GXMnow
Solar Wizard
- Joined
- Jul 17, 2020
- Messages
- 2,740
I had some time to do some garage work today. I had the XW do a little bit of grid charging as another storm is moving in. With just enough sun out to run the house on the Enphase system, and the XW sitting in "No Float", I figured it was a good time to check on battery health. Remember, I am using Lithium Ion NMC cells from a Chevy Bolt EV. The clouds did their odd thing again where I got some extra reflected light on the panels a few times. The peak power on the Victron system actually hit 1,900 watts out of the 2,000 watts of panels, but it was only for a couple minutes at 1:04 PM. The Enphase system only shows the average of each 15 minutes, so it is only showing a maximum of 2,500 watts out of 4,800 watts of panels. I got a total of 9.4 KWHs out of the Enphase system (1.958 sun hours), and 4.01 KWHs out of the Victron (2.005 sun hours). So once again, the DC system does better when the production sucks.
Back to the battery testing from earlier...
It seems either the JK-BMS voltage calibration drifts with the temp, or it just plain drifts. The measured voltage at the cells at very little current was off by 0.2 volts again. And the Victron battery voltage reading matched my Fluke meter. So I did the voltage cal in the JK and now it is within better than 0.05 volts. My Fluke is only a 4 digit, so I could not go any further. The XW still shows higher than the Victron, but the JK-BMS reading is closer now. I verified a few of the cell voltages in the old battery case with the JK-BMS and after the cal, the ones I can get to easily were all within 0.002 volts of the reading in the iPhone app. At less than 4.1 volts I get the third decimal. So I see up to 4.099 volts, but the next step goes to 04.10 volts. So I trust the JK readings now, especially from cell to cell. And it is reporting the lowest to highest cell, at absorb voltage were within 0.003 volts from lowest to highest at 4.043 to 4.046 volts.
Now it was time to check the two strings in the new battery cabinet. I can't easily get to the actual cell terminals, so the best way I have to check the cells is from the BMS balance connectors. I have to turn off the battery disconnect, then unplug the connector from the BMS and measure each adjacent wire pair. The meter leads end up very close to touching to probe those pins. I need to come up with a better way. I managed it without shorting a cell, but you have to be slow and careful. These are on the dumb Daly BMS units, so they don't have any data reporting, and they only have a 60 milliamp passive balance function. When a cell is above about 3.85 volts, and more than 0.010 volts higher than any other cell, it turns on the 0.060 drain on that high cell. I measured the upper bank first and was quite surprised by the result. The lowest cell was #3 at 4.040 volts and 3 other cells were all the same highest voltage of 4.046 volts. So just 0.006 volts from the lowest to highest cell. That means the balancer is not even turning on to do anything. It only works at greater than 10 mv.
The lower rear bank is a bit more difficult to measure as the balance leads are going further down, so it does not have as much slack to pull it out to get the meter leads on it. It took me a bit longer to get all the cell readings, but I did it again with no shorts. The balance was not quite as good, but still not a problem at all. The highest cell was just 1 mv higher than the upper string at 4.047 volts. Several cells were at 4.045 and a few were down at 4.041 volts. But there was one straggler down at 4.038 volts. That makes the spread from highest to lowest at still only 0.009 volts (9 millivolts). My guess is that the balancer did have to pull it in and stopped trying when the difference went under 0.010 volts. So that is where it is at.
Unlike LFP cells, there is not a sharp voltage knee at the top. There is a subtle knee where the voltage des rise a little faster, but it is down around 60% state of charge. The slower rising voltage seems to give the small balance current enough time to keep a weak cell from running and hitting the high voltage cut off. On the low side, the weakest cell will go below the knee first, then the voltage drop slows a little and the others will catch up a bit before they are all below that knee. And then the voltage of the whole battery falls slower until 10% left, and then they do fall off a steep knee like LFP cells. I have never run mine down low enough to see it actually happen. The XW stops inverting at about 40% remaining.
The Victron was still pushing some charge current, so when I had the 2 new strings disconnected, the old string got the full charge current of maybe 4 amps. I only had the disconnects open for about 15 minutes, so the voltage didn't rise much with out dismal sun, but it was enough that when I reconnected a new string, the old JK-BMS charging current fell to under 1 amp and the new string took all of the charge power. The difference in voltage was not enough to cause the old battery to discharge into the new battery, because the Victron charge controller was pushing enough current that it only took a couple minutes for the new batteries to get up to the same voltage as the old strings and the current was very evenly shared once again.
