X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Originally posted by ericy View Post
    I turned off the inverter and went away for 10 minutes or so. When I came back, it only saw 19V. I have 39 panels, so I would have expected 39V. Not sure what this implies - broken string?
    Solaredge in residential is normally one panel (module) to one optimizer.
    So you should have 39 optimizers and with 39 of them, I would expect a string of 19 and a string of 20.
    So if you didn't have them tied together inside the inverter, I'd expect to see one of them with 19V, the other with 20V. But the +'s and -'s are connected together inside your inverter, so I would expect you would see either 19V or 20V (and I don't know which)

    I would expect to see 4 wires coming into the inverter from the roof. (A pair of wires for each string.) But it is possible to combine the two strings before they come down from the roof.

    Comment


    • #17
      Originally posted by Ben25 View Post
      The safety voltage will only be an indicator of the longest string in series, not total optimizers. There are probably 2 strings connected in parallel.
      Good point.

      I opened the inverter today - there was indeed some amount of condensation fogging the window for the LCD display. And here was some amount of dried scale in the bottom of the case around where the seal for the front door is supposed to seal, but aside from the fog on the window, there really wasn't much other moisture that I could see.

      I opened up the DC disconnect box - it was far cleaner - no real sign of scale in there.

      I opened one of the conduit LL boxes where the conduit/wires go into the crawlspace - there was dried scale in there as well, but it was currently dry. I did not see any wire nuts in there.

      It appears that the string that is down really consists of 2 substrings. One of the substrings is on the side of the house where the working string is. The other substring is on the same side of the house as the inverter - thus to wire that substring, a conduit goes up from the DC disconnect box into the attic. The two substrings are essentially wired together in series in the DC disconnect box to form the 2nd string (that is down).

      I ran out of my allotted time for the day (and it is supposed to rain later today), so I brushed out the scale, and did my best to clean up the seal, buttoned everything back up again, and turned everything back on again. There was no real change in how things were working, but I didn't really expect an improvement. Well, the LCD display is now readable - that definitely improved.

      The next time I mess with this thing, I am going to try and connect the string/substrings one at a time and see what the isolation values look like. This would also let me clearly identify which optimizers/panels are on which string - something that is currently in doubt. I suspect one of the substrings is in good shape, and I could probably bypass the bad substring and get another 7 panels up and running.

      It isn't obvious how the DC wires are locked into the DC disconnect - I need to figure out the proper way to release those from the terminal block before I make my next attempt.

      Comment


      • #18
        Hi ericy.

        Perhaps you've already found this document, but if not:

        (www).solaredge.com/sites/default/files/application_note_isolation_fault_troubleshooting.p df

        Be safe! Or as we say in our robotics lab, "Don't get dead!" (-: Thought I think Mike Rowe said it best, "Safety Third!"

        Seriously, be very careful.

        -Jonathan

        Comment


        • #19
          Originally posted by ericy View Post

          It appears that the string that is down really consists of 2 substrings. One of the substrings is on the side of the house where the working string is. The other substring is on the same side of the house as the inverter - thus to wire that substring, a conduit goes up from the DC disconnect box into the attic. The two substrings are essentially wired together in series in the DC disconnect box to form the 2nd string (that is down).
          If they are wired in series then it is ONE string not two sub strings. There is no requirement that the modules be near each other for a string, just that all the modules are wired in series.


          Originally posted by ericy View Post
          The next time I mess with this thing, I am going to try and connect the string/substrings one at a time and see what the isolation values look like. This would also let me clearly identify which optimizers/panels are on which string - something that is currently in doubt. I suspect one of the substrings is in good shape, and I could probably bypass the bad substring and get another 7 panels up and running.

          It isn't obvious how the DC wires are locked into the DC disconnect - I need to figure out the proper way to release those from the terminal block before I make my next attempt.
          keep in mind there are minimum string lengths. Minimum recommended is 8 PV modules on the older systems, where generally 6 is absolute minimum on the older ones (the string will not be able to get to operating voltage with less).
          OutBack FP1 w/ CS6P-250P http://bit.ly/1Sg5VNH

          Comment


          • #20
            Originally posted by JSchnee21 View Post
            Hi ericy.

            Perhaps you've already found this document, but if not:

            (www).solaredge.com/sites/default/files/application_note_isolation_fault_troubleshooting.p df

            Be safe! Or as we say in our robotics lab, "Don't get dead!" (-: Thought I think Mike Rowe said it best, "Safety Third!"

            Seriously, be very careful.

            -Jonathan
            Yeah, I read that thing a couple of days ago. The first thing I did when I started this was to read up on how to make sure things are in a safe state before I even undo the screws for the cover.

            If you have the right equipment, they make it sound like a straightforward process to find these faults. But I don't have the right kind of insulation test meter for that job, so that much is academic.

