Enphase microinverter efficiency

[fresnoboy]

Efficiency is technically power in to power out. If the panel CAN generate 400 watts, but is actually generating 290 watts, and 280 watts of that goes to the grid - the inverter's efficiency is 96.5%.

Because it's more cost for a limited increase in power.

I'll use my system as an example. It's an older Sunpower system with 9680 watts worth of panels. However, the two inverters, together, are rated at 8700 watts. Why is this OK? Because most of the time the panels generate closer to 7000 watts and the inverters have no problem using all of that. The only time I've ever seen close to 8700 watts are right after rain, when the panels are clean and cool and there's a little cloud lensing going on.

In your case, depending on many things (angle, insolation, typical weather, temperatures) a 400 watt panel will produce about 280 watts at full sun. That means the Enphase will work well for you. Under certain conditions (the first sunlight after a rainstorm) the panel will generate close to 400 and you will clip - but those cases will be relatively rare.

Want to see this for yourself? You can get a panel and a curve tracer (they are getting cheaper now) or just a panel and a variable resistor and mount it outside your house. Measure the panel's power periodically and see what you get. This will give you a good sense of how much power you will see vs the panel's actual rating. Make sure you leave the panel outside for a while so it picks up a "typical" layer of dirt, as all panels do.

Interesting - definitely something to model in PVWatts (thanks for the pointer). I'll try and find the proper inputs in terms of panel geometries and such and run the numbers.

Also, how big an impact is dirt on the panels? It's a brand new house build, so I could put run water up to the roof line and put some spray assembles on them so I would rinse the panels once a week or so. Not sure about the cost tradeoff, but because its not yet in construction, it's very inexpensive to add.

Are there other things I could do during the build of the house and roof to improve efficiency, other than redesigning the roof?

thx
mike

[jflorey2]

Also, how big an impact is dirt on the panels? It's a brand new house build, so I could put run water up to the roof line and put some spray assembles on them so I would rinse the panels once a week or so. Not sure about the cost tradeoff, but because its not yet in construction, it's very inexpensive to add.

I wouldn't recommend that for three reasons.

1) In PG+E territory, water is often in short supply.
2) When it's _not_ in short supply it means it's been raining, which means the panels get cleaned anyway.
3) The water there is very hard and is as likely to leave hardwater deposits as it is to clean the panels. (Rainwater has no minerals in it, so it works a lot better.)

[J.P.M.]

Interesting - definitely something to model in PVWatts (thanks for the pointer). I'll try and find the proper inputs in terms of panel geometries and such and run the numbers.

Also, how big an impact is dirt on the panels? It's a brand new house build, so I could put run water up to the roof line and put some spray assembles on them so I would rinse the panels once a week or so. Not sure about the cost tradeoff, but because its not yet in construction, it's very inexpensive to add.

Are there other things I could do during the build of the house and roof to improve efficiency, other than redesigning the roof?

thx
mike

I've got an interesting side hobby that's become an ongoing project. I attempt to measure how dirt (fouling) affects my array's output and performance and what effect rain may have on fouling.

Without rain, I believe my array fouls at a rate such that performance degrades, very approximately at a rate of ~ 0.6% - 0.8% per week, +/- a bit depending on wind and a bunch of other atmospheric variables, like A.M. fog that cakes up the dust, wind, blown dust, especially from Santa Anna's, etc.

A decent rain of ~~ 0.1" to 0.3" or so restores ~~ 2/3 to 34 of the performance that was lost due to dirt accumulated since the last rain or cleaning.

A hose rinse, with no soap at a rate of ~~~ 3/4 - 1 gal. of H2O per panel seems to restore ~~ 3/4 to 90%+ of the performance lost to fouling or a bit more than a decent rain might do.

After measuring performance several hundred times over the last 3 years under a variety of conditions, clean, dirty and in between, I've found that hard water spots from hose water no soap cleaning/drip/air dry do not impair performance over cleaning and squeegeeing/wiping dry to remove spots.

One area still difficult for me to get a handle is whether or not that weekly performance roll off rate of ~~ 0.6% to 0.8% or so without rain stays reasonably "constant" over time or, as I suspect - but have no confirming data yet - tends to become asymptotic after some length of time, say after 8 - 12 weeks or so and depending on other weather conditions besides rain. Hard for me to accept that performance would roll off, say, 0.7%* 52 weeks = 36. % /yr. For one thing, skylights in the desert don't get that dirty.

