Battery banks, parasitic loads and low sun hours

[MurphyGuy]

This might irritate you but when I turned the Sol-Arks back on to get the batteries back to 100% I noted a 5% drop in SOC (until hubby gets back that's all I have). If that SOC was correct (beginning and ending) that would mean I dropped around 2.7 kwh in about 23 hours. I too look forward to confirming these numbers.

Can you log into your BMS interface and post a screen shot of you cell voltages on one of your 5.4kW packs? I don't need every pack, just pick one at random.
You are basing you 5% calculations on the assumption that each of your packs is 5.4kWh in capacity and that may not be accurate. While the pack might have that much capacity, full capacity and usable capacity are usually 10% to 20% different. Charging a large battery pack up to full voltage and draining it to empty is harmful to the battery's life expectancy, and while it can be done, most manufacturers set BMS limits to chop off a little at the top and a little at the bottom. I can see the voltage specs using the Fortress Power spec sheet, but nothing on their website indicates if your packs are 15s or 16s configs.

For now, lets assume 16s config and use your numbers from your 5% SOC drop. Assuming your 10 packs are 5.4kwh each, that's 54000 watts.. x 0.05 = 2700watts-hours of capacity loss (2.7kWh). So 2700 divided by 24 hours = 112.5 watts per hour, or a bit over 10 watts per pack.. Assuming the BMS is doing some balancing, that's more reasonable to expect.

Would you say the figure showing harvested PV is likely accurate on a Sol-Ark?

Yes.. those numbers are generally accurate no matter what inverter is being used.

Until our conversation, I thought all of the Sol-Ark reported data was accurate. I'm pretty sad that it's not. I'm beginning to dislike my choice in inverter as well.

It kind of depends. I'm not sure who's reporting SOC information in your case but I suspect your Sol Arks have nothing to do with it other than acting like a user interface for the battery BMS's. In other words, the batteries are reporting SOC and the Sol Arks are just displaying the information for you. (not entirely sure on that one). It is possible that the inverters are attempting to report their own calculated SOC based on your input parameters but I suspect that isn't the case.

 

[MurphyGuy]

All 4 of my devices that report SOC differ from each other, the longer between 100% charged day's the worse it gets. Simple voltage is by far the best metric dispite what Will says ?. You just have to remember you are dealing with millivolts here not 1 decimal place.

And that is even more true for phosphate cells that products like fortress power use.. It's not so bad with regular lithium ion.

 

[SHTF Power Anyway]

Can you log into your BMS interface and post a screen shot of you cell voltages on one of your 5.4kW packs? I don't need every pack, just pick one at random.
You are basing you 5% calculations on the assumption that each of your packs is 5.4kWh in capacity and that may not be accurate. While the pack might have that much capacity, full capacity and usable capacity are usually 10% to 20% different. Charging a large battery pack up to full voltage and draining it to empty is harmful to the battery's life expectancy, and while it can be done, most manufacturers set BMS limits to chop off a little at the top and a little at the bottom. I can see the voltage specs using the Fortress Power spec sheet, but nothing on their website indicates if your packs are 15s or 16s configs.

For now, lets assume 16s config and use your numbers from your 5% SOC drop. Assuming your 10 packs are 5.4kwh each, that's 54000 watts.. x 0.05 = 2700watts-hours of capacity loss (2.7kWh). So 2700 divided by 24 hours = 112.5 watts per hour, or a bit over 10 watts per pack.. Assuming the BMS is doing some balancing, that's more reasonable to expect.



Yes.. those numbers are generally accurate no matter what inverter is being used.

It kind of depends. I'm not sure who's reporting SOC information in your case but I suspect your Sol Arks have nothing to do with it other than acting like a user interface for the battery BMS's. In other words, the batteries are reporting SOC and the Sol Arks are just displaying the information for you. (not entirely sure on that one). It is possible that the inverters are attempting to report their own calculated SOC based on your input parameters but I suspect that isn't the case.

