Peak temp was 62C with output of 1920 watts AC. Xantrex might want to revise the derating chart in the manual.

Sean

What are you trying to tell us ?

Can you explain in detail please?

According to the manual (page A-3), the inverter will derate as the temperature increases, starting at about 30C (3000W) and taper to 0 watts at 70C. At 62C the line appears to be slightly below 1000W.

Mine was processing twice the manual wattage and appeared to not derate. I know it has been stated on this forum that the GT performs better than specifications, but to actually see it in production impressed me. I guess I'm just used to products that don't meet the manufacturer's specs.

Sean

Thanks for your explanation, I thought your we thinking the opposite

The GT3.0 won't even begin to derate until the heat sink hits 70C. So the full 3200 watt (dc) limit is available

Also I know that they have continued to tweak the design and will be upping the AC rating even higher.

So yes, the GT has been real-world conservative rated so end-user will get the performance promised under even the most stressful environments.

I was out at 2 adjacent sites with GT3.0 Inverters. One has 36 Kyocera KC170GT (170W) panels connected to 2 GT3.0 Inverters (2 x 18) and the other has 32 Kyocera KC190GT (190W) panels connected to 2 GT3.0 Inverters (2 x 16). The strings were sized based on outputs from the Xantrex string sizing program based on San Diego, CA temperature inputs.

The California Energy Commission (CEC) rating for the above systems (per Inverter) are 2552 Watts (18 KC170GT panels to a GT3.0) and 2536 Watts (16 KC190GT panels to a GT3.0) respectively. Generally, I am happy if I see a system outputing its CEC rating.

It was a cool cloudy March morning in San Diego, CA but nothing extraordinary. The open circuit voltage was around 515 Volts (for the 18 Kyocera KC170GT systems). As the clouds would clear briefly, all 4 inverters would be putting out 2900-3000W which means the systems were outputing about 18% higher than their CEC rating. I have never seen a system running at so high above the CEC rating. Even if the inverters were running at 95% efficiency, the panels must have been putting out close to their STC values. Solar Guppy has stated that the Xantrex GT3.0 Inverter reports the correct output to within 1%.

My questions for which I would welcome comments are:

(1) What is an "Edge of Cloud" effect and was this one? It lasted for a long time, about 10 - 15 minutes until the next cloud came in.
(2) How critical is the 600 VDC upper voltage limit for the GT3.0 inverter and am I dangerously close? My guess is the ambient was around 18 degrees Celcius and maybe historically it could get -10 degrees Celcius. At -0.12 V/DegreeC, it could be getting close to 600 VDC open circuit. Kyocera Panels are listed as +10%/-5% which presumably applies to the Voc also. With a batch of +10% Voc panels, I would be over the limit.
(3) With a 2500W Inverter, I would be throwing away the excess power. Seems like it is a good general design idea to size the panel strings at about 80% of the inverter capacity. You'll never throw away excess power and would avoid de-rating in hotter ambients.

Anyways, I was impressed with the performance of both the GT3.0 Inverter and Kyocera Panels. Since then, the KC170GT and KC190GT panels have been replaced with the KC175GT (175W) and KC200GT (200W) panels that look identical. I wonder if they made any change or just found that they were performing at the higher specification.

Clouds, being white are giant reflectors. Now there are not focused but do end up reflecting/redirecting/absorbing light that would normally be in an area under the cloud and in a defuse way reflect this to other areas on the ground. Its totally random, but in partly cloudy weather, 1.2 to 1.3X the solar irradiance can be seen and converted by the solar pv panel. Its usually only for a few minutes and typically is between the peek-a-boo the sun plays with the pv panels.




I'm not sure what the "I am" part as the inverters you mentioned we not yours?

the +/- 10% is almost always due to current variations, not voltage. I have looked at 60+ PV flash data tester results and never did the voltage vary much. The Voltage of Silicon Junction is pretty much near constant, its the size and chrystaline structure variability that causes the current to vary. Now, I'm talking about at the same give tempature here and its IS the PV voltage that causes the large voltage variations due to tempature ... got that?

As for 515 VDC in cool conditions, thats about as close as I would like to see and I wouldn't add panels in series to this, but it should be fine. The limits are real, the Mosfets are 600V rated, and this is the reverse breakdown voltage value, also the input caps are 600V rated. Now both the Mosfets and Caps might handle brief periods above this but it should be avoid at all costs!

If your planning a system, I strongly recommend you target a lower VOC. ALL string inverters are more efficient at the lower voltages, this is due to switching losses in the Mosfets, which increase as the voltage increases. The GT is about 1-1.5% better at 200V vs 500V and its the same ball park for all string inverters

Thanks for the comments Solar Guppy. Everything was clear.

The "fit" of Kyocera Panels, the only ones I have been able to get for a decent price the last 2 years, with the Xantrex GT3.0 has necessitated 500 Voc strings. I could load them less, but often the competition is stiff and it becomes a $/Watt game. More inverters would raise the $/Watt.

The situation may have changed with the release of the GT3.3 and GT3.8 inverters. However, I still don't think it is a bad idea to stay with the larger strings, and live with the 1-1.5% efficiency loss which you mention. My experience is that most retrofit sites in the city are subject to some shade at some point in the day (from a tree, chimney, etc.). For the site I mentioned, even if many of the panels (up to 7) are shaded the system will MPPT.

Even if a site has full sun now, there is no guarantee that it will not be subject to some shade in 5 - 10 years. The price of these systems is such (especially with the current PV panel crunch) is that people need trouble free good performing systems for many, many years to realize payback.

I wasn't suggesting you lighten the load, just use three strings instead of two and adjust the lenghts accordingly

As for the partial shading, unless the two ( or three ) strings get shaded the same, losing more than two panels will pretty much wipeout the generation from the remaing string, actually, the GT will mppt and both strings will be comprised output wise so being close to the bottom maybe better even in that condition

Just some things to concider

jpinon,
even with the larger inverters available there may be another problem to stand in your way if putting more in series. i tried to verify this on kyocera's site, but didn't see this listed. it's the voltage rating of the wires/box for the pvs. usually this is about 600v max.
as to the higher voltages during edge of cloud you should be fine, but this could go to or exceed 600v if the pvs are cold or cool. if you look at your pv specs from kyocera they show how the voltages are higher on the graph with a 25 degree c temperature. that temperature is 77 degrees fahrenheit. that's the module temperature and even when it's below freezing outside module temps are higher due to alot of solar absorption. even the flow of electrons can raise it a tad too. if cold enough like in early morning with high amounts of sun your voltage will be high. add to that the edge of cloud effect and it could go over. fear not as it was engineered for you as being a good working system by xantrex so if that really rare time happens and is long enough i'd say they are liable for the damages, but you may have to be able to prove that's what happened.
p.s. you may be losing some in efficiency with higher voltages in the inverter, but you are gaining by percentage slightly because of lower wire losses. i also believe the higher voltage systems will start delivering power faster and longer compared to the smaller voltage systems.