midday output to zero

[dochubert]

Hi,
I have a 10 year old grid tie system in san diego area; Xantrex gt3.0 with 10 kyocera 190 watt panels. When I first installed it, it was kicking out 2100w at peak sun, sometimes more. After break in it was a good 1800w and a little more on good clear days after some rain. pretty normal i'm thinking. Problem is, that output started dropping. And it just kept dropping.
At a point that would put the system at about 7 to 8 yrs old, I emerged from my fog to realize that i was only getting about 50% or around 900-1000 watts instead of 1800-1900.

I thought maybe the connectors were at fault. These are 2005 panels which had mc3 connectors. When the system stops producing around mid day and output drops to zero watts, I would go up on the roof and 'exercise' the mc3 connections between panels. Usually then on coming down and checking, I would have output again. Eventually, I got an mc4 crimper and a set of mc4 connectors and changed them all out. No difference that I can tell. So that wasn't it.

About a year or so ago I got a deal on 4 235w panels, so i added them to the series string. Got my output back up to 2000 or so watts (not counting the drop out times that still happen)
The drop outs seem to most reliably happen on the hottest sunniest days, and usually from noon or so to as late as 3pm. The GT3.0 inverter says "Insufficient Solar Energy". The output reads 0w but the array voltage is still there. Just no watts. This to me says bad connection or open circuit in a panel

Now I realize it could possibly be the inverter with the problem rather than solar panels, which is one reason I'm sharing this problem here. Thoughts and ideas appreciated.
I lean towards thinking the GT3 is working correctly and there are 1 or more intermittent opens in 1 or more solar panels.
I ordered a clamp on DC ammeter, which I'm hoping will help corner the problem(s) when it arrives.

At some point, I am going to have to probably replace those ten panels. But Hey! Don't I have a 20 year warranty? If one or more panels is only putting out half power, I would think the warranty would cover that. I bought the ten panels and the GT3 from Arizona Wind and Sun, so at least its somebody still in business. I cringe to think I paid $800-825 each for those panels. How many 250w panels could I buy with more than $8000 today?

Anyway, while i wait for that ammeter, maybe someone here will point out something i missed.

Thanks for listening
I'm new here so hope i'm in the right place

[sensij]

What voltage are you getting when you are measuring it?

Same error reported in this thread:

http://www.solarpaneltalk.com/showthread.php?18476-What-is-min-PV-Power-%28-Watts%29-that-is-needed-fire-up-a-Xantrex-GT-3-0

I found in the manual where it discusses that error, but it isn't clear what electrical condition triggers it.

[dochubert]

I'm reading normal open circuit voltage which is about 400-425 when its saying insufficient solar energy. My gt3.0 requires a minimum 190 (or maybe its 195) volts to operate. The original ten panels was easily above that and with the added 4 panels the minimum wouldn't be a problem. But an intermittent open (or high enough resistance) possibly triggered by a threshhold high temperature would drop input volts to below minimum, causing the insufficient solar energy error state and zero output. At least thats my current theory.
What do you think?

[sensij]

There are other errors that should be showing up indicating low voltage or pv array disconnected. If you go through the fault log, is anything else there?

One trouble shooting step might be to disconnect the inverter and short the array (connect + to -). Watch the Isc with your ammeter. If it still drops out, you know for sure there is something wrong with a panel.

[Mike90250]

I'm thinking high heat is causing a panel or 2 to go bad. You may have a warranty claim.

To test, open the circuit, and isolate each panel, and try to measure Isc for each panel. You can get good Voc from the array, but a bad panel will have high resistance and as soon as the inverter loads the array down, the voltage crashes.

Use a sacrificial shorting cable with your Ampmeter, MC connectors are NOT designed to quench an arc, and if you open or close them while energized, you will fry the connectors. Uh oh, you replaced the factory connectors on the PV cables, you likely voided the panel warranty

[sensij]

I'm not so sure that a module failing under high heat would qualify for the performance warranty, if it generates the right amount of power under the lab test conditions. That seems more like a defect that was covered only by the 1 year materials/workmanship warranty.

[dochubert]

I had not thought about shorting the whole string to see what i get there. Might try that. I did read some other threads with similar issues, including one about testing panels.
I'm thinking about making a solar panel tester out of an 8 ohm dump resistor i have and a voltmeter wired up to some leads with mc4 connectors to load test each panel on a hot day. (and a blanket to shade the panel while hooking/unhooking test connections--system off of course) I'm figuring I will get my test stuff ready, wait for a drop out so I know conditions are right for it, and then shut down the system and start testing panels.

I don't know how to access the fault log in my GT3.0. When I tap the panel it cycles thru the various readings, but if it stores errors, it doesn't show any by tapping. I haven't looked at the manual for awhile, but don't remember anything in it about a fault log or how to access it. Doesn't mean its not there, just don't remember seeing anything.
When the drop out occurs, the insufficient solar energy reading pops up and after awhile it apparently clears (the panel cools maybe?) and the watt readings start climbing back up to 'normal', but once it clears, i can't see any errors.

Yeah, I probably screwed myself out of any warranty claims by changing those connectors out. Any excuse to not back a product these days, but i'm trying not to pre-judge them as I haven't given them a fair chance yet. Still have to prove a fault - and back to the load test...