Bottom line for my array, if I ever stop measuring instantaneous performance, when it doesn't rain, I'll spay the array with a hose about 1X/month IN THE EARLY A.M. BEFORE THE GLAZING HEATS UP and let it drip dry. Doing that and allowing such rain as does occur around here to clean the array will mean I'll perhaps incur an average fouling penalty of something like 3% or maybe a bit less.

[fresnoboy]

I wouldn't recommend that for three reasons.

1) In PG+E territory, water is often in short supply.
2) When it's _not_ in short supply it means it's been raining, which means the panels get cleaned anyway.
3) The water there is very hard and is as likely to leave hardwater deposits as it is to clean the panels. (Rainwater has no minerals in it, so it works a lot better.)

Water is in short supply, but there is a high watertable in my area, and I may put in a small well for irrigation of the plants. I haven't done the make vs buy analysis on that yet comparing the cost to city water, but others in my area have done that. If I do go that route, it is just the cost of electricity to run the pump, which would be running anyway for irrigation.

Good point about not watering when the glazing is hot! Maybe at sunrise...

I think the right spray head would save water over a hose rinse from the ground, but that is something I haven't really thought about yet. It's interesting that dust, etc... fouling has that kind of impact on yield, but I don't think it can add up like that. In my area (SF peninsula) there is not that much dust compared to the valley, but it's something to think about.

Are there any professional studies on this point?

Thx
mike

[J.P.M.]

Water is in short supply, but there is a high watertable in my area, and I may put in a small well for irrigation of the plants. I haven't done the make vs buy analysis on that yet comparing the cost to city water, but others in my area have done that. If I do go that route, it is just the cost of electricity to run the pump, which would be running anyway for irrigation.

Good point about not watering when the glazing is hot! Maybe at sunrise...

I think the right spray head would save water over a hose rinse from the ground, but that is something I haven't really thought about yet. It's interesting that dust, etc... fouling has that kind of impact on yield, but I don't think it can add up like that. In my area (SF peninsula) there is not that much dust compared to the valley, but it's something to think about.

Are there any professional studies on this point?

Thx
mike

Some studies, but they range from no more than bogus adverts to sell panel cleaning at very non cost effective prices (perhaps interestingly, one such stab is a 12/29/2012 post on this forum), to grad theses white collar welfare rehashes of generalized data. Still , there a few attempts under varying conditions and environments that seem objective and interesting. Results seem to range from roughly 1.4 %/yr. to 80% or so. They're all over the place. PV is not a super refined branch of technology, and array fouling is probably the least developed backwater of the discipline.

See Google for all types of stuff. Ignore the adverts, and take all the rest as approximations.

The most that can be said in a few words is that panel fouling is highly variable depending on climate, local conditions, including soil and vegetation (e.g., pollen), site conditions, including panel tilt and proximity to dirt sources, and other factors.

Most "valid" studies that deal with fouling treat it as sort of an assumed mostly constant rate process, and describe how performance is reduced as f(time).

I'd put my data and resulting SWAGs as representative and practical of what is perhaps common in a semi rural environment.

Bottom line for me: Hose the panels off 1X/month if it doesn't rain and reduce expected performance by ~ 3 % or so due to fouling. Hose after sunrise, not before. You'll need light to see what you're doing. use ~ 1 gal./panel. H2O and don't bother with soap or rise agents or rags. I couldn't measure a difference in results from many trials and different methods, but suit yourself.

FWIW, one thing I think I did find is that the human eye is a poor indicator of the quantitative effects of dirt on a solar array. Basically, what looks real dirty may result in no more than a few % of performance impairment.

[NEOH]

Of course, you need to over-size your Solar Panels with an Enphase Inverter.
If you think it attaching a 280 Watt Solar Panel to a 280 Watt Enphase Inverter is optimal, well then you are just kidding yourself.
Anybody that actually has solar panels, an inverter and a daily log will know how low real PV production actually is.
And I don't care about any theoretical graph, it does not represent reality, reality reigns supreme.

The question from the OP was ... "... Why would you attach a 400 Watt Solar Panel to a 280 Watt Enphase Inverter? ..."

