Sudden battery voltage drop, only when epever connected

[DaniSe]

Measure voltage drop across components and wires while current is high. Set meter to dc voltage and one test lead on each end of breaker for example. I had a breaker once brand new had several volts across it.

With the connections now tight, there is no loss across the breaker and the BMS low voltage disconnect issue has solved.

Now it is back to working out why the controller disconnects under higher rates of charge - I am sitting here monitoring the system today: The inverter gave two alarm beeps in succession and after the second I saw the MT-50 give the over volt disconnect though I didn't catch the voltage readout. As I am writing this, it cut again - no OVD on the MT-50 but the battery voltage read 27.3 (happened again while writing, at 27.1) and on the PV shutoff the voltage dropped to 26.0

I think tonight I am going to bring the array down to one panel to limit the charge while I sort out what the issue is.

Since I have seen so many varying opinions on the topic I was wondering if someone could recommend charge parameters for LiFePO4 that are geared towards optimal battery health. At the moment I am trying to run load at around equal to the charging amperage to keep any overcharging from happening - is that logic sound?

 

[acdoctor]

If I were you the very next thing I would do is turn off SCC and use a meter to check / monitor solar input voltage at the SCC terminals.

 

[DaniSe]

If I were you the very next thing I would do is turn off SCC and use a meter to check / monitor solar input voltage at the SCC terminals.

Could you walk me through this? By turning off, do you mean disconnecting from the battery (I understood this to be something to avoid)?

I ended up covering the array and disconnecting PV entirely for the moment - I am really trying to not damage anything.

 

[acdoctor]

I have a breaker between panels and SCC. I would turn it off and through the day check that voltage on the input side of the breaker. I have Schneider equipment. My panels are wired 3S, VOC is 47 SCC is 150 max VOC. On cold clear morning's I get a DC over volt warning. Usually when I see it the over volt is over but the warning remains. My SCC’s just shut down during the fault. My panels are solarworld 320 watt. Either you are having higher VOC than SCC can handle or SCC is not outputting correct voltage. After proving VOC is ok check SCC output but be sure to use a meter and not what the SCC says that could be flawed.

 

[DaniSe]

Your thinking is sound.
Either the load is too much for the batteries. Or, something is wrong with the batteries or connections.

I braved opening the batteries up today to see if there is any obvious damage. I -think- the lead cell from the positive terminal is possibly bulged? Images attached, I have not taken the cells out individually to inspect yet this is a first look.

There also seems to be some light colored... something between the first and second cell. There is also a scorch mark on the terminal of this battery as well.

Should I further disassemble the potentially faulty battery? Will that void any chance I have at a warranty if the cell was faulty from the start? We have been trying to sort the higher voltage disconnect since day one, we checked SOC to be same before connecting in series and today when measured they both read at 13.19

 

[timselectric]

I wouldn't disassemble any further. And try to get a warranty replacement. It appears that this one slipped through quality control. In general, the build quality looks good on all. I would like to think that they would want this one back.

 

[acdoctor]

Use extreme caution not to short leads to the case! These aren’t lead acid. No I wouldn’t disassemble any farther.
What is strange is no matter how much I zoom in I don’t see the “strange white substance “ on whole battery picture. That gap is the only one that looks to be completely closed. Strange white substance looks a little like aluminum corrosion. I do see more gap and strange yellowish between others?
Can you visual inspect the back of the main pos. Terminal for more or less signs of over heating. It could be all on your connection.
Use your meter to test cell voltages. You can do it without disconnecting anything. Can you charge with something other than SCC ? If so do that while checking individual cell voltages, particularly cell in question. Depending some on the charge current,cell voltage shouldn’t change much between say 20 and 80%.
If you have a cell that causes the bms to go into charge protection mode, depending on the bms design it could cause you to see the DC over voltage discussed above.

 

[acdoctor]

Reading the entire thread I think you have a cell that’s capacity is significantly lower than the others. And is opening bms on high cell protection during charge and low cell protection with loads. In addition to checking cells as stated above I would run loads and check individual cell voltages.

 

[HRTKD]

The spacing between the cells is not uniform. I don't know if it was that way from the start or not.

 

[DaniSe]

Use extreme caution not to short leads to the case! These aren’t lead acid. No I wouldn’t disassemble any farther.

Noted, will use extreme caution whenever the top is open.

What is strange is no matter how much I zoom in I don’t see the “strange white substance “ on whole battery picture. That gap is the only one that looks to be completely closed.

Somehow some mislabeled images that I meant to replace ended up in here - there were two extra images of the good battery with both labels. Removed all extra images, should only be four now!

The white substance is partway down and can only be seen if the light and camera are at the exact correct angle. In the photo from a distance, the shadow obscures visibility.

Strange white substance looks a little like aluminum corrosion. I do see more gap and strange yellowish between others?

I will look again, that was the only place I noticed it.

Can you visual inspect the back of the main pos. Terminal for more or less signs of over heating. It could be all on your connection.

Will do.