I am not meaning to ignore this request. I didn't see it until this morning. My neighbors shanghai'd me yesterday after my farm chores were done.
I went out this morning and stood there while my inverter recharged my batteries to exactly 100% SOC and, while I was watching them hit that mark, I was attached to the BMS and had all that data on the screen. I chose the "Record" option and assumed it saved 10 files showing the detailed status of each battery as it did the other day when I performed this function. I came inside and it's not there so I'm going to have to try again. In the meantime, I did get a picture of what the Sol-Arks showed when they reached that 100% SOC. This system is still not connected to any loads so the numbers we're seeing on Master and Slave for PV uptake so far today was what was needed to recharge the system from it's own power consumption (parasitic load). We stopped uptake yesterday around 6 pm our time. It's now 11 am. Pretty big numbers required to bring the bank back up to that 100% mark, I'd say...

 

[MurphyGuy]

No problem, just make sure you get the BMS records from those batteries and then disconnect everything from them so the only parasitic loads are the batteries themselves.
24 hours of letting them sit should do it. Record battery bank voltage every hour, or every 2 hours. Just so long as it is consistent time intervals. I usually set an alarm on my phone when I do that kind of stuff so I don't forget.

 

[SHTF Power Anyway]

Can you log into your BMS interface and post a screen shot of you cell voltages on one of your 5.4kW packs? I don't need every pack, just pick one at random.
You are basing you 5% calculations on the assumption that each of your packs is 5.4kWh in capacity and that may not be accurate. While the pack might have that much capacity, full capacity and usable capacity are usually 10% to 20% different. Charging a large battery pack up to full voltage and draining it to empty is harmful to the battery's life expectancy, and while it can be done, most manufacturers set BMS limits to chop off a little at the top and a little at the bottom. I can see the voltage specs using the Fortress Power spec sheet, but nothing on their website indicates if your packs are 15s or 16s configs.

For now, lets assume 16s config and use your numbers from your 5% SOC drop. Assuming your 10 packs are 5.4kwh each, that's 54000 watts.. x 0.05 = 2700watts-hours of capacity loss (2.7kWh). So 2700 divided by 24 hours = 112.5 watts per hour, or a bit over 10 watts per pack.. Assuming the BMS is doing some balancing, that's more reasonable to expect.



Yes.. those numbers are generally accurate no matter what inverter is being used.

It kind of depends. I'm not sure who's reporting SOC information in your case but I suspect your Sol Arks have nothing to do with it other than acting like a user interface for the battery BMS's. In other words, the batteries are reporting SOC and the Sol Arks are just displaying the information for you. (not entirely sure on that one). It is possible that the inverters are attempting to report their own calculated SOC based on your input parameters but I suspect that isn't the case.

I was going to attach the CSV file but it's not an allowable format.

 

[SHTF Power Anyway]

No problem, just make sure you get the BMS records from those batteries and then disconnect everything from them so the only parasitic loads are the batteries themselves.
24 hours of letting them sit should do it. Record battery bank voltage every hour, or every 2 hours. Just so long as it is consistent time intervals. I usually set an alarm on my phone when I do that kind of stuff so I don't forget.

I have a record of every battery with what I think are pretty critical details and it appears to record data every single second. It hit 100% SOC per the Sol-Arks AND the BMS software which I was connected to at the time it hit 100%. However, the csv file with the records never showed that point that it hit 100%. And I can't find a way to post it to get that data to you so you can see what I see. The files are far too wide for a pdf to do much of anything. They're a csv format. I really think you'd find that data useful. Suggestions?

 

[TorC]

You could just rename the .csv to .txt and the site will take it. That's basically all a CSV is, except the contents are supposed to follow a format. IIRC, a zip file would also work.

 

[MurphyGuy]

I have a record of every battery with what I think are pretty critical details and it appears to record data every single second. It hit 100% SOC per the Sol-Arks AND the BMS software which I was connected to at the time it hit 100%. However, the csv file with the records never showed that point that it hit 100%. And I can't find a way to post it to get that data to you so you can see what I see. The files are far too wide for a pdf to do much of anything. They're a csv format. I really think you'd find that data useful. Suggestions?

Please ignore anything coming from the inverters.. just BMS data please.

What is the battery pack voltage
Please just copy and paste the cell level voltage from maybe one or two random packs.

Make sure inverters are DISCONNECTED from batteries as soon as they are charged to 100%. Unbolt one of the battery cables if you have to.

 

[SHTF Power Anyway]

Please ignore anything coming from the inverters.. just BMS data please.

