I just saw that the CEC rating of the GT3.0 has stayed at 94% efficiency and 3000W continuous output. Congratulations to Xantrex! This justifies much of what I have read on this web site.

My question is the string sizing for the GT 3.0. I saw the Xantrex web site string sizing program. I want to put 18 Kyocera 187W panels in a 2 x 9 config but the Xantrex program does not recommend this above 77F because the string voltage will drop below the 195 Vdc lower MPPT tracking threshold. This seems quite conservative to me compared to other string sizing programs.

The specs for that panel are 32.5 Voc at 25C with -.123V/C. Even at 75C (a pretty hot roof temperature), the Voc should be 26.35 Vdc and a 9 panel string should be 237.15 Vdc which is well above the 195 Vdc lower threshold of the inverter. Am I doing something wrong or can I safely use this configuration? Thanks in advance for any input.

The GT3.0 is the only inverter that didn't drop in it rated output and efficiency ... Can't wait to see PVPowered , I measured 93.2 best case with my Yokogawa (they have been claiming 97.5% which is pure fiction)

On your panels ... it vmp your concerned about , that's about 80% of VOC , so take VOC * thermal derating , which for Si panels is .08volts per "C" above standard conditions (25C) * 0.8 = vmp at tempature

Or take the actual VMP given and do the 0.08V per C above 25C

Using your VOC number

32.5 - (.08 * 50) = 28.5 VOC * 0.8 = 22.8 VMP (mppt operational point)

22.8 * 9 = ~205 VDC


I aggree with xantrex , not enough margin for hot roof installations. It will work , but in hot weather you might lose some watts , where is the installation located ?

Hello Solar Guppy,

Thanks for the input. It was the 0.8 factor I was missing. The ambient can be expected to reach 104 F in the summer. The site is actually comprised of 36 Kyocera 187W panels (spread over a south and west facing roof) connected to 2 GT3.0 inverters. I wanted to have 18 on each inverter but that is not going to work for reasons we discussed. I am now thinking of 16 (1 string) on the south facing side and 20 (2 x 10) on the west facing roof. 10 in a string should put it always operating in the MPPT range. The loading is high but because of the west orientation it should never exceed the 3000W output capability of the inverter. Do you think that will work?

Regarding the new CEC list, when can we expect numbers on PV Powered or Sharp or even the Sunny Boy 2500? Or have they missed the deadline?

Your 16X1 and 10X2 would work .....

On the 16X1 your working voltage will be on the high-side , My math check for cold conditions , its the same .08 Volt/C from 25C ...

So for a cold morning , say minus 5C thats 30 * .08 = 2.4 + 32.5 = ~560V , so I think that will be OK

On the 2X10 , thats 3740 STC watts ... I looked up the specs on the 187 , 26.1 Volts at 7.17 amps. In Hot weather , the power available would be ~3200 watts DC peak , more in cooler weather , but keep in mind this is only when the sun angle is perfect (on a few days a year in roof mount installations) , you have wire and inverter losses , so I'd say its OK as well , you will be running the inverter in the 2600-2700 range most of the time , no problem , but there will be a few time on cold days when some of the peak power can't be harvested becuase you will run into the 3200 watt DC limit of the GT.

Overall , I agree it looks like the best fit for this installation

I read on the re-wrenches list that SMA will have everything listed with 2 weeks , there was quite a rush to get all the productes tested by the deadline , I know 2 weeks ago Xantex had every UL/CSA inspector in the US and Canda available working on getting there products test data verifed.

Despite what some re-wrenches are saying , this is the first real attempt to get a level , honest playing feild on inverter products. You will find most products will have a lower efficiecy and or lower power rating and no longer are these values taken on a manufactures say so. Its all to lab standards now , cerfiied/calabrated test equipement on a standard , defined way to test and measure.

On PvPowered , I haven't heard anything ...

This industry has long been overdue for this , I think the CEC/Sandia is a huge step in the right direction !!.

Just saw that the PV Powered is rated 95% efficient and 2800W output. What do you make of this? My impression is that it is higher than some expected but lower than their Sales rep said it would be.

I'm really waiting for the SMA 1800 and 2500 and to see if they dropped 2 points like the 1100 and 700.

I'd have to see the raw data and also what the equipment was used to measure with.

My tests were last June, using a Yokogawa WT2030, which is spec'ed to 0.08% in measurement accuracy. I also tested the Fronuis 2500 and 3000 and the new numbers match exactly to what I measured , which was 91-93% depending on the operational input voltage.

These are weighted tests and the PVPowered unit has a 150V smaller operational range which will help there numbers as compared to the GT3.0 and the SMA 2500, the higher operational range drags down the final weighted efficiency number.

It's also possible the unit I tested was below par or that the latest unit is using better FET's , a better wound transformer ... without dissecting the actual test unit , one will never know why I guess

I thought it was telling the PVP2800 208Volt unit is rated at 94% and the PVP2800 240Volt unit was 95% , such a small change in the VAC showes there very close to the 94.5% window

One Last note is 94.49 is 94 , 94.5 is 95 on the CEC scale ... so I'll look with interest at the underlying data when it becomes available

Just so you know, the current CEC website has a bug that will not allow the decimal place to be displayed.

The testing protocol specifies a minimum accuracy of +/-1%, and we requested that data be rounded to the nearest 0.5%. More than a third of the inverters have 0.5% numbers that are being truncated. The PVP 2800 208V is one of them (actually 94.5%).

There is also a note that these data are under review along with some other test results.