And the answer is ...
Because you will generate more Kilowatt-Hours each year.
A 400 Watt PV Panel would help to maximize the time the Enphase Inverter would generate 280 Watts.

At my location, I would not have any clipping for 151 Days - none, nadda, zip, zero - So, where is the graph for that?
The 280 Watt Enphase Inverter would harvest ALL of the energy created by a 400 Watt Panel all 91 days of winter ( extra 40% more ) - because the sun is too low.
The 280 Watt Enphase Inverter would harvest ALL of the energy created by a 400 Watt Panel in the 60 days of "summer" ( extra 40% more ) - because too hot or cloudy.

My inverter does not turn on at 0 Volts, it needs a minimum Voltage and minimum Watts to operate.
So I think a 400 Watt Solar Panel could start the inverter earlier in the morning and keep the inverter on later in the evening, so the base of the curve could be wider, too.

The "clipping" graph is appropriate (accurate) for only about 1% or about 4 days were my PV Array is generating 95% or more peak energy.
How many replies are so worried about "4 Days" and have completely ignored the other 361 days?
4 Days - who cares?

Let me be clear ...
I do not care that during 1% of the days per year, I will have only a minor increase in production due to clipping.
I do care about the other 99% days where my energy production will be significantly increased - 40% of the days will harvest at a full 40% increase.

If the increase in yearly production justifies the increased PV Panel cost and roof area required, then do it ...

[AzRoute66]

[...] One more, south facing ratio 1.0... output is 1.82 kWh / kW... suggests the clipping penalty to energy production is ~4%. [...]

I just couldn't believe that, so I ran it myself with the same results. Damn, another bar bet I would have certainly lost. There really is no substitute for experience. The odd thing is, I expected the clipping to be worse in summer, it was not. It occurred in all months, in fact the worst month I see without pulling out my solar powered desk calculator is March, which speaks to the 30 degree tilt. That shifted right when I ran it at 10 degree tilt, but not a lot. The moral of the story must be, "Don't expect more than 280/400 = 70% instantaneous (or less) out of your array very often."

[J.P.M.]

Of course, you need to over-size your Solar Panels with an Enphase Inverter.
If you think it attaching a 280 Watt Solar Panel to a 280 Watt Enphase Inverter is optimal, well then you are just kidding yourself.
Anybody that actually has solar panels, an inverter and a daily log will know how low real PV production actually is.
And I don't care about any theoretical graph, it does not represent reality, reality reigns supreme.

The question from the OP was ... "... Why would you attach a 400 Watt Solar Panel to a 280 Watt Enphase Inverter? ..."

And the answer is ...
Because you will generate more Kilowatt-Hours each year.
A 400 Watt PV Panel would help to maximize the time the Enphase Inverter would generate 280 Watts.

At my location, I would not have any clipping for 151 Days - none, nadda, zip, zero - So, where is the graph for that?
The 280 Watt Enphase Inverter would harvest ALL of the energy created by a 400 Watt Panel all 91 days of winter ( extra 40% more ) - because the sun is too low.
The 280 Watt Enphase Inverter would harvest ALL of the energy created by a 400 Watt Panel in the 60 days of "summer" ( extra 40% more ) - because too hot or cloudy.

My inverter does not turn on at 0 Volts, it needs a minimum Voltage and minimum Watts to operate.
So I think a 400 Watt Solar Panel could start the inverter earlier in the morning and keep the inverter on later in the evening, so the base of the curve could be wider, too.

The "clipping" graph is appropriate (accurate) for only about 1% or about 4 days were my PV Array is generating 95% or more peak energy.
How many replies are so worried about "4 Days" and have completely ignored the other 361 days?
4 Days - who cares?

Let me be clear ...
I do not care that during 1% of the days per year, I will have only a minor increase in production due to clipping.
I do care about the other 99% days where my energy production will be significantly increased - 40% of the days will harvest at a full 40% increase.

If the increase in yearly production justifies the increased PV Panel cost and roof area required, then do it ...

I'm having a hard time understanding where your coming from, as I am a lot of the logic behind some of the posts on this thread.

There's too much in your post that, to me anyway, seems based on incomplete understanding of some solar basics, starting with what seems to be the notion that an individual panel needs, somehow, to be sized to fit the micro attached to it. That seems a backward way to go about it to me.

