Intermittent Ground Fault with Classics on PV

Started by malkierian, April 14, 2022, 05:37:53 PM

Previous topic - Next topic

malkierian

For background once again, I have 6 Classics networked together. 5 are charging from PV, one from wind. I have GFP turned off on the wind Classic, and the GFP jumper installed only in the Master Classic (for FollowMe, where the rest are all followers). I've been having this problem for a couple months now, where I would get a ground fault in the system, with seemingly no correlation between instances. Sometimes it would happen multiple times in a day, sometimes it would go for days without a problem. Every time I notice it, I would go out and turn the classics off and back on and the fault would be gone. Because of the intermittent nature of it, I'm having trouble tracking it down. Just today, I managed to get out when it had just happened (I think), and I could move the jumper back and forth between different classics, after resetting the fault without the jumper in them, and have it set off again immediately. However, while doing this, none of the PV sources were connected (positive breakers were open. Negatives were still connected). A few minutes later, after another power cycle of all classics, the fault was gone, with nothing else done on my end. Where can I go to start tracing this problem?

Also, in the manual, it mentions that battery negative and ground shouldn't be bonded in the system except through the jumper, but the manual for my inverters says that battery negative should be grounded. I know I must be missing something between the two, perhaps some clarification could be given here? (for reference, I'm talking about the second to last bullet point on page 6 of the the manual rev 1933)
1st System: 12.5kw on 5 Classics (3 150s, 2 250s), 1 classic 250 with wind turbine, 2 stacked Magnum 4024PAE inverters.

2nd System: 1250W solar on 1 Classic 150, two DIY3024 running separate breaker boxes, one hooked up to Magnum output for AC charging.

boB

Hi malkierian...

How the Classic's internal GND Fault works is,  when you have the GF jumper installed, it connects a 0.5 amp resettable fuse between the common negative and the chassis GND.   If anything should cause that fuse  (PTC)  to open, the Classic detects that and trips a fault.

One thing that can cause this, and this is how we test the circuit in house, is to have a positive voltage (like PV+) touch the Chassis GND connection which will open the PTC...   That is, if there is over 1/2 amp of + current possible to trip the PTC fuse.

If you were to bond the battery negative to GND somewhere, that would short out the PTC fuse and no fault would be detected.

Thanks for the note about the manual, page 6.  I will let the others know and we can put that into a list of changes, if it isn't there already.

One thing that might cause intermittent ground faults is that jumper, which might be making a bad connection... It has to make a good connection, otherwise the internal logic that tries to pull a logic HIGH (a GND fault), will not be held LOW by the good connection that the jumper and PTC and traces provide.   The connection between negative and GND.

So, that could be one possibility.   That is why we use gold covered connectors and jumpers

One thing you might try, IF you are willing to live without one of your 6 controllers for a short time to experiment, is to just turn off the PV to your Classic with the GFI and see if it faults.  Turning off the PV+ will remove the possibility of a real GND fault there...  If it still faults, then it is more than likely just a bad connection with the jumper or something like that.

If it IS that Classic with a bad jumper connection, you could try another of the 6 Classics to see if that one works better.  Disable the first GFI Classic and enable the 2nd GFI Classic and make sure it has its jumper on.

Please keep us posted.

boB


K7IQ 🌛  He/She/Me

malkierian

#2
Alright, so a pretty massive update. I had another day where it was tripping over and over again, so I was able to do some further diagnosis. I moved the jumper between classics and was still getting the fault, even with the previous ones turned off. So I did the only other logical thing. I started disconnecting wires one at a time to try to isolate the issue. It finally got to a point where I had a couple different controllers completely unplugged, including ground, except for the battery+, ethernet and the three communication ports (front panel and follow-me ports). One time, I forgot that I had done that and tried to turn one of those controllers on, and it actually turned on. Unplugged the communication cables and it turned off. At first I thought there were neutral connections going through there, but surely it was running from battery to ground using the ground in the comm ports, right?

After that, I completely un-networked all of the classics (including removing ethernet cables) and set them with sane internal measurement cutoffs for charging, put the GFP jumpers back on all of them, and let the system run for a few days, and when I went out to check, only one was actually faulted. So, the only think I could think of, is that when there was a fault on that one controller's ports, whether it was on or off, that fault was passed along via the comm ports because of those ground connections, thus creating faults in the rest of the controllers. That's the best I could come up with, and it still doesn't make any sense. Any thoughts?
1st System: 12.5kw on 5 Classics (3 150s, 2 250s), 1 classic 250 with wind turbine, 2 stacked Magnum 4024PAE inverters.

