Odd voltage behavior (huge drop) with replacement Sunny Boy inverter... ???

[Chris_NH]

Hello all,
I've got a 6 year old solar system comprised of 2 banks of panels. One bank of a dozen 245w 24v panels in series at about 250v and one bank of 14 100w 12v panels in series at 269v. Originally, both fed into a Magnatek Aurora inverter and all worked flawlessly for years. The Aurora finally died last year so I got a great deal on a new Sunny Boy 3000, but since it couldn't handle the 2 banks of separate voltages like the Aurora could I hooked up the bigger bank of 24v panels to it and everything there is great. I recently bought a used Sunny Boy 1800U to again utilize the bank of 12v panels, but am running into a strange situation.

The voltage for the 12v panels in series is 269v like it always has been, but when I hook it up to the Sunny Boy 1800U the voltage drops to 2.3v and the 1800U obviously won't operate. I disconnect the wires from the 1800U and the voltage reads 269v again. I initially thought the 1800U was DOA, but I hooked up the other bank of panels and it functioned. It gave me a GFDI fuse error, so it appears to need a new GDFI fuse, but other than that it was functioning.

Anyone have any idea why hooking up that one bank of panels to the 1800U would result in the massive voltage drop?

I feel like I'm missing something obvious, but this knucklehead obviously need some guidance on this.

Many thanks,
Chris

[Mike90250]

I wonder if something went wrong with the string of panels. Can you connect them to the working inverter ? 260V seems low for a GTI. are you sure it's not undervoltage ?

[Chris_NH]

I can try connecting them to the working inverter to see what happens.

The operating range is 140v - 400v, so the 260v can't be the issue.

[bcroe]

I would try swapping panels and inverters to see which component is common to all failing
arrangements. The inverter could be shorted. Or the panels may have a fault, a high
resistance that passes voltage but practically no current under load.

I use this test light, just some series 75W bulbs, to make sure the panels have some
current capacity. good luck, Bruce Roe

PVtestLt.JPG

[Sunking]

Sounds like a short circuit to me. Bet if you used a current meter, you would see something close to Isc from the panels at close to 0 volts. When open circuit your are seeing Voc.

[Mike90250]

In my opinion, you should use a voltage stabilizer for best use, you can use inverter[1] to use or use voltage stabilizer also.

gee, can you sketch up how it would be wired and explain how it would help the original poster ? Since we are dealing with lethal voltages here, all testing must be safe.

[jflorey2]

The voltage for the 12v panels in series is 269v like it always has been, but when I hook it up to the Sunny Boy 1800U the voltage drops to 2.3v and the 1800U obviously won't operate. I disconnect the wires from the 1800U and the voltage reads 269v again. I initially thought the 1800U was DOA, but I hooked up the other bank of panels and it functioned. It gave me a GFDI fuse error, so it appears to need a new GDFI fuse, but other than that it was functioning.

Anyone have any idea why hooking up that one bank of panels to the 1800U would result in the massive voltage drop?

You have a bad (high resistance) connection somewhere. This is the opposite of a short. The high resistance allows your meter (which presents almost no load to the array) to read the correct open circuit voltage, but pull any significant current at all and the voltage drops to close to zero.

To debug:

Get some sort of load (like Bruce made) and apply it. In your case it could be two light bulbs in series. Attach it. Now connect one side of your meter to the negative return from the array, and start moving the other lead to each panel junction in turn, starting from the "bottom" of the array. You will see a "normal" rise in voltage (i.e. about 18 volts per panel) until you get to the problem connection, where you will see a large and dramatic change in voltage. You will probably need very long leads for this.

Needless to say, employ good safety practices as you are doing this - 250VDC is a dangerous voltage to be working with.

[Chris_NH]

You have a bad (high resistance) connection somewhere. This is the opposite of a short. The high resistance allows your meter (which presents almost no load to the array) to read the correct open circuit voltage, but pull any significant current at all and the voltage drops to close to zero.

To debug:

Get some sort of load (like Bruce made) and apply it. In your case it could be two light bulbs in series. Attach it. Now connect one side of your meter to the negative return from the array, and start moving the other lead to each panel junction in turn, starting from the "bottom" of the array. You will see a "normal" rise in voltage (i.e. about 18 volts per panel) until you get to the problem connection, where you will see a large and dramatic change in voltage. You will probably need very long leads for this.

Needless to say, employ good safety practices as you are doing this - 250VDC is a dangerous voltage to be working with.

I was thinking of waiting for the next overcast day, or first thin in the morning, and getting up in the roof with a working small 12v Chinese GTI on a long extension cord and hooking up one panel at a time until I found one or more that wasn't putting anything in to the grid. Even overcast with almost no sun each panel aught to put 5 or 10 watts into the grid. I have extra 100w panels and could replace any outliers.

Would this accomplish same thing?

Thanks for all the input.
Chris

[jflorey2]

I was thinking of waiting for the next overcast day, or first thin in the morning, and getting up in the roof with a working small 12v Chinese GTI on a long extension cord and hooking up one panel at a time until I found one or more that wasn't putting anything in to the grid. Even overcast with almost no sun each panel aught to put 5 or 10 watts into the grid. I have extra 100w panels and could replace any outliers.

Well, you'd still need a meter to measure the output of the inverter, and some of them take a while to start up - so it could take a while. If you are going to go panel to panel a 12V light bulb might be a better choice - even a watt or so would give you some visible light.

You should also be prepared for the problem to be _between_ two panels, at the connection (assuming you are using MC4 connectors or similar.) So that approach might not find the problem.

[Chris_NH]

Well, you'd still need a meter to measure the output of the inverter, and some of them take a while to start up - so it could take a while. If you are going to go panel to panel a 12V light bulb might be a better choice - even a watt or so would give you some visible light.

You should also be prepared for the problem to be _between_ two panels, at the connection (assuming you are using MC4 connectors or similar.) So that approach might not find the problem.

OK, I do have a 12 volt light bulb, so will rig something up with that. Hopefully it's something simple at one of the MC4 connectors and not a panel issue.