Analysis of Microinverter Clipping

[nsgoldberg]

How much clipping on an instantaneous basis ? How do you measure the input to the panel(s) ?

Not measuring instantaneous clipping. I'm relying on the graph that loads from Enphase's app. I know it's clipped when the max output for the day is 5.76kW. (240W x 24 panels)

On a couple of the days, a small plateau is visible at the peak, where the parabolic curve has it's top cut off.

[nsgoldberg]

FYI, new panels tend to put out over nameplate for a few weeks after install. Once that initial bump goes away they settle down for the long run degradation. If you live in cold area the panels also can put out over nameplate in cold conditions with snow on the ground. Possibly the combination of the two caused a but more clipping?

Southern California. No snow on the ground here.
Looking at the daily estimate from PVWatts, it appears the clipping is factored in. During summer months, I should see 3-4 hours a day where it hits the maximum output. Not ideal, but the math still holds true, as long as the PVWatts estimate is correct. We have had cold but sunny days, so I'm sure that's contributing.

[J.P.M.]

How much clipping on an instantaneous basis ? How do you measure the input to the panel(s) ?

[peakbagger]

FYI, new panels tend to put out over nameplate for a few weeks after install. Once that initial bump goes away they settle down for the long run degradation. If you live in cold area the panels also can put out over nameplate in cold conditions with snow on the ground. Possibly the combination of the two caused a but more clipping?

[nsgoldberg]

You raise good points and your analysis is sound. The only thing worth adding is that panels degrade with time, so today's 330 watt panel will be roughly a 300 watt panel in 10 years, depending on a zillion little details, such as temperature, sunlight exposure, manufacturing quality, materials, etc.

That's true, and I had considered that as well. At 25 years, the panels are guaranteed to put out 80.7% of the original amount, or 266W. At that point, the effect is negligible.

At half of that, 12 years, the guarantee is 89.1%, or 294W. Likely also negligible. So after paying back the initial cost of the higher capacity micros, 8 years, the return on your investment may quickly deteriorate. I say "may", because I'm going off the minimum spec in the warranty. It *should* be higher than that.

[bob-n]

You raise good points and your analysis is sound. The only thing worth adding is that panels degrade with time, so today's 330 watt panel will be roughly a 300 watt panel in 10 years, depending on a zillion little details, such as temperature, sunlight exposure, manufacturing quality, materials, etc.

[nsgoldberg]

From what I've read, clipping is negligible in the long run. However, after less than a week's output, I see some clipping occurring already. Nothing big, at least not so far. I started wishing I had done a bit more research on the microinverters and used the IQ7+ (290W) instead of the IQ7 (240W) with my 330W panels. My AC/DC ratio is 1.38.

However, I started doing some math. Inputting this data into PVWatts, I get an annual difference of 239kW between the two microinverters. At $0.20/kWh, that's about $50/yr. Not big, but at only $15 more for the IQ7+, that's an 8 year payback on the difference between them, assuming electric rates don't increase (which they will). Keep in mind, the string is limited to 13 micros vs 16, so some cable planning needs to take place beforehand.

A better option would be to spend the extra $360 on an additional panel and microinverter, assuming you have the room to do so. This gives an increases the PVWatts estimate of 582kWh/yr (vs the 239kWh of the IQ7+).

Bottom line, after initially being upset and reading a bunch of articles and forums on clipping, it's best just to run the PVWatts estimate. If I were to do it over again, I'd use the IQ7+ and center feed the array to make using two strings easy. An 8 year payback on a system that will be in place for 20+ years is worth it in my opinion. But, the difference really is negligible, and if you have the space and a set budget, adding just one more panel is a better ROI than paying for a higher rated micro. I'm less familiar with string-based systems, but I assume the same logic would follow. YMMV.