If one of the big goals, among several, is to maximize system output to meet the design duty in the most cost effective way possible, then, if micros are the inverter type of choice, it would seem the goal would be to see which micros are available for the best match with the choice of panel, not only size wise to minimize clipping to the limits of cost effectiveness, but also with a careful eye toward inverter size vs. the cost of any differential clipping imposed by changes in the ratio max. panel output for the application vs. any inverter max. output.

An absurd situation: Suppose, as you write, one does need to oversize a panel for a specific inverter. How much to oversize the panel ? If I have some old Enphase 190's (still working ?), and I mate them to, say, new 300 STC Watt panels, is that a better choice than using, say 200 Watt panels ?

I fully appreciate, as do others, your information that, because panels rarely, if ever, produce their STC rated output, clipping is usually not a big a issue. However, I've found it's a bit less black and white than you seem to be making it.

Mated to the same panel in an application, provided the panel's likely max. output is greater than the rated output of the larger micro, the smaller micro will have a higher probability of clipping than a larger micro. Whether or not that reduction and its probability in output is acceptable is what needs to be determined and will vary with each application.

Another example, using Sensij's and my location as he's done in a prior post to this thread, suppose my need is to offset 14,800 kWh/yr. With no clipping, and to keep things simple, but realistic, I'll define annual output at 1,850 kWh/yr, for a 30 deg. tilt, south facing array with no shading, and ave. fouling at 3 %. My choices are limited to 360 or 400 STC Watt panels and either 250 or 280 Watt micros. With no clipping, I'll need either 14800/665 = 22.26 = 22 ea., 360 Watt panels , or, 14,800/738 = 20.05 = 20 ea. 400 Watt panels.

Running PVWatts w/10 % system losses and the 4 combinations of panel/micro:

360 W panel output w/no clipping = 665 kWh/yr., = 14,630 kWh/yr.
400 W panel output w/no clipping = 739 kWh/yr., = 14,780 kWh/yr.

360 W panel clipping w/250 W micro = 29.4 kWh/yr., = 29.4*22 = 647kWh/yr., or 4.4% annual loss , requiring 1 extra panel/micro to meet the duty and make up the clipping loss.
360 W panel clipping w/280 W micro = 6.3 kWh/yr. = 6.3 * 22 = 139 kWh/yr. lost, probably requiring no extra panels.

400 W panel output w/250 W micro = 66.8 kWh/yr., = 66.8 * 20 = 1,336 kWh/yr. = 9.04 % annual loss from clipping, requiring 2 extra panels/micros to meet the duty and make up the clipping loss.
400 W panel output w/280 W micro = 30.4 kWh/yr., = 30.4 * 20 = 608 kWh/yr. = 4.1 % annual loss from clipping, requiring 1 extra panel/micro.

As for actual output being different than model estimates, provided model inputs reflect what's actually present in the application, actual to modeled varation is mostly due to weather variations and observed time period than anything else. I've got everything my array's monitor put out in 5 minute increments since startup almost 4 years ago, as well as the log, in 1 minute increments of what my Davis instrument located on my roof about 4 feet north of my array has recorded since array startup, including horizontal solar irradiance., which I then transpose to POA (plane of array) irradiance using a modification of the HDKR model I've developed over many years.

As best as I can figure it, depending on some basic level of sophistication, the model's (including stuff I've written over the years) mostly all seem to do a pretty good job of estimating likely output of a PV system provided they get and start with a realistic and reasonable representation of the system they're modeling. If that's done, and the modeler has a good understanding of how a PV system interacts with the atmosphere, including how an array interacts and handles the instantaneous as well as hourly to seasonal effects of insolation and other environmental effects, annual output can be estimated with about the same probability and characteristics as predicting the weather.

Take what you want of the above. Scrap the rest.

[J.P.M.]

I just couldn't believe that, so I ran it myself with the same results. Damn, another bar bet I would have certainly lost. There really is no substitute for experience. The odd thing is, I expected the clipping to be worse in summer, it was not. It occurred in all months, in fact the worst month I see without pulling out my solar powered desk calculator is March, which speaks to the 30 degree tilt. That shifted right when I ran it at 10 degree tilt, but not a lot. The moral of the story must be, "Don't expect more than 280/400 = 70% instantaneous (or less) out of your array very often."