We have rain and clouds for 2 more days. I will be doing a small grid charge each night until we get some sun again on Wed. Last night it charged for 8.5 hours, but at just 7 amps. At an average voltage, just under 55 volts, that is less than 3.3 KWHs of battery charging from the grid. But it also causes all the house load to come from the grid overnight as well. So it was charging at 4 amps instead of discharging at 17 amps. Running the house was another 7.8 KWHs of grid power last night. Oh well..
Back to the battery testing from earlier...
It seems either the JK-BMS voltage calibration drifts with the temp, or it just plain drifts. The measured voltage at the cells at very little current was off by 0.2 volts again. And the Victron battery voltage reading matched my Fluke meter. So I did the voltage cal in the JK and now it is within better than 0.05 volts. My Fluke is only a 4 digit, so I could not go any further. The XW still shows higher than the Victron, but the JK-BMS reading is closer now. I verified a few of the cell voltages in the old battery case with the JK-BMS and after the cal, the ones I can get to easily were all within 0.002 volts of the reading in the iPhone app. At less than 4.1 volts I get the third decimal. So I see up to 4.099 volts, but the next step goes to 04.10 volts. So I trust the JK readings now, especially from cell to cell. And it is reporting the lowest to highest cell, at absorb voltage were within 0.003 volts from lowest to highest at 4.043 to 4.046 volts.
Now it was time to check the two strings in the new battery cabinet. I can't easily get to the actual cell terminals, so the best way I have to check the cells is from the BMS balance connectors. I have to turn off the battery disconnect, then unplug the connector from the BMS and measure each adjacent wire pair. The meter leads end up very close to touching to probe those pins. I need to come up with a better way. I managed it without shorting a cell, but you have to be slow and careful. These are on the dumb Daly BMS units, so they don't have any data reporting, and they only have a 60 milliamp passive balance function. When a cell is above about 3.85 volts, and more than 0.010 volts higher than any other cell, it turns on the 0.060 drain on that high cell. I measured the upper bank first and was quite surprised by the result. The lowest cell was #3 at 4.040 volts and 3 other cells were all the same highest voltage of 4.046 volts. So just 0.006 volts from the lowest to highest cell. That means the balancer is not even turning on to do anything. It only works at greater than 10 mv.
The lower rear bank is a bit more difficult to measure as the balance leads are going further down, so it does not have as much slack to pull it out to get the meter leads on it. It took me a bit longer to get all the cell readings, but I did it again with no shorts. The balance was not quite as good, but still not a problem at all. The highest cell was just 1 mv higher than the upper string at 4.047 volts. Several cells were at 4.045 and a few were down at 4.041 volts. But there was one straggler down at 4.038 volts. That makes the spread from highest to lowest at still only 0.009 volts (9 millivolts). My guess is that the balancer did have to pull it in and stopped trying when the difference went under 0.010 volts. So that is where it is at.
Unlike LFP cells, there is not a sharp voltage knee at the top. There is a subtle knee where the voltage des rise a little faster, but it is down around 60% state of charge. The slower rising voltage seems to give the small balance current enough time to keep a weak cell from running and hitting the high voltage cut off. On the low side, the weakest cell will go below the knee first, then the voltage drop slows a little and the others will catch up a bit before they are all below that knee. And then the voltage of the whole battery falls slower until 10% left, and then they do fall off a steep knee like LFP cells. I have never run mine down low enough to see it actually happen. The XW stops inverting at about 40% remaining.
The Victron was still pushing some charge current, so when I had the 2 new strings disconnected, the old string got the full charge current of maybe 4 amps. I only had the disconnects open for about 15 minutes, so the voltage didn't rise much with out dismal sun, but it was enough that when I reconnected a new string, the old JK-BMS charging current fell to under 1 amp and the new string took all of the charge power. The difference in voltage was not enough to cause the old battery to discharge into the new battery, because the Victron charge controller was pushing enough current that it only took a couple minutes for the new batteries to get up to the same voltage as the old strings and the current was very evenly shared once again.
We have rain and clouds for 2 more days. I will be doing a small grid charge each night until we get some sun again on Wed. Last night it charged for 8.5 hours, but at just 7 amps. At an average voltage, just under 55 volts, that is less than 3.3 KWHs of battery charging from the grid. But it also causes all the house load to come from the grid overnight as well. So it was charging at 4 amps instead of discharging at 17 amps. Running the house was another 7.8 KWHs of grid power last night. Oh well..
