@jlankford 

Thanks for your thorough response. I understand what you're saying, but in one sense, I don't care which is out of whack, one battery, or the others, the voltage or SOC--I just want the damn things balanced so I have a sense of how much capacity I have left. I have a significant chunk of change invested in these and the inverter, and I just want the features they have to work well enough that I can go a couple hours without feeling like I need to check on them.

But, I see that going off the voltage, they are all pretty close--however, once they start to discharge, that one battery that the SOC always starts dropping first, and fastest, is always the lowest voltage by just a little, and also show the highest current discharge--at low loads by .5-1.5 amps, higher load 2+amps. Further, the battery that shows the highest SOC, is always the last to register any number of amps in discharge--usually 0 at low loads, often half of that registered by what I'm calling the lowest battery by SOC. And, as the SOC values between the high and low batteries grows, so does the difference in voltage between them.

I'm reluctant to extrapolate the SOC based on voltage, and then bank on that--being cautious because I don't want to tank the battery. I believe there is some difference in how they are discharging, and that 2 batteries are misbehaving, with opposing performance--ignoring the SOC. I've done as suggested by 2 different folks at EG4, and I don't see that anything has changed significantly. I am going to check on the continuity between the batteries and bus--but I'm not really expecting to learn anything there. I wiped all contact areas with denatured alcohol before assembling, and torqued to spec. Unless one of the battery cable lugs was crimped poorly, there shouldn't be any significant difference.

I am also concerned about the generator charging based on SOC, as you describe testing. As you point out, the total battery bank SOC will register a higher number than the individual batteries. I was already doing the math to try to estimate when I might need to start generator charging and it seems like 60-65% SOC, if I want to be safe--based on SOC, but not really sure exact SOC... And, if you need to do this, it'll be kicking the generator on earlier, and thus more often than I should really need with the 5 batteries--maybe I should've only bought 2--and spent the rest on propane.

Another thing to note, when I have been at the cabin, the numbers I'm seeing on monitor center are consistent with those displayed on the individual batteries. I tend to be more familiar with the monitor center, than the 6000xp display, mostly because I use it more, being away from the cabin for 2 weeks at a clip--and only being there for 2 days at a time. Plus, I haven't found the 6000XP to be very user friendly.

Right now, I have a frustrating situation where the cbin indoor temp is approaching freezing, and not having a reliable sense of the SOC, I'm reluctant to run the heat, but I'm going to have to because the temps if going to go down to around 5 degrees tonight, and it'll probably drop to 10-15 inside. The battery manuals don't recommend charging batteries below freezing--right now, they're about 33 degrees. We might actually get some sun tomorrow, but even if we have full sun, it won't offset what the mini-split will use togetthe temp up to around 40. I might just put it on standby, until I can get up there after Christmas.

Thanks, Brad

Posted by: @brdfrd

↑

@jlankford 

Thanks. So, if I had a higher continuous draw--maybe at least 200 watts or more--the discharge may be more likely to be spread across the batteries, balanced? When I was at the cabin, running the mini split at +/- 1000 watts, that 4th battery still discharged faster, although they all discharged.

 

I think the discharge is spread across the batteries regardless, but most of them are simply not measuring it.

When you say that one unit discharges faster than the others at higher loads, are you talking about SOC numbers, or discharge current? If all batteries display the same rested voltage (remember my previous suggested test), then when there's a load high enough to be registered by all batteries they should all be displaying the same discharge current +- a small margin. If you are still trying to compare SOC numbers, then unless your higher load is constant there are still going to be time periods of lower current where Batt_ID_4 is going to drop its SOC faster than the others. So, the SOC divergence would continue during the higher usage, i.e when the mini-split cycles off.

If Batt_ID_4 is really discharging faster than the others based on amperage numbers, then I would reexamine all of the battery connections, bus connections, etc. If you have an accurate meter you can measure continuity at every cable interface and confirm it's 0.0 ohms. A better check would be to pull a high steady current load from the system and measure voltage drops across each cable interface.

<edit>

Don't rely on the web monitor for comparing battery discharge current - the lag and update period is too slow. You need to be at the rack waking up each battery and looking at what its display reports.

@brdfrd 

FWIW, here is the status of my system right now. This is exactly 36 hours after the last full charge took place. With 1,800 watts of EPS standby consumed per day, my system should be down 2,700 Wh right now, which is 8.9 % of my total capacity. Note the data shows my Batt_ID_04 is down by exactly how much its 9% contribution should be, while the others still show 100%.

It appears that Batt_ID_04 is showing me a SOC value which can be used to know my entire battery bank state. Not in all cases, but in this very slow discharge case it is. It's possible that your Batt_ID_04 is giving you a similar representation of your entire bank status, while all the other batteries' SOC values are off in the weeds. Note that my inverter is telling me I have 99% SOC and 591 Ah available, when in fact the amount is 91% and ~546 Ah. You can't depend on the totals your inverter is displaying during this slow discharge scenario.

I've spent enough time measuring and characterizing this behavior to prove what I am saying is true for my system. I've almost reached the point of understanding what is happening enough to start ignoring it. All of my batteries are working together at both low and high discharge rates. The BMS SOC calculations are very misleading, which in turn causes the inverter's total system capacity status to be misleading.

Posted by: @brdfrd

↑

@jlankford 

Thanks. So, if I had a higher continuous draw--maybe at least 200 watts or more--the discharge may be more likely to be spread across the batteries, balanced? When I was at the cabin, running the mini split at +/- 1000 watts, that 4th battery still discharged faster, although they all discharged.

Yes, that’s correct. If you start with a fully charged system, all reporting 100%, and then give it a constant 200 watt discharge load, I predict you will see all the batteries reporting nearly the same SOC as the system discharges. And it doesn’t need to be a constant discharge. It can be a random load as long as the floor doesn’t go below 200 watt. Next time at your cabin fully charge the system, find some constant 200w load you can turn on, and then let the mini-split cycle on and off as needed. I expect after a day you will be satisfied that what I’m theorizing is true.

Note that the discharge is evenly spread across all batteries even at low discharge. It's only the calculated SOC values that are not "evenly spread."