What is the battery pack voltage
Please just copy and paste the cell level voltage from maybe one or two random packs.

Make sure inverters are DISCONNECTED from batteries as soon as they are charged to 100%. Unbolt one of the battery cables if you have to.

My apologies for the significant delay in response here. I know how hard you're trying to help me. I had a call from a good, disabled friend in East Liverpool, OH and it took 2 hours for her husband to translate everything she had to say. So.... back to the solar.

Here's the file renamed as a txt. We took readings of the following at the time it hit 100% SOC per the BMS software:

With a Fluke T5-1000: 55 volts
the measurement was taken at the connection of the battery cables to the Sol-Ark

With a Fluke 789 Process Meter: 54.83 DC which dropped quickly to 54.79 DC within seconds.
Again, the measurement was taken at the connection of the battery cables to the Sol-Ark

The attached file (which I finally found) was generated at the time we took the above readings. We have not taken any readings since because I didn't get the instructions to do so due to that dang phone call. But I will start taking the readings at set intervals going forward and will be that info posted tomorrow morning. My husband is out confirming that there is no draw from the inverters at all. Both are off but, if there is a draw, we'll let the batteries recharge to 100% again and will start the analysis over.

 

[SHTF Power Anyway]

My apologies for the significant delay in response here. I know how hard you're trying to help me. I had a call from a good, disabled friend in East Liverpool, OH and it took 2 hours for her husband to translate everything she had to say. So.... back to the solar.

Here's the file renamed as a txt. We took readings of the following at the time it hit 100% SOC per the BMS software:

With a Fluke T5-1000: 55 volts
the measurement was taken at the connection of the battery cables to the Sol-Ark

With a Fluke 789 Process Meter: 54.83 DC which dropped quickly to 54.79 DC within seconds.
Again, the measurement was taken at the connection of the battery cables to the Sol-Ark

The attached file (which I finally found) was generated at the time we took the above readings. We have not taken any readings since because I didn't get the instructions to do so due to that dang phone call. But I will start taking the readings at set intervals going forward and will be that info posted tomorrow morning. My husband is out confirming that there is no draw from the inverters at all. Both are off but, if there is a draw, we'll let the batteries recharge to 100% again and will start the analysis over.

I better clarify that the Sol-Arks were powered off when the measurements were taken.

 

[MurphyGuy]

I better clarify that the Sol-Arks were powered off when the measurements were taken.

I'm honestly not sure if powered off is good enough. My television has a parasitic draw even when off.. 3 watts or something like that. But we can roll with it.. Even if the inverters have a small draw when off its not going to be 30 watts x 10 modules.

Just looked at your text file and converted it over to something readable. I'm posting a sample screen shot of the cell balance voltages so everyone else can see it as well. Couldn't capture everything, but the sample is typical of the rest of the readings. Biggest differential, after a quick look, seems to be 20 mV, which isn't bad at all.

 

[MurphyGuy]

Now we wait. Post some updates on battery bank voltage if you can.. maybe later tonight or in the morning. Time + Voltage.. You need to post time and voltage every time you give take a reading. If the parasitic draw is bad, we'll know in 12 hours.. But if the parasitic draw is what it should be, and what I think it's going to be, then it will take a full 24 to 48 hours to even begin to see it via battery pack voltage readings.

All lithium batteries, in fact all batteries period, exhibit a behavior called "hysteresis". When under discharge, the battery voltage will read lower than it actually is, and when being charged, the voltage will be higher than it actually is. Hysteresis increases with current, so the bigger the draw, or the faster the charging, the more hysteresis you will experience.

Under draw, some people call the discharge hysteresis "Voltage sag", and with a lead acid battery, it can be huge.. not so much with a lithium battery, but it still shows up when we're looking at voltage readings with two or three digits below the decimal. And the same is true for charging hysteresis.. You might charge the battery up to 55.32 volts and come back in an hour after it's been sitting and find the voltage has sunk to 55.30 or 55.29. Different lithium configurations experience different levels of hysteresis, but it is always there...

To get accurate battery voltage from a lithium battery, it must sit without being charged or discharged for about 30 minutes to an hour.. Actually, its more like 24 hours, but the bulk of the hysteresis effects will flatten out after about 30 minutes and for our purposes, that's good enough.