2nd System: 1250W solar on 1 Classic 150, two DIY3024 running separate breaker boxes, one hooked up to Magnum output for AC charging.

boB


Wow, good job on trying to troubleshoot.

I am still a bit confused about what could be happening ?  The Ethernet connections are isolated BTW.  In fact, we don't even use the metal surround on the jack on the Classic anymore.  So unless the isolated magnetics in the Ethernet jack has broken through somehow, that should not be the problem.

Need to ponder a bit.  Any others on this forum are welcome to ponder with me if wanted.   :)

boB
K7IQ 🌛  He/She/Me

malkierian

I wasn't talking the ethernet ports when I was mentioning what would allow them to power on, I was talking the master in/slave out ports under the display port. That line of three RJ12 ports. Also, it didn't matter which classic was networked to the one that was completely disconnected, as long as its battery negative was connected, it would allow the other with battery negative to power on.

So like I said on the last one, only one Classic was registering as faulting, so I switched the PV source from that one to one of the other ones I think is still working to see if it was the PV array or the Classic itself. Did that this morning. So far no faults, but it can take days for one to show up, so we'll have to wait and see for now.
1st System: 12.5kw on 5 Classics (3 150s, 2 250s), 1 classic 250 with wind turbine, 2 stacked Magnum 4024PAE inverters.

2nd System: 1250W solar on 1 Classic 150, two DIY3024 running separate breaker boxes, one hooked up to Magnum output for AC charging.

boB


Thanks for the clarification.

Yes, you certainly need those battery negatives connected together on the Classics.

Those GND connections on the master/slave jacks will try to take any current if there was any high current to be had and might not be so good for the cable etc...

Keep us posted.

K7IQ 🌛  He/She/Me

malkierian

Well, I'm happy to say this problem is resolved, and let me tell you, it was a journey. Let's start from the beginning.

About a year and a half ago, we got a wind turbine that we hooked up to our system. About the same time, my dad reworked the wiring in the power shed before insulating the walls and putting wall and ceiling panels in. Keep that aside for the moment. At the same time, about a year ago, we got a set of batteries we could hook up as a secondary power source (we have two DIY inverters hooked up to that collection, as well as a singular Classic that is using power from 4 combined arrays, from a combiner box that is right there in the shed (3 of the 4 arrays were already wired into there on an Outback controller before I added one more array and hooked them all up). Before I got the batteries hooked up properly (custom copper bars connecting the necessary series and parallels), that charge controller was also hooked up to our main battery bank. My dad had associated the start of the problems with the installation of the wind system. That turned out to be a red herring that took me a while to work past.

Now, two separate things happened that probably wouldn't be critical if they had happened separately. First, there was apparently some confusion as to what wires were the negative wires for two of the four arrays going to the secondary bank. So, when my dad did the rewiring, he took those two negative wires and connected them to ground, thinking they were grounds. Like I said, not super bad, those two arrays just wouldn't work. Or so you would think. Because of that rewiring my dad did, there were still a couple wires over by the combiner box that I could reasonably assume were the negative wires for those two arrays. So, OK, I hooked them up. One was a dead wire, made no connections. However, the last one wasn't a dead wire, wasn't a negative. It was ground. It was connected to ground right next to the charge controller that was having the most troubles with ground faults. Again, it wasn't an issue, necessarily, because that array just wouldn't work. And so, the kicker is that those two arrays were actually still providing power to the secondary battery charger. Through ground. Through the connection right above that charge controller. Which would explain why it was so intermittent, because it would require some kind of special circumstance in the combiner box array to trigger something in the main array.

So now I've hooked everything back up (I had to switch the controllers between the main one on the main bank and the one on the secondary bank to try to diagnose faulty hardware, since that former main controller was 5.5 years old). Ethernet, FollowMe, all arrays, all controllers, even the wind system. Only finished it up this morning, but so far everything's back to peak operation, no ground faults. As my dad said when I told him, it looks like we found the smoking gun, and I agree with him, so I'm confident the problem is actually resolved, and after a couple of days with the front plates still off, I'll put everything back to normal and finally be able to completely clean up in there.
1st System: 12.5kw on 5 Classics (3 150s, 2 250s), 1 classic 250 with wind turbine, 2 stacked Magnum 4024PAE inverters.

2nd System: 1250W solar on 1 Classic 150, two DIY3024 running separate breaker boxes, one hooked up to Magnum output for AC charging.

boB

Kind of weird but I am glad you found that connection !

Intermittent problems are the hardest !   Sometimes I have to beat on things to get them to either act up or behave !

boB
K7IQ 🌛  He/She/Me