Grid Tied Grief!

[sunnyguy]

According to the inverter the voltage seems good at 385-390 and you are correct that the house system is fine.

Definitely a head scratcher. So string voltage and grid AC voltage are roughly the same for the house system and the ground systems? Panels/combiners/inverters are the same? Only difference is feed in transformer? Has anyone looked at the AC waveforms on a scope?

It doesn't seem like it is output limited if the output scales linearly with # strings connected. It is hard to believe anything could current limit the strings without melting/burning in the process. So that suggests an mppt problem but only on 4 of 6 inverters with identical settings? Strange.

[J.P.M.]

[QUOTE=gtuplin;n325104]

I just went through this thread again and note that there is no mention of array orientation(s) (tilt & azimuth), any shading or estimates of performance. OP: Since we're all throwing darts, any chance things have some weird set of orientations ? Also, were you given, or did you do anything in the way of performance estimates ? Also, any way to estimate insolation levels during all this underperformance ?

Maybe I'm stupid, or going blind, or both, but I seem to be missing some basic information here. Knowing array orientations would allow a dart throw estimate of clear day performance. That might be helpful in the analysis. Not meant as an insult, but do you know that array orientation (tilt & azimuth) and shading has a bearing on performance ? Or that system output is variable ?

Good call, you are right I never provided that info.
The entire array is situated almost perfect due south and the angle was set at 44 degrees as a permanent angle and was calculated for best February to October outputs based on our location. The winter was not a concern as it is primarily light rain cloud and short days. There is a slight shade now in the early morning but only effects a few panels then clear until sun down.
I can only compare this system to the running system and it is pretty much the same orientation.

Thank you. Have you checked PVOutput for similar orientations in your area ? FWIW, 44 deg. tilt is probably closer to year round optimum at your lat. With as little winter sun as you get, I might have gone a bit lower on the tilt, but I wasn't the designer.

[gtuplin]

Oh boy is it a head scratcher but something completely by accident occurred that may help to resolve this. A bit cool again this AM but clear and readings seem the same as last few record keeping days so I decided that it was a good morning to clean modules while I hang out so away I go with a bucket of soap water and my squeegee. I cleaned one complete set went to check readings and things were about the same as yesterday but no surprise the cleaned panels were performing slightly better(never underestimate clean panels) so back to the next row and so on. Part way through the water was dirty enough to change so back to the inverters for data readings then off to the house for fresh water but for some reason I dumped the water on the ground under the cabinets that holds all the electrical components and it ended up pretty close to a direct shot at the ground cable to the large ground plate. Holy mother of God all hell breaks loose! all 4 inverters crash from about 2800w down to practical shut down , stall at 400-800w then as fast as that occurred back on line and major jump to the highest numbers I had ever recorded (still not expected output however) of 4100w-4500w! I stood and watched in amazement for some time and it continued at these number for awhile then as time progressed it began a very slow decline.

Just theory at this point but the system is built on very dry rocky terrain, perhaps the grounds all be them everywhere are not working! we have wet dew in AM so that might account for better morning reading and as the day goes on the dirt dries and removes what little ground ability it has, is this possible? I have a friend who is an electrical engineer and conversed with him (wish he lived closer) and he has asked me to check for AC current between the DC positive at inverter and ground to see if there is AC being picked up by the grounding system from the overhead high power lines. Seems like a plan so will do that soon.
Does any of this make sense?

[jflorey2]

According to the inverter the voltage seems good at 385-390 and you are correct that the house system is fine.

OK, good. 390 volts is 32.5 volts at each panel. Expected Voc with a 20C rise (pretty average for a low wind, ~70F day) would be about 35 volts, so that's a strong indication that the array is functioning correctly and is just very close to its Voc operating point (i.e. the inverter is drawing very little current.)

[gtuplin]

Oh boy I have a big problem if I'm doing this correctly!!
So I took the tester as described set on AC volts then did test black probe on service ground then red on DC positive and it reads 53 V AC !!! yikes! I then tested DC negative to system ground and have reading of 53.4v AC. Is this possible? I tested each inverter all the same?
If this is the case I think that I will try to add a ground rod about 30meters away in a dampish bog (dry now for summer but digging down its damp) then run a line and attach to ground rod at service, if this works great if not I don't know!!
Am I on the right track?

[Robert1234]

Why did you test the DC lines with the meter set on AC? (Not sure I understand what you are looking for with that test?)

[Mike90250]

Sounds like you got a big clue with the grounding, and are likely missing a large neutral wire

[adoublee]

I'm wondering if the inverters are set up to be AC coupled and are limiting output based on frequency on the AC line as it appears Growatt inverters are capable of frequency shifting to control power output. I would expect this to be unrelated to your water line of thinking, but rather other loads on the AC line causing the frequency to change and therefore inverter output. Improper AC voltage does not seem like something that would otherwise cause reduced output, as UL1741 protections should take the grid-interactive inverter offline completely.

[sunnyguy]

That is interesting. Since you have a neutral connection to the transformer it seems to imply that the new transformer is also poorly grounded. Then the grounds are acting like antennas for the overhead transmission lines and disrupting the inverters ability to lock onto the grid waveform? If the bucket test is repeatable then looks like you may have found the issue.

