massive bypass numbers are most likely the good packs. They are dumping energy (bypassing) on purpose to match the leaky packs with lower SoC.

The counts of course need to be rest after the problems have been sorted out. Otherwise it'll get very confusing trying to identify new problems developing.

I'm sure Pete is already on it, but I'd immediately take out #45 - it has self discharging cell(s) for sure. And all other packs with less than 200mAh bypass are very suspicious, too.
 
Last edited:
Since Pete has a 14s system the cells are grouped in packs of 14, so cells 42 to 56 are in "the pack with the problem" & yes agree it looks like cell #44 is a heater/bad SD(self discharge) & I'd also be very suspicious of cell #51 for the same.
Of the rest, only cell 2 cells stand out: #24 (looks like it might be low capacity) & #28 (likely has an SD cell or two).
 
Question with heaters, I understand that when charged they will immediately self-discharge, by shorting out and generating heat. Now the question is does it also continue to take the energy from the other cells, therefore depleting the entire pack?

If so, then it's probably a good idea to routinely take packs out of service (especially during summers when there's plenty of sun), then let it sit idle for a week to see if any of the packs starts to drain. There should never be any drain in a week, so any drop in voltage would be suspect.
 
Question with heaters, I understand that when charged they will immediately self-discharge, by shorting out and generating heat. Now the question is does it also continue to take the energy from the other cells, therefore depleting the entire pack?

If so, then it's probably a good idea to routinely take packs out of service (especially during summers when there's plenty of sun), then let it sit idle for a week to see if any of the packs starts to drain. There should never be any drain in a week, so any drop in voltage would be suspect.
This is my point. Once a balance has been done (for a few days) then balance can be turned off - no balance at all for months. In my case, I can go 5 months before a max difference of 60mv creeps to 90mv among 84 packs - e.g. healthy packs don't need daily balance. Then, a few days of auto-level bring it back to 60mv max difference.

So I'm looking at the idea of fixing/replacing any packs that require balance more than once a month as a criteria. Maybe once a month is too aggressive or not aggressive enough - depends on what is wrong with the pack. So far, I haven't had a pack to 'fix' - so I'll have to see as time goes on.
 
Huh that is kinda opposite of I would have thought. I was thinking that lower capacity packs would bypass more since they would reach higher voltage sooner it would start to bypass. but yes after changing packs one would need to backup the data then reset it to get an Idea of the Health of the entire system.

Also we only balance during charging and not while in discharge
 
Correct. I do same. I balance 2 times per year in average you could say. Though in summer when i got excess energy i do it more often just because i can. But i live more than 6 month during winter without balancing.
Balancing is only needed if
1. You have packs that self discharge
2. You push huge currents in or out of the packs that the losses (IR) affects it unevenly between the packs.

The later is generally not a problem on larger packs. The first is easiily spotted that a pack differ in voltage in top range over time and or by checking balanced current. In passive systems thats basically the packs never being balanced/bleeded.
 
45 and 47 are the worst, I've addressed them twice taking them from service 'fixing' them then putting them back into service and they go south again fast.. Now some things to mention - I know about the problem and check almost weekly, there isn't any noticeable heat, they have lower overall AH storage & 45 is #1 cell.. ie Positive closest to the inverter so has the heaviest load, they are also the 1st batteries I made years ago. This weekend they are having a retirement party. So I'll drop from 40kwh storage to 30
 
Also If I was using an active balancer I would never see this kind of data accumulate over time. Another thing is that the first 42 cells for the past year or more are all linked so three longmons per 'battery'. Another issue is I mess with my batteries ... well used to so lots of long mons have come off and put back up in a different order. Firmware updates might mess with that data also i guess. (I do lots of testing)
 

Attachments

  • batrium.JPG
    batrium.JPG
    454.1 KB · Views: 105
What we need is an active balancer with tracking.
Most active balancers are dumb, but I wonder if the Heltec BMS with active balancing records balance info.
Technically this data is found in a BMS that records the voltage. You can see which cell packs that are always Lagging to full charge.

Some problems with balance could also be due to different chemistries charging quicker than others even with the packs being of the same capacity.
 
Not sure if this would work.
Use a programable BMS. Turn off balance
Have Electrical jacks on the packs so the programable active balancer could be used once or twice a year. Then removed.?
later floyd
 
>What we need is an active balancer with tracking
Batrium auto-level is an 'active balancer' and the mah per longmon are tracked as shown above. So you could achieve this goal with Batrium.
 
Back
Top