The reason the hysteresis takes time to "flatten out" is because even after you remove the load, or stop charging, the ions inside the battery are still moving around looking for a comfortable place to settle down, put their fuzzy slippers on, and watch TV.

When I said "Get a good meter", you kind of went overboard with that 789 eh? Cranky, I wish I had one of those.

Your Fluke 1000 is not appropriate for this work.. Keep using the 789

 

[SHTF Power Anyway]

Now we wait. Post some updates on battery bank voltage if you can.. maybe later tonight or in the morning. Time + Voltage.. You need to post time and voltage every time you give take a reading. If the parasitic draw is bad, we'll know in 12 hours.. But if the parasitic draw is what it should be, and what I think it's going to be, then it will take a full 24 to 48 hours to even begin to see it via battery pack voltage readings.

All lithium batteries, in fact all batteries period, exhibit a behavior called "hysteresis". When under discharge, the battery voltage will read lower than it actually is, and when being charged, the voltage will be higher than it actually is. Hysteresis increases with current, so the bigger the draw, or the faster the charging, the more hysteresis you will experience.

Under draw, some people call the discharge hysteresis "Voltage sag", and with a lead acid battery, it can be huge.. not so much with a lithium battery, but it still shows up when we're looking at voltage readings with two or three digits below the decimal. And the same is true for charging hysteresis.. You might charge the battery up to 55.32 volts and come back in an hour after it's been sitting and find the voltage has sunk to 55.30 or 55.29. Different lithium configurations experience different levels of hysteresis, but it is always there...

To get accurate battery voltage from a lithium battery, it must sit without being charged or discharged for about 30 minutes to an hour.. Actually, its more like 24 hours, but the bulk of the hysteresis effects will flatten out after about 30 minutes and for our purposes, that's good enough.

The reason the hysteresis takes time to "flatten out" is because even after you remove the load, or stop charging, the ions inside the battery are still moving around looking for a comfortable place to settle down, put their fuzzy slippers on, and watch TV.

When I said "Get a good meter", you kind of went overboard with that 789 eh? Cranky, I wish I had one of those.

Your Fluke 1000 is not appropriate for this work.. Keep using the 789

MURPHY! I THINK I FOUND IT!!!!!!!

Take a look at the files for batteries 3, 6 & 8. Make note of the unit current. Then look at battery 5. I had my Pop with me (good 82 year old neighbor who was an electrical engineer). I was showing him the BMS software screens and noted that every battery in our string was showing a unit current of -0.3 except #5 which was showing a 0.5 (positive, not negative like the others). I said, "OMG, Pop, is battery 5 bad?" He said he felt we'd had a bad battery this whole time!

If this is the case, I'm beyond angry because I kept trying to get help from Fortress only to get no answers, straw men, and bad answers and an insult along the way... $35k investment in their products. I cannot tell you how mad I am.

My husband had that Fluke 789 for work. Like I said, he was an electrician for over 40 years. He has lots of equipment and what he doesn't have, my neighbor does. It's just that I don't know how to use it myself and my poor hubby is struggling these days with simple instructions. :-(

We'll continue with the experiment but I wanted you to know I had Pop come over and work with Kevin to eliminate the very small draw the Sol-Ark was still taking though powered off. That adjustment was made right between 3 & 4 pm PST.

The voltage readings so far have been as follows:
2:43 pm 53.27
3:43 pm 53.25
4:43 pm 53.24
5:43 pm 53.23

The draw down is certainly a lot slower than the initial drop we were seeing.

 

[SHTF Power Anyway]

You could just rename the .csv to .txt and the site will take it. That's basically all a CSV is, except the contents are supposed to follow a format. IIRC, a zip file would also work.

Thank you for the tip! I used to be smart and tech savvy! I'm embarrassed at how far I've fallen from grace! lol

 

[SHTF Power Anyway]

I'm honestly not sure if powered off is good enough. My television has a parasitic draw even when off.. 3 watts or something like that. But we can roll with it.. Even if the inverters have a small draw when off its not going to be 30 watts x 10 modules.

