What Happened - EG4 18Kpv and 2 All Weather Wall Mounts

[garybryan33]

It's been overcast and raining for the past four days, so I've closely monitored my off-grid system.

With each passing day, I saw the SOC dip and bounce back as follows,

on day one, from 100 to 82 overnight,

on day two, it crept up to 95 and dropped to 77 overnight,

on day 3 it crept up to 88 and dropped to 68 overnight.

On day four the rain was heavy and the clouds thick, there was going to be some cloud break the next day so I had high hopes that the batteries were going make it given that they dropped about 20% each night the past few nights. I figured they would dip to about 48% at worst before the solar panels kicked in.

The last time I checked SOC was 59% and a half hour later the system was offline. I went out to physically check, it was powered down. I cycled the system, it powered up and shut itself down again. Both batteries read 45.8 volts on the display with 0% SOC they also say "on standby"

I can't believe they went from 59% to zero in a 1/2 hour. I thought the inverter cut off power at around 20% to protect the batteries. Is there a setting that I have wrong?

Needless to say, I am a bit vexed at the moment. Any suggestion on how to resolve this would be greatly appreciated. BTW, the system has been running without issue for about 8 months.

 

[wpns]

Closed loop comms between the inverter and the batteries?

How are you reading SOC?

 

[shadowmaker]

45,8V is 2,86V cell voltage which is under 10% SOC.

 

[42OhmsPA]

What are your individual cell voltages when the batteries are full and empty?

Sounds to me like cells aren't balanced and the lowest cell(s) triggered a low voltage disconnect.

 

[rogerheflin]

In general most BMS's if not fully charged or fully discharged for a few days lose track of SOC. I have had mine (after not being full or empty for several days) go from 20-30% SOC to 0% soc once the BMS realized via a low voltage that the SOC was wrong. For the low limit I started using voltage because SOC being off by 20% is enough to cause this to happen, and it seems quite easy to get off by 20-30%. with SOC.

SOC is supposed to work by counting current in/out but it seems that there is some significant inaccuracy in how most of the BMS makers do this, best guess is their current meter is not very accurate across its full range of current and if not very accurate at lower current (say 20-30A) that most of our systems likely spend most of their time at then SOC rapidly becomes very misleading/wrong.

 

[garybryan33]

Closed loop comms between the inverter and the batteries?

How are you reading SOC?

I have been monitoring via the soc% on the app and confirmed on the inverter display. Here is a picture of the Soc when it dropped off, it looks like it was around 53%. That said, this morning we have clearing skies and it looks like the batteries are taking charge.

 

[garybryan33]

45,8V is 2,86V cell voltage which is under 10% SOC.

Math does not lie! But the drop was abrupt which raises concern

 

[garybryan33]

What are your individual cell voltages when the batteries are full and empty?

Sounds to me like cells aren't balanced and the lowest cell(s) triggered a low voltage disconnect.

On average, the batteries drop to about 60 SoC overnight; by noon, they are at 100 SoC. I thought the system balanced the cells automatically when it reached 100 SoC.

 

[wpns]

On average, the batteries drop to about 60 SoC overnight; by noon, they are at 100 SoC. I thought the system balanced the cells automatically when it reached 100 SoC.

It’s supposed to. Mine certainly do. Can you see the individual batteries in the monitoring website?

 

[garybryan33]

In general most BMS's if not fully charged or fully discharged for a few days lose track of SOC. I have had mine (after not being full or empty for several days) go from 20-30% SOC to 0% soc once the BMS realized via a low voltage that the SOC was wrong. For the low limit I started using voltage because SOC being off by 20% is enough to cause this to happen, and it seems quite easy to get off by 20-30%. with SOC.

SOC is supposed to work by counting current in/out but it seems that there is some significant inaccuracy in how most of the BMS makers do this, best guess is their current meter is not very accurate across its full range of current and if not very accurate at lower current (say 20-30A) that most of our systems likely spend most of their time at then SOC rapidly becomes very misleading/wrong.

Thank you for your input. When I first launched the system I applied the settings via the phone app, but when I access the settings on the desktop the current settings do not display. One of the first things I had to do was, switch the frequency from 50 to 60 to get my well pump working so I know the app is making changes. What are best practices when it comes to properly inputting settings?

 

[42OhmsPA]

On average, the batteries drop to about 60 SoC overnight; by noon, they are at 100 SoC. I thought the system balanced the cells automatically when it reached 100 SoC.

