So I just got 7020 watts of panels and I'm generating 6000 watts AC this morning - cool!

I'm wondering how much power is normally lost in the wires, and wondering if my inverter is properly sized or if I'm throwing watts away - at 6000 watts?

The inverter is rated for "8000 watts DC" and "6000 watts AC"?   Specs of the inverters here.  I have the 6000AUS: https://www.wholesalesolar.com/cms/solaredge-se3000a-us-u-inverter-specs-3099568964.pdf


I ended up using QPro panels with a SolarEdge inverter and optimizers on each panel.

I talked to a bunch of installers and  finally just went with the guys who seem to have a good reputation around here.  The price was $3.34 per watt installed.

I could have gone SolarWorld + enphase micro inverters for about $3.50.  I don't know if it would made any difference.

I like the idea of getting the inverter off the hot metal roof  and into my nice cool , dry basement.

So its 260 * 27 panels for 7020 watts.  It will be interesting to see if if this gets us to "net zero".

Thank you for helping me through everything.

There's very little power lost in the wires, most if the losses are in the inverter magnetics & semiconductors. If you're burning any appreciable power in the wires they would be warm or even hot to the touch. The cabinet & convection cooling air coming out of inverter may feel pretty warm if it's dissipating 100-200 watts (which it would be at 6kW-AC output), but not scorching hot.

It's highly unlikely that a 7kw array could be actually delivering 7kw-DC on any February morning in VT, since your solar intensity is lower than the rated-tested 1 kW/m^2 spec at this time of year, even with the enhancement of snow reflection. Operating in the upper 75% of rated power most inverters can be running in the 95%+ conversion efficiency range, but substantially lower efficiency when operated in the lower 10% of it's range (say, on cloudy day).

In VT you may slightly exceed 1 kW/m^2 at mid-day on bright sunny days in June & July, but not in winter and certainly not in the morning. So you're not delivering 7kW-DC into the inverter- it's probably getting ~6150-6250 W-DC in to get 6000 W-AC out. Mid day this time of year it might even hit 6300 W-AC if you have dead-clear air and some snow-reflection enhancement. Don't be surprised to see peaks north of 7500W on the cusp of the summer solstice on cool-but-sunny days.

An 8kw inverter would be right-sized for a 7kw array in VT, but would be undersized if your array were located on a mountaintop in Peru where solar intensity can reach well above the rated intensity (due to latitude & altitude, and colder temperatures that are conducive to higher efficiency.)

Thanks for the explanation.  I didn't think there would be much loss in the wires.

So do you think I should have the next size up inverter?  Sounds to me like I could be throwing power away in the summer by limiting myself to 6000 watts AC power in the inverter, if we're already hitting 6000 AC watts on a sunny day in February, but I imagine there's more to this , probably relating to power output on less than ideal days , and inverter efficiency.

From SolarEdge (inverter) perspective, the max DC input for my inverter is 8000 watts, so I'm well below that.  The max input for the next inverter up is 10,000 watts. 

Solar Edge seems to suggest what they picked is fine but not necessarily the best choice:

http://www.solaredge.com/files/pdfs/inverter_dc_oversizing_guide.pdf
(If you read their stuff, their word "oversize" , should be replaced with "overpowered" - the inverter is not oversized, the input power is. )


The main reason to oversize an inverter is to drive it to its full capacity more often. This will maximize power output in low light conditions, thus allowing the installation of a smaller inverter for a given DC array (or alternately installation of more DC power for a given inverter). Oversizing the inverter is typically not a requirement, however an experienced PV designer may choose to oversize the inverter in order to maximize the power production, due to the following:
 Actual PV module power vs. module nominal power
  Financial considerations

It looks like the SE7600A-US would have been the right choice for a 7kw array in VT. (I hadn't really read the spec prior to posting.)

The maximum AC output of the SE6000A-US is in fact 6000 watts, despite being rated for 8100 watts DC on the input (which it never actually takes in, or it would catch on fire.) The 6000 watts is also it's nominal AC output. Most nominal ratings are a bit lower than peak ratings, which seems to be true for the SE7600A-US, which has a nominal output of 7600 watts, but can deliver peaks of 8300 watts (or 700 watt higher than it's nominal rating.) This is a more appropriate sizing.

If the system is actually delivering the 6000 Watts in February (in the morning, no less), it's DEFINITELY robbing the system of throughput. Even on the brightest day in June/July you won't be hitting the peak 8300 AC out of the SE7600A-US, but you might beat it's nominal 7600 watts for a few hours per year.