Just looked at your text file and converted it over to something readable. I'm posting a sample screen shot of the cell balance voltages so everyone else can see it as well. Couldn't capture everything, but the sample is typical of the rest of the readings. Biggest differential, after a quick look, seems to be 20 mV, which isn't bad at all.

I thought my "team" got the Sol-Arks completely off. The Sol-Arks are both powered off and the breakers where the batteries connect are also off. The current reads 0.2 on the master and we momentarily saw 0.9 on the slave. It then read zero when we tested a few more times. Again, both are powered off and have the battery breakers off. Battery 5 now shows, like the rest, a -.03 unit current on the BMS software. I have left the laptop connected to the BMS and it should be recording so we have a much longer period we can look at to see if battery 5 is the problem. I have really faulty laptops at this point to work with. The one they loaded the software on for the BMS has bad ports. I tried loading the software on other laptops I have available and cannot get the connection to work for some reason. I have to use ethernet as I have pretty nasty reactions to emf these days so I can't just monitor with wifi. That problem is just icing on today's cake! lol

 

[SHTF Power Anyway]

Missed our exactly hourly time by a few but here's the next reading:

6:52 pm 53.22 volts

 

[RV10flyer]

I have 10 diy batteries or 143kWh, 2X SA12K, JK BMS with LCD’s on all winter. I use a low current 100A clamp on meter to test idle loads and direction of flow on everything on my home, shop and van systems. Especially solar components. When I purchase something, the idle and working current gets put on a label. You will be surprised to find similar junk electronics idle consumption can vary drastically between different brands. My system idle consumption is about 4kWh/day. I produce 5-130kWh/day. The meter would have found your problem in a few minutes.

 

[SHTF Power Anyway]

I have 10 diy batteries or 143kWh, 2X SA12K, JK BMS with LCD’s on all winter. I use a low current 100A clamp on meter to test idle loads and direction of flow on everything on my home, shop and van systems. Especially solar components. When I purchase something, the idle and working current gets put on a label. You will be surprised to find similar junk electronics idle consumption can vary drastically between different brands. My system idle consumption is about 4kWh/day. I produce 5-130kWh/day. The meter would have found your problem in a few minutes.

He's using a Fluke 36 Clamp Meter on the lowest setting (200A)

 

[SHTF Power Anyway]

He's using a Fluke 36 Clamp Meter on the lowest setting (200A)

How much PV do you have? Where are you??? I want all that production! Wow!

We did build some DIY batteries with EVE cells but they're for a 12 volt system. We were going to 100% DIY the batteries but chickened out due to hubby's memory issues.

 

[SHTF Power Anyway]

Now we wait. Post some updates on battery bank voltage if you can.. maybe later tonight or in the morning. Time + Voltage.. You need to post time and voltage every time you give take a reading. If the parasitic draw is bad, we'll know in 12 hours.. But if the parasitic draw is what it should be, and what I think it's going to be, then it will take a full 24 to 48 hours to even begin to see it via battery pack voltage readings.

All lithium batteries, in fact all batteries period, exhibit a behavior called "hysteresis". When under discharge, the battery voltage will read lower than it actually is, and when being charged, the voltage will be higher than it actually is. Hysteresis increases with current, so the bigger the draw, or the faster the charging, the more hysteresis you will experience.

Under draw, some people call the discharge hysteresis "Voltage sag", and with a lead acid battery, it can be huge.. not so much with a lithium battery, but it still shows up when we're looking at voltage readings with two or three digits below the decimal. And the same is true for charging hysteresis.. You might charge the battery up to 55.32 volts and come back in an hour after it's been sitting and find the voltage has sunk to 55.30 or 55.29. Different lithium configurations experience different levels of hysteresis, but it is always there...

To get accurate battery voltage from a lithium battery, it must sit without being charged or discharged for about 30 minutes to an hour.. Actually, its more like 24 hours, but the bulk of the hysteresis effects will flatten out after about 30 minutes and for our purposes, that's good enough.

The reason the hysteresis takes time to "flatten out" is because even after you remove the load, or stop charging, the ions inside the battery are still moving around looking for a comfortable place to settle down, put their fuzzy slippers on, and watch TV.

When I said "Get a good meter", you kind of went overboard with that 789 eh? Cranky, I wish I had one of those.

Your Fluke 1000 is not appropriate for this work.. Keep using the 789