The clipping, such as it occurs at all, will be worst when the per panel output is the highest. That usually, and most commonly occurs when the incidence angle on the array is close(r) to 90 deg. and the atmospheric temps./wind vectors combine to keep panel/cell temps. low(er). The confluence of those types of events at usually occur, in the U.S. anyway, in the spring.

[bcroe]

Of course, you need to over-size your Solar Panels with an Enphase Inverter.
If you think it attaching a 280 Watt Solar Panel to a 280 Watt Enphase Inverter is optimal, well then you are just kidding yourself.
Anybody that actually has solar panels, an inverter and a daily log will know how low real PV production actually is.
And I don't care about any theoretical graph, it does not represent reality, reality reigns supreme.

The question from the OP was ... "... Why would you attach a 400 Watt Solar Panel to a 280 Watt Enphase Inverter? ..."

And the answer is ...
Because you will generate more Kilowatt-Hours each year.
A 400 Watt PV Panel would help to maximize the time the Enphase Inverter would generate 280 Watts.

At my location, I would not have any clipping for 151 Days - none, nadda, zip, zero - So, where is the graph for that?
The 280 Watt Enphase Inverter would harvest ALL of the energy created by a 400 Watt Panel all 91 days of winter ( extra 40% more ) - because the sun is too low.
The 280 Watt Enphase Inverter would harvest ALL of the energy created by a 400 Watt Panel in the 60 days of "summer" ( extra 40% more ) - because too hot or cloudy.

My inverter does not turn on at 0 Volts, it needs a minimum Voltage and minimum Watts to operate.
So I think a 400 Watt Solar Panel could start the inverter earlier in the morning and keep the inverter on later in the evening, so the base of the curve could be wider, too.

The "clipping" graph is appropriate (accurate) for only about 1% or about 4 days were my PV Array is generating 95% or more peak energy.
How many replies are so worried about "4 Days" and have completely ignored the other 361 days?
4 Days - who cares?

Let me be clear ...
I do not care that during 1% of the days per year, I will have only a minor increase in production due to clipping.
I do care about the other 99% days where my energy production will be significantly increased - 40% of the days will harvest at a full 40% increase.

If the increase in yearly production justifies the increased PV Panel cost and roof area required, then do it ...

The "one design fits all" idea works if you are in orbit. Those of us living with seasons, clouds, and other
restrictions are in a different situation. Your approach is not so different from the one I am using with extra
panels facing other directions. The rooftop micro options are much less than the ground mount strings
here. My only question is, would your approach here work better than mine? Bruce Roe

[tyab]

Great discussion. Here is real example. Enphase M250 micro, has a normal max output of 240W but can peek to 250W. Connected to 260W panels. System is under 10 months old so minimal loss from age. Hosed off the panels early in the morning so they were mostly clean. Ground mount so lots of free air space under panels help cool the temps.

Here is a 24 hour video showing the conditions (midnight to midnight): https://www.wunderground.com/webcams...e=20170917.mp4
Other than a few high clouds that drifted across the day was one of the nicest days we had in some time. Cooler and clear, haze from Railroad fire mostly gone, high temps of 82.2F. System is right at 3200' elevation so much better sky's than what you will find in the valley.

I pulled down the data from pvoutput (it is close enough to what Enphase is showing its good enough for this example), and sucked it into excel.

Yesterday the average output per micro over the entire day was 121.2W, so under 50% output
Between the times of 11pm and 2pm the average output was 202.2 W so around 80% output.
Max peak output was 217.2W at 1:10pm and it was near that value for about 30 minutes.

I can count on one hand the number of times I have hit 250W - all of them were after snow storm - all ground covered with snow - freezing temps and completely clear skies - and in all cases it lasted under 30 minutes. Sunglasses mandatory type of days. Other than that - never hit that 250W mark.

Thus expect that 400W panel on GREAT days to give you under 200W on average for the day. Maybe peak at 320W (above the IQ6+, so some clipping) but your daily output will be higher so it does not matter if you clip. As already stated in great detail by other posters - it all boils down to cost and your available roof space to decided what size panel to use.

[ButchDeal]

Great discussion. Here is real example. Enphase M250 micro, has a normal max output of 240W but can peek to 250W. Connected to 260W panels. System is under 10 months old so minimal loss from age. Hosed off the panels early in the morning so they were mostly clean. Ground mount so lots of free air space under panels help cool the temps.