Thanks for the comments and suggestions. I appreciate it

[inetdog]

I'm thinking about making a solar panel tester out of an 8 ohm dump resistor i have and a voltmeter wired up to some leads with mc4 connectors to load test each panel on a hot day. (and a blanket to shade the panel while hooking/unhooking test connections--system off of course) I'm figuring I will get my test stuff ready, wait for a drop out so I know conditions are right for it, and then shut down the system and start testing panels.

I hope that 8 ohm dump resistor is very high power.
If you end up putting 10A through it, that would be 80 volts (close enough to zero for an Isc test) and 800 watts, at least for a short time.

[solarix]

I had to replace a whole array of 16 Kyocera panels one time that failed due to bad solder connections inside the panel. They were made in 2001 and Kyocera claimed it was just a bad batch and happily replaced them under warranty (screwed me on getting paid for doing the work though). I would not be surprised if that Xantrex inverter was the problem either. Probably out of warranty though.

[dochubert]

Well, the resistor is 600w. I've never seen 10 amps from those panels on their best day so surprised to read that someone thinks i could get 10. I was expecting maybe 7a at most from a panel still working properly. The short circuit current rating is 7.96a. I can double the wattage on my load just in case but wasn't expecting more than 300w max so figured I was safe with 600w
Vmp 26.1 and Voc is 32.5
Am I wrong?

[inetdog]

Well, the resistor is 600w. I've never seen 10 amps from those panels on their best day so surprised to read that someone thinks i could get 10. I was expecting maybe 7a at most from a panel still working properly. The short circuit current rating is 7.96a. I can double the wattage on my load just in case but wasn't expecting more than 300w max so figured I was safe with 600w
Vmp 26.1 and Voc is 32.5
Am I wrong?

For crystaline panels a Vmp which is roughly 80% of Voc is normal over a wide range of insolation. What does the panel label say (or can't you get at them to read it?)

[dochubert]

I apologize if I'm missing something here, but I seem to be missing something here.

I don't see why a 600 watt resistor won't do to test a 187 watt panel. I plan to test each panel one at a time with my dump load resistor, reading voltage across the load and current with a clamp on dc ammeter that i don't have yet. I will wait for a drop out to happen as the trigger to start my testing. The system will be then turned off and I'll use a blanket to shade the test panel while connecting/disconnecting the load. The bad part is that it will be a cooking hot day to be up on that roof screwing around with mc4 connectors, but those are the typical days I get drop outs.
The voltage should be the telling point with this test. I hope and expect at least some panels to give good readings under individual test loading. Those that differ significantly likely are pulling the rest of the system down. Some hot days, like today, I got no drop outs at all unless they timed them between my spot checks. If by some magic the panels all test out good I then go back to being stumped. Guess I will have to consider the xantrex as culprit at that point

Hopefully I described my plan more coherently.
If not then I am missing something and will just wait for a light to dawn (or something...)

[bcroe]

I apologize if I'm missing something here, but I seem to be missing something here.

I don't see why a 600 watt resistor won't do to test a 187 watt panel. I plan to test each panel one at a time with my dump load resistor, reading voltage across the load and current with a clamp on dc ammeter that i don't have yet. I will wait for a drop out to happen as the trigger to start my testing. The system will be then turned off and I'll use a blanket to shade the test panel while connecting/disconnecting the load. The bad part is that it will be a cooking hot day to be up on that roof screwing around with mc4 connectors, but those are the typical days I get drop outs.
The voltage should be the telling point with this test. I hope and expect at least some panels to give good readings under individual test loading. Those that differ significantly likely are pulling the rest of the system down. Some hot days, like today, I got no drop outs at all unless they timed them between my spot checks. If by some magic the panels all test out good I then go back to being stumped. Guess I will have to consider the xantrex as culprit at that point

If your panels are delivering power, its the inverter. If they are not, the voltage will nose dive. I would
connect my test light to the panel output. it is 4 each 75 watt bulbs in series, well insulated. It will
draw several hundred watts, to prove if there is enough power to run the inverter or not. They should
be brighter if the inverter shuts down. I would expect the inverter to go into a startup cycle after
dropout, eventually turning on and dimming the lights.

If the lights go out, check the PV current. If maxed, there is a short. If zero, the array is failing to deliver.

Once you establish inverter or array, I would not waste time trying to test every panel. Insert a 3 way MC4
at the string mid point and bring a wire down to a voltmeter. Running you should see half the array voltage
from the tap to either end. When it quits, you may see a big change in those voltages. If one of them has
dropped greatly, or even reversed polarity, there is a fault in that half limiting output. You can divide that
section again with another 3 way, and eventually move them in to the exact location of the fault. Do pay
attention to polarity, as the bad section may have the right voltage at the wrong polarity. Bruce Roe

[sensij]

Hopefully I described my plan more coherently.
If not then I am missing something and will just wait for a light to dawn (or something...)

I would really suggest doing full array Isc testing before going through the work to test each indvidual panel.

[dochubert]

First, thanks to everyone who commented and/or made suggestions. I have learned a lot just reading your comments.
I still haven't received my dc ammeter so not really ready to go, but am currently leaning toward adding the splitter mid array as a voltage tap. Voltage readings using the tap during a dropout should clear up most of the questions while narrowing down the faulty panel.
Narrowing the dropout fault down will speed things up considerably.
Besides, I can always go back to a panel by panel test. I suspect more than one bad panel, which means eliminating the drop outs will be great, but I won't be done.
Definitely a big step in the right direction, tho