Grid tied solar array has voltage but no current

[lindq530]

I have a roof mounted grid tied system: 20 each Sharp ND-L3E1U panels (123W panels connected in series) and a Sunny Boy SWR2500U inverter. It's worked flawlessly for the last 15 years.

A couple of weeks ago I noticed that the the inverter display indicated that the array was producing 395V and 0 watts (I use the old fashion monitoring system of looking at the inverter display once a week or so). The voltage was consistent with historic values, but obviously the watts was not. Everything appeared to be working (the inverter wasn't showing any error codes) - except that I wasn't getting any power.

I tried "rebooting" the system by turning off the AC and DC inputs. That didn't changing anything. I was fortunate to get an identical SB2500 inverter a few years ago form a working system that was being upgraded - I replaced the inverter and the results were the same - about 395 volts and no watts. I tested the continuity of the wires from AC connection to the house panel and the inverter to the roof - the wires are OK. I checked the line voltage - its OK (244V (122V each leg) and 60 Hz).

I visually inspected the panels and wiring connections and did not see any obvious issues.

The last thing I've tried was disconnecting each panel from the array and testing Voc and Isc using my multimeter. Every panel's output voltage was close to the the manufacturer's spec of 21.3V. Current output varied from 3-6A for each panel (I tried to measure the output in direct sun, but that was hard due to time constraints - spec is 8.1A). I have not tested each panel under load (because I'm trying to avoid that until its absolutely necessary).

I've read on this and other forums that I might have a panel that has failed. If so, the failed panel is letting a small amount of current pass so I get voltage, but not enough current to pass to produce measurable watts.

Any suggestions about what is causing the issue?

If it is a bad panel, any suggestions on how to identify the bad panel without dismantling and load testing each panel?

Thanks in advance for your comments and suggestions.

[littleharbor]

Can you get a good close up visual on the panels? If there is a bad one you might be able to see something out of the ordinary. Look for small discolored areas , scorch marks on the cells and all visible tabbing and buss ribbon areas. Sometimes the burnt area is very small which would be all it takes to show voltage but be incapable of carrying current.Of course you might find something much more obvious like a shattered panel.

[sensij]

Do those panels have bypass diodes? If so, a cell failure should be bypassed, and not knock out the whole string. Have you tried measuring Isc for the string, and not just the individual panels?

[lindq530]

Can you get a good close up visual on the panels? If there is a bad one you might be able to see something out of the ordinary. Look for small discolored areas , scorch marks on the cells and all visible tabbing and buss ribbon areas. Sometimes the burnt area is very small which would be all it takes to show voltage but be incapable of carrying current.Of course you might find something much more obvious like a shattered panel.

I visually inspected the panels and couldn't find anything obvious (like discolored areas, corroded busses, or cracked glass).

But, that's a great suggestion to double check. It's much easier than disassembling the array.

[lindq530]

Do those panels have bypass diodes? If so, a cell failure should be bypassed, and not knock out the whole string. Have you tried measuring Isc for the string, and not just the individual panels?

I presume there are diodes, but the junction boxes are potted so I don't have access to them (if they do indeed exist).

I did not check the Isc of the string. I will do that tomorrow morning and report what I find.

[sensij]

I presume there are diodes, but the junction boxes are potted so I don't have access to them (if they do indeed exist).

I did not check the Isc of the string. I will do that tomorrow morning and report what I find.

Please don't make or break the Isc connection while the sun is out. At full string voltage, that is a much more serious arc.

[Mike90250]

be careful about disconnecting MC connectors with power applied. there should be no current flowing, the inverter and disconnects turned off to make it safe to disconnect

[lindq530]

Please don't make or break the Isc connection while the sun is out. At full string voltage, that is a much more serious arc.

Thanks for the safety lesson.

Is the proper way to measure the full string Isc:
Connect the multimeter when it's dark outside, e.g. 5AM.
Leave it connected all day and note the Isc when the array is producing max current, e.g. full sun on the panels.
Disconnect the leads after the sun sets, e.g. 8PM.

[bcroe]

Is the proper way to measure the full string Isc:
Connect the multimeter when it's dark outside, e.g. 5AM.
Leave it connected all day and note the Isc when the array is producing max current, e.g. full sun
on the panels. Disconnect the leads after the sun sets, e.g. 8PM.

A lot of work has been done here on panel wiring under the lights, for safety and avoiding damage.

I have a couple ways to deal with measuring current. A quick way is to use a DC clamp on meter.
A more permanent and accurate way is to wire a shunt into the wiring, then connect your DVM to
it as needed. Bruce Roe

[lindq530]

Update:

I measured the Isc of the total array (18 panels are still on the roof - I disconnected two panels to provide safe access) - measured about 6.5 amps with my clamp on meter. Voc was about 330V. That matched panel specs.

I finally resorted to the hardest option yet - bypassing individual panels. I thought I'd find a bad panel - when the bad panel was removed from the array I'd get watts. I used a jumper cable to bypass each panel individually. Each time I would bypass a panel, I would restart the system (it took about 4 hours). Each time the system voltage would be about 18 volts lower (as expected because of the array missing a panel) and still zero watts. It appears its not a bad panel.

When I measure the voltage of the array leads (with the inverter connected and on) I get:
Pos-neg: 330V
Pos - ground: 330V
Neg - ground: 0.0V
It doesn't appear I have a ground fault.

Everything points to the inverter being bad. I'm having trouble accepting that because one of the first things I did was replace the inverter. It seems very unlikely that two separate inverters would have the same issue.

I'm out of ideas.

HELP!

[inetdog]

Any chance that the grid tie inverter is not drawing any power from the panel because of a problem on the AC side, such as not being connected or not recognizing the AC as valid within its operating range? Wrong Hz setting?
Similar to a charger not drawing current from the array because it cannot send current to the battery.

[lindq530]

Any chance that the grid tie inverter is not drawing any power from the panel because of a problem on the AC side, such as not being connected or not recognizing the AC as valid within its operating range? Wrong Hz setting?
Similar to a charger not drawing current from the array because it cannot send current to the battery.

I checked the line voltage and frequency - both are within specs for the inverter (measured 244V (122V each leg) and 60.5 Hz). The Hz is on the edge of the inverter's specs (60.5 Hz)).

The inverter has an error code that should be displayed if the AC side is the problem.

I measured the current in the AC lines when the inverter is operating - it's 0.48A - every time I measure it. It's almost zero when the inverter is idle.

[lindq530]

I had a terrible thought this evening - what if my replacement inverter actually has the same issue as my existing? It's a very low probability, but I can't dismiss the possibility. That would explain why all my testing thus far hasn't identified a problem.

Does anyone have a suggestion about how to have my inverter professionally tested?

[sensij]

Pos-neg: 330V
Pos - ground: 330V
Neg - ground: 0.0V
It doesn't appear I have a ground fault.

In the first post, you said you had 395 V. Where did the other 65 V go?

Is the 0.48 A you measure in operation in the direction of consumption or production?

[lindq530]

In the first post, you said you had 395 V. Where did the other 65 V go?

Is the 0.48 A you measure in operation in the direction of consumption or production?

The voltage is reduced because I had to remove 2 panels to safely get access to the array on the roof and one panel was bypassed when I took the measurement (bypassed so I could test each individual panel for failure).

I'm not sure of the direction. I'll have to use my clamp on meter to determine if it's being consumed or produced by the inverter.