[gtuplin]

Why did you test the DC lines with the meter set on AC? (Not sure I understand what you are looking for with that test?)

The electrical engineer as me to check because of the entire system being install and contained under the high power transmission lines that it is possible that everything array connected metal wise could pick up ac power from above if not grounded properly. Both possitive and negative on the DC side tested to system ground show excessive ac voltage, his thought is it is creating a unique or distorted wave and confusing electronics.
Going to buy some ground rods and drive them into some wet smampy soil and run cable to see if a correction or change is noted.

[foo1bar]

<quuote>The electrical engineer as me to check because of the entire system being install and contained under the high power transmission lines that it is possible that everything array connected metal wise could pick up ac power from above if not grounded properly.</quote>

53V AC between the DC wire and ground under a transmission line.
And there's no high-impedence path to ground?
That may or may not be a problem.
My guess is not a problem - that the voltage is there when you're using the voltmeter, but as soon as there's a connection to the wire the induced voltage drops to hardly anything. ie. that the induced power is some small amount - and at near-zero curent, there's 53V; but as there gets to be a smaller resistance more current flows and the voltage drops to a really small amount.


Adding a ground rod/ground wire that connects to the existing ground rod shouldn't cause any problems (nor would I expect any real benefits)

[jflorey2]

So I took the tester as described set on AC volts then did test black probe on service ground then red on DC positive and it reads 53 V AC !!! yikes! I then tested DC negative to system ground and have reading of 53.4v AC. Is this possible?

Sure, if the inverter is isolated. In an isolated inverter the DC side "floats" with respect to the AC side, so unless you explicitly ground one side or another it will float to whatever external influences it sees are. In your case it would be capacitively coupling to the power lines.

However your inverter is transformerless, which means that it is not isolated - so the array is referenced to the AC side of things. And that is a problem, because you are seeing significant real energy being coupled into the inverter.

If this is the case I think that I will try to add a ground rod about 30meters away in a dampish bog (dry now for summer but digging down its damp) then run a line and attach to ground rod at service, if this works great if not I don't know!!
Am I on the right track?

You can try that experiment, but if your utility has a grounding problem of that magnitude (i.e. the neutral isn't near ground) they will have to fix it. If the service there is as bad as it seems to be, you run the risk of returning significant neutral current through your ground.

A suggestion - if you do sink that second grounding rod, first measure the voltage between the new rod and the incoming neutral. Do this BEFORE connecting the rod to anything else. If you see a high voltage, then that's a big problem with your service.

[DanS26]

Grant, in your first post you stated that the POCO required you to mount the arrays on wooden posts. Did they state why? Sinking more ground rods may not be wise....I'd check with a knowledgeable POCO engineer before proceeding with that experiment.

Just how high a voltage are those overhead lines? Counting insulator discs may give you a clue. We usually use as a rule of thumb for determining the number of porcelain/glass suspension disc insulators: 10kV for each disc (standard 5-1/4' x 10").
For typical system voltages in North America, this would be:
69kV: 4-6 discs; 115kV: 7-9 discs; 138kV: 8-10 discs; 230kV: 12 discs; 345kV: 18 discs; 500kV: 24 discs.

[gtuplin]

Well exciting in theory but it did not resolve the issue ! I will try to answer questions in bulk from your post
Ok so my understanding is that is a 230 KV line I will count insulators or check on line.
The wood was to limit the amount of metallic parts under the lines to reduce inductive voltage I think.
If I had a utility issue would that not also affect my home and the other system or are you thinking the issue is from my new transformer inward to solar?
The voltage did not change with the additional grounding.
I have been unable to recreate the water test but the ground is still damp but do plan to revisit that.

Now a question or two.
So I have 53 V DC from DC+ and DC- (checking AC) to ground but as stated above it may disappear under a load to zero I have not tried this but very possible but when I test DC+ to earth ground (DC setting on meter) I get 112V DC and then checking DC- to earth ground I get 271v DC. Is this normal?
Now because I am not an electrician I am just asking, would it be considered wrong to connect DC- to earth ground? To be clear it is not now but was wondering.

[gtuplin]

Grant, in your first post you stated that the POCO required you to mount the arrays on wooden posts. Did they state why? Sinking more ground rods may not be wise....I'd check with a knowledgeable POCO engineer before proceeding with that experiment.

Just how high a voltage are those overhead lines? Counting insulator discs may give you a clue. We usually use as a rule of thumb for determining the number of porcelain/glass suspension disc insulators: 10kV for each disc (standard 5-1/4' x 10&quot.
For typical system voltages in North America, this would be:
69kV: 4-6 discs; 115kV: 7-9 discs; 138kV: 8-10 discs; 230kV: 12 discs; 345kV: 18 discs; 500kV: 24 discs.

There are two lines side by side one one large metal tower and the second appears older. The oler one is three single lines suspended by 6 discs. The one on towers 6 lines also suspended by 6 discs. I would think that makes it a 230kv plus a 115 kv so 345kv?