The 1 kW/m^2 is roughly the mid-day mid-summer solar intensity of most temperate climates at about 48-50 degrees north latitude. You're a handful of degrees south of that, which will deliver a bit more under the right weather patterns. Higher summer temperatures will reduce the summertime output by quite a bit, but you'll still get significantly more than 7kW-DC during peak hours out of it, particularly in it's early years. (It'll degrade a few percent over a couple of decades.)

Definitely undersized, and your February AM measured output proves it. I'd take up the issue with the installer- they SHOULD swap it out on their dime if they're straight shooters. If there is a data logger on the system and you can track the AC output (rather than just the daily kwh), you'd have a very good case if it ever came down to a legal fight. There's just no way it should be power-limiting the output AT ALL in winter, even if it may be acceptable to give up 5% of the power on 1% of summer days rather than stepping it up to an even bigger size.

Bright late winter days have the best output and it generally goes down as the temperature goes up (even if intensity is higher).

You can probably remove ~1000 watts of panels from the current string and put them on micro-inverters (I have Enphase). I'd wait and see how often you max out 6K.

I just got up and running literally last night.
According to the monitoring, we were maxed out at 6000 watts for about 3 hours today (11am-2pm).

This felt like a very sunny day and I don't know, maybe a smaller inverter does better on a marginal day and the net AC output is actually higher with the smaller inverter? 

For example  SolarEdge says "In many cases, having more DC power than the inverter AC power, may increase power output in lower light conditions".
And goes on: "The main reason to (have more DC power than inverter AC power) is to drive it to its full capacity more often. This will maximize power output in low light conditions"

Here's the output for today - slow start I suspect is snow on the panels melting off.
https://c2.staticflickr.com/2/1615/25162962612_db5ddab58f_b.jpg

I will ask the installer what they think.

The remainder of the sentence after your quote is clearer:

This will maximize power output in low light conditions,thus allowing the installation of a smaller inverter for a given DC array (or alternately installation of more DC power for a given inverter).

In other words, it's all about not overspending on the inverter to capture panel output that is rarely there. Your design is probably pretty close to optimal ($/watt) over the seasons/life of the panels.

jonr: There are many bright mid to late spring days in VT when the outdoor temps are fairly cool (high-40s & 50s- no wait, those were the temps in central VT just yesterday! ) , which is where I'd expect the annual peak output to occur for patonbike.

I like the idea pulling 3-4 panels and putting them on micro inverters, if that would work with the array as configured.

patonbike: If the incoming power is 10% of the maximum rating a lot of the time, then maybe you would get more annual power out of it, but given the amount you've been clipping in the past few days at April or May-like outdoor temperatures makes me think that's not the case for you.

If you lived in foggy-dew Vancouver B.C. there may be an argument for sizing it the way yours is, since a large fraction of the time they're in "bright clouds" conditions.

According to weatherspark data clouds didn't break until about 11AM in Montpelier today, but it's been a nice cool 17-18 F & sunny outside.

But if the installer has actual data sets from nearby installations with similar levels of inverter under-sizing that show higher annual output than with a slightly bigger inverter, there's nothing quite like real data to settle it.

If you take a peak at the efficiency curves of the SE6000 on p4, you'll see it's falling off an efficiency cliff crossing the 95% level practically free-fall below 400 watts of output:

http://www.solaredge.com/files/pdfs/application_note_solaredge_inverters_efficiency.pdf

Unfortunately they don't have a curve for the SE7600, but assuming it scales, the knee in the curve would start at about 500 watts of output. The question becomes how many daylight hours would the DC input be under 525 watts but over 420 watts, and is the 1-2% improvement in that band with the smaller inverter worth giving it up a bunch on the high end? You probably gave up more than a kwh today, maybe more than two (but not three). I simply don't know how many hours of cloudy hours operation it takes to make up for that loss, and the marginally higher low-light output of the smaller inverter.

Make sense. I'll just see what they say. These guys do a lot of installs and should have a lot of data.

Pretty cool stuff though.

At least in one case, it is late March for the worst clipping due to inverter under-sizing. Poor installers - they can always be accused of under-sizing either the panels or the inverter (a perfect match changes with the weather). Note, it's not about maximizing power output, it's about maximizing $ ROI.

That's the clipping loss, but where's corresponding data set for the higher low-light efficiency gain on the same system?