Here is a 24 hour video showing the conditions (midnight to midnight): https://www.wunderground.com/webcams...e=20170917.mp4
Other than a few high clouds that drifted across the day was one of the nicest days we had in some time. Cooler and clear, haze from Railroad fire mostly gone, high temps of 82.2F. System is right at 3200' elevation so much better sky's than what you will find in the valley.

I pulled down the data from pvoutput (it is close enough to what Enphase is showing its good enough for this example), and sucked it into excel.

Yesterday the average output per micro over the entire day was 121.2W, so under 50% output
Between the times of 11pm and 2pm the average output was 202.2 W so around 80% output.
Max peak output was 217.2W at 1:10pm and it was near that value for about 30 minutes.

I can count on one hand the number of times I have hit 250W - all of them were after snow storm - all ground covered with snow - freezing temps and completely clear skies - and in all cases it lasted under 30 minutes. Sunglasses mandatory type of days. Other than that - never hit that 250W mark.

Thus expect that 400W panel on GREAT days to give you under 200W on average for the day. Maybe peak at 320W (above the IQ6+, so some clipping) but your daily output will be higher so it does not matter if you clip. As already stated in great detail by other posters - it all boils down to cost and your available roof space to decided what size panel to use.

This is 100% Enphase argument and makes since only if you are stuck using Enphase limited capabilities. If you have 400w pv modules, you will get more production with an inverter that will not clip and frankly it will cost you less money.

[max2k]

...
Between the times of 11pm and 2pm the average output was 202.2 W so around 80% output.
Max peak output was 217.2W at 1:10pm and it was near that value for about 30 minutes. ...

I see similar pattern here- on the best days 85% of STC on non clipping system with central inverter. This is purely due to panels heating up- right after the moment a cloud passes by and they cool down they jump to 96% STC and then decrease back to 85%. This is for 7,320W STC system which never goes above 6,250W on the normal day: PowerOutput.png

[J.P.M.]

Great discussion. Here is real example. Enphase M250 micro, has a normal max output of 240W but can peek to 250W. Connected to 260W panels. System is under 10 months old so minimal loss from age. Hosed off the panels early in the morning so they were mostly clean. Ground mount so lots of free air space under panels help cool the temps.

Here is a 24 hour video showing the conditions (midnight to midnight): https://www.wunderground.com/webcams...e=20170917.mp4
Other than a few high clouds that drifted across the day was one of the nicest days we had in some time. Cooler and clear, haze from Railroad fire mostly gone, high temps of 82.2F. System is right at 3200' elevation so much better sky's than what you will find in the valley.

I pulled down the data from pvoutput (it is close enough to what Enphase is showing its good enough for this example), and sucked it into excel.

Yesterday the average output per micro over the entire day was 121.2W, so under 50% output
Between the times of 11pm and 2pm the average output was 202.2 W so around 80% output.
Max peak output was 217.2W at 1:10pm and it was near that value for about 30 minutes.

I can count on one hand the number of times I have hit 250W - all of them were after snow storm - all ground covered with snow - freezing temps and completely clear skies - and in all cases it lasted under 30 minutes. Sunglasses mandatory type of days. Other than that - never hit that 250W mark.

Thus expect that 400W panel on GREAT days to give you under 200W on average for the day. Maybe peak at 320W (above the IQ6+, so some clipping) but your daily output will be higher so it does not matter if you clip. As already stated in great detail by other posters - it all boils down to cost and your available roof space to decided what size panel to use.

Good info. Thanx. Any data on P.O,A. irradiance ?

[tyab]

Yes max2 - this thread was from the OP trying to understand why one would pair a 400W panel with a IQ6+ micro (see post #1 and your reply and others on page 1 and his questions). And due to RF noise that interferes with his hobby he has ruled out DC optimizers. String inverters are an option but unfortunately its getting harder in CA to find installers that will iinstall new residential rooftop string inverters due to 690.12 - and he has questions about how much RF noise is generated by those string inverters. The OP's options are limited due to his unique RF requirements for a new system on his new (under construction?) home. I thought that a real life output example may help him understand why clipping is not that big of an issue even though many mathematical explanations has already been given.