            Comment


            • #21
              Originally posted by ButchDeal View Post

              If they are wired in series then it is ONE string not two sub strings. There is no requirement that the modules be near each other for a string, just that all the modules are wired in series.




              keep in mind there are minimum string lengths. Minimum recommended is 8 PV modules on the older systems, where generally 6 is absolute minimum on the older ones (the string will not be able to get to operating voltage with less).
              I wasn't quite sure what the correct terminology is, so I used "substring" to simply mean a portion of the string that isn't physically adjacent to the rest of the string.

              I recall reading about the minimum string lengths now that you mention it. Sigh.

              Comment


              • #22
                Minor update. Got a hold of the installer - found out that in the winter they lay off most of their staff, so they only have a skeleton crew. I don't know how it works when they staff up again in the spring. Got a hold of the CEO of the install company - he talks a good game on the phone - said he would come out today (weather permitting). I am not optimistic that he will show up. He was talking about installing a cable to bypass the non-reporting optimizers, which I suppose might be a starting point for tracking down the problem.

                I still need this company for a couple of things. I am still awaiting the SREC to transfer (bought the house on 9/28/18) - that seems to be stuck in the bureaucracy up at the Delaware PSC, but someone at the installer company is walking me through the process. And I want to get an installer account of my own (I saw the sub-installer thread here), which would let me fully administer the site myself. For example, if I ever needed to replace an optimizer, I would want to be able to make the change on the monitoring site myself and not depend on this guy to do it for me. I think I might start out on Monday by contacting SolarEdge to make sure the info in the other thread is still current. But I will probably need help/permission from the installer to make this happen.

                I am more than happy to hire someone to work on the system. All I ask is that they return phone calls, show up when they are supposed to, know what they are doing, and not try and BS me.

                Comment


                • #23
                  I need to go out and buy a lottery ticket. Repeated badgering of the installer finally paid off. He showed up today - found a bad connector on one of the strings. I no longer have non-reporting optimizers (I had given him the ID of the non-reporting optimizers, which let him focus on that part of the array).

                  I believe the system is *probably* working better - I will keep an eye on it for a while before I close the book on this issue. If I look at the production for last week, I would typically see a peak in the neighbhorhood of 2-3kW (at 2:15PM, typical values are around 2.0kW). The system was off for about 2 hours today - when it came back on (at 2:15PM), it was generating 5kW. And the last reported status of all of the optimizers shows that none of them were in SafeDC mode.

                  I don't have historical data for Jan - the previous owner would go to Florida, and disconnect the Comcast, so no monitoring history. But I do have data for Nov of 2015 (when the system was operating correctly) - in late Nov I would see production of right around 5kW at that time of day, with a peak of about 7kW at about 11:45AM.

                  I believe I still have a bit of a ground fault - not enough at the moment to cause an error, but the ground resistance fluctuates a bit and currently is at about 2000kOhm. For a single-phase inverter, it ought to report a fault if the number drops to 600kOhm.

                  And I still want the installer to grant me more access to the thing so I have better visibility as to errors and whatnot, but he can do that from his office.

                  Comment


                  • #24
                    Just to bring everyone uptodate - in addition to the bad connector, we had a bad optimizer (harder than it should have been since his original map was screwed up). I don't know if there are shortcuts to fixing a map, or whether one needs to take down the panels one-by-one and check the ID of each optimizer. But that didn't involve me needing to do anything, and he eventually found the faulty optimizer.

                    Just this week he replaced the bad one with a new optimizer, and I am again back to 100% with 39 panels/optimizers online.

                    He also granted me more access - I now have "Charts", "Reports" and "Alerts" tabs. Not the same as an installer account, but a huge improvement nonetheless.

                    I still have occasional ground faults which come up during heavy rainstorms. That's the last issue that remains for me, but it only comes up once every couple of months, and when it does happen, the condition only lasts for a few hours. That one could be a pain to track down, but if I see it happen, I might turn the system off and check the junction boxes to see if there is water in there. I can do at least that much.

                    Comment


                    • #25
                      Originally posted by ericy
                      I still have occasional ground faults which come up during heavy rainstorms. That's the last issue that remains for me, but it only comes up once every couple of months, and when it does happen, the condition only lasts for a few hours. That one could be a pain to track down, but if I see it happen, I might turn the system off and check the junction boxes to see if there is water in there. I can do at least that much.
                      You should not have ground problems with your array, I would take that as a sign of
                      sloppy wiring. Problems might be obvious, moisture getting in. Current flows with sun.

                      I once opened the 2 pole disconnect switch, and measured extremely low leakage of
                      the system to ground. One way to do this with the switch open is put a light bulb
                      between ground and the + or the - lead. Measure the voltage across the bulb in either
                      setup, should be near zero. If none found, reduce bulb to night light size and try again.
                      Finally I got it down to a 1000 ohm resistor, which reads 1mv for each micro amp of
                      current. Only a handful of microamps flowed in my rather extensive ground array.
                      Bruce Roe
                      Last edited by bcroe; 02-23-2019, 01:09 PM.

                      Comment

                      Working...
                      X