Use your meter to test cell voltages. You can do it without disconnecting anything. Can you charge with something other than SCC ?

I didn't realize that, I will check them. I only have the SCC to charge however.

If so do that while checking individual cell voltages, particularly cell in question. Depending some on the charge current,cell voltage shouldn’t change much between say 20 and 80%.
If you have a cell that causes the bms to go into charge protection mode, depending on the bms design it could cause you to see the DC over voltage discussed above.

As you say, out of balance cells would seem to explain a lot of my system symptoms. Thank you again for all this help, I would be at a loss trying to diagnose this on my own.

 

[DaniSe]

The spacing between the cells is not uniform. I don't know if it was that way from the start or not.

There are foam spacers between all but the first two - I only know this after watching an SOK battery review video that included disassembly.

 

[DaniSe]

@acdoctor After testing the voltages on the cells :

-The battery without possible visible imperfections read 3.32 across the board

-The suspect battery read 3.32 on the two farthest from the + lead and 3.32 or 3.33 on the two closest, it flickered between the two as I held it.

No visible terminal damage on the underside, and while trying to get a better look down between - there may actually be a foam insert there. I suppose it is possible the white substance could be adhesive but I don't know why one would have the insert and not the other.

Spent the last hour doing one more test charging with the SCC. Only charging one battery at a time. First one panel (40.1 Voc / 10.36 Isc), then when all seemed well progressing to two in series. No disconnects. To note, charging both batteries in series on even one panel resulted in disconnect yesterday.

-PV Voltage readings in match the MT-50 at the SCC terminals (75.5V)
-Voltage out on MT-50 reads 14.1V
-Voltage out at SCC terminals reads 14.05V
-Voltage at battery terminals reads 13.91V
-There is a .05V loss over the 40amp breaker (is this within tolerance?)

I am so turned around trying to find the actual issue here. Is the controller not wanting to auto detect voltage for the parameters after the initial shutdown, and any consequent shutdowns another possible clue? If the controller is the fault, what might cause the controller to suddenly leak enough voltage at night to result in the BMS engaging LVD?

 

[acdoctor]

Cell voltages look great at that state of charge. If that was at rest could be anywhere between 50 and 80%. I wouldn’t charge separately but parallel or you will get them out of balance. 0.1 volt can be a lot of capacity difference.
Have you reconnected all of the panels? Can you fully charge both in parallel?
Can you monitor cell voltage in the top 10%
Doing this will ensure batteries are equal. When the sun goes down if you have a DC amp meter check to see if there is any current from batteries to scc .
Disconnect panels from SCC and batteries from SCC and reconnect batteries to 24 volts and connect inverter and use it through the night.
This process should weed out a lot of variables.

 

[acdoctor]

How can MT-50 see panel voltage I thought it was your shunt/ battery monitor? Where is the 40 amp breaker is it between the panels and SCC?

 

[DaniSe]

How can MT-50 see panel voltage I thought it was your shunt/ battery monitor? Where is the 40 amp breaker is it between the panels and SCC?

The MT-50 is connected to the SCC and displays incoming PV current on far left, and charging current in center, and load on right. I have a separate shunt for the battery monitor.

I checked the cell voltages after bringing both through a short absorption at 14.00 individually (Your reply came after one had already been charging). All cells reading 3.34, except for one reading 3.32 in what I am now referring to as the suspect battery (middle cell on the negative side). Is that difference significant enough to cause my problems?

And is that difference enough to not combine the batteries in parallel at this point in order to charge together?

 

[timselectric]

Wouldn't hurt to go a day without that one connected, if possible. It may provide you with the information you are looking for.

 

[acdoctor]

No not 1 cell I would parallel and charge to SOK specs or at least 14.2 to 14.4. What have you been using with old SCC?

 

[acdoctor]

You can’t really tell what the cells are doing unless you are measuring in the upper knee of charge while still charging. Once charging ends they will fall back to at least 3.4 volts.

 

[DaniSe]

Have you reconnected all of the panels? Can you fully charge both in parallel

With the charging amps rating of max 50, I can only charge the one battery at 12V on two panels I believe. It was getting 45 odd amps in with the two in series. The batteries in parallel would be able to handle 100 amps charging current, is that correct? But in series still only 50?

Can you monitor cell voltage in the top 10%
Doing this will ensure batteries are equal.

I don't believe I would be able to with what I have (my only measuring tool at present is the multimeter and the software in the SCC) - what is the usual process to monitor the cells while charging?

When the sun goes down if you have a DC amp meter check to see if there is any current from batteries to scc .

Don't have an amp meter, is there one you would recommend? I feel like I should own one after all this.

Disconnect panels from SCC and batteries from SCC and reconnect batteries to 24 volts and connect inverter and use it through the night.
This process should weed out a lot of variables.

Thanks for these, I will keep sleuthing to narrow it down.

 

[acdoctor]

I don't believe I would be able to with what I have (my only measuring tool at present is the multimeter and the software in the SCC) - what is the usual process to monitor the cells while charging

I meant with the meter at the individual cells.