It should, however they could need an extended time at 55.2V or higher to balance out depending how bad the deviation is.

 

[garybryan33]

It’s supposed to. Mine certainly do. Can you see the individual batteries in the monitoring website?

If available I can't find that page. It just shows the battery as a collective whole, I believe.

 

[garybryan33]

It should, however they could need an extended time at 55.2V or higher to balance out depending how bad the deviation is.

They generally stay at 100 soc from noon till dusk, on average 6+ hours in summer and at least 4 hours during the winter.

 

[wpns]

If available I can't find that page. It just shows the battery as a collective whole, I believe.

If everything is at the latest firmware revision, and you set the Lithium Brand to 1 you should be able to see the batteries individually on the monitoring website.

 

[wpns]

They generally stay at 100 soc from noon till dusk, on average 6+ hours in summer and at least 4 hours during the winter.

That should be more than enough time for balancing.

 

[42OhmsPA]

That should be more than enough time for balancing.

Agreed.
Still would like to see cell voltages / deviation when "full and empty"

 

[zanydroid]

Total voltage graph as a timeseries for the past week might be helpful. And maybe with the summer as a comparison (I hesitate to call it a control)

Power consumption graph as well for the day where the battery nosedived, maybe there was an unaccounted load that clicked on.

How does one pull the cell-level data from this hardware?

 

[wpns]

How does one pull the cell-level data from this hardware?

You need to talk directly to the batteries, usually with the BMS Tools app on a windows machine. I’d reach out to @EG4TechSolutionsTeam to see if they can offer any advice.

 

[shadowmaker]

Math does not lie! But the drop was abrupt which raises concern

Well sadly my math can, but lifepo can't drop like that unless it reaches "lower knee" around 3,0-3,1V cell voltage (about 10-20% SOC) or batt has bad bottom balance (for 45,8V resting voltage that would mean there are many "bad" cells which is unlikely but possible). I suspect there's error with determining SOC%, but then I'm usually wrong.

Like others have already said you need cell level monitoring to figure this out.

 

[zanydroid]

You need to talk directly to the batteries, usually with the BMS Tools app on a windows machine. I’d reach out to @EG4TechSolutionsTeam to see if they can offer any advice.

(It's a pre-purchase question) Does that require unhooking any comms cable / buying a comms dongle to go straight from PC to the battery? Or does the inverter proxy from a network connection -> the battery?

If the BMS determines during top-balancing and SOC calibration that it is way off, how does it interact with the communicating inverter (both the logic and the config parameters that impinge on this) to resume charging so it can determine the real 100%? Presumably the inverter needs to be smart enough to do something like, turn off all SOC-based charge stop logic, and go to CC/CV for a while.

 

[garybryan33]

Agreed.
Still would like to see cell voltages / deviation when "full and empty"

I will be downloading the BMS tools app, I will update as soon as I can,

 

[garybryan33]

Total voltage graph as a timeseries for the past week might be helpful. And maybe with the summer as a comparison (I hesitate to call it a control)

Power consumption graph as well for the day where the battery nosedived, maybe there was an unaccounted load that clicked on.

How does one pull the cell-level data from this hardware?

Agreed.
Still would like to see cell voltages / deviation when "full and empty"

Might I have created this issue? In July and August, I got an intermittent notification that the system voltage was high. I have included that particular day (Aug 15) in the attachments. The other is a sample of the readings after a settings change.

I went into the settings and changed the cutoff voltage to a lower setting. I never got the high voltage reading again, but in looking at the history the system did not go above 55.2 volts to balance the cells.

For the life of me, I cannot remember what setting I changed. Looking at the settings today they all look correct.

 

[42OhmsPA]

As long as they got to 55V on a regular basis they should have balanced.

I didn't download your txt files (weird about attachments on my main device...) I don't own an 18KPV.
I charge to 55.2V or 55.6V depending on the time of year. Winter months make it difficult to keep packs in balance.

 

[EG4TechSolutionsTeam]

What battery protocol is configured for the All-Weather batteries, and which lithium brand is selected on the inverter?

 

[Quattrohead]

On the website portal, look for the "System Charge Volt Limit(V)" setting under "Application Setting"

I had mine set to 56v to be cautious and it worked fine for a couple of months, then I noticed my system SOC would only go to 99%
1 of the batteries was only getting to 97% but the other 6 were at 100%.
Changed this setting to 57v and got to 100% SOC today