Please help me decide on which way to go. (enPhase vs SolarEdge vs LG)

[russ]

Any time you see a report made and paid for by a company beware.

The daily data we have from the logs shows a much different story.

I have a 10KW Enphase I installed and a 10KW Solar Edge installed. Both have zero shading and perfect south views at the same angle.
In the winter the Solar Edge makes 6 to 8kw a day more due to the panel power being so much higher in the cold and the Enphase flat tops.

Watching all these systems on line shows the real picture.

Comparing two systems at even slightly different locations is meaningless. You are spreading the snake oil you are worried about - nothing more.

[sensij]

Rather than believing blanket statements about which will perform better various conditions, it can easily be modeled using PVWatts or SAM. The amount lost to clipping can be estimated and valued at the appropriate cost per kWh. The OP in this case is not in the mid-west, and data from those systems in that location seems odd to bring up.

[n9lri]

That test was rigged to generate a particular outcome, as many are. There should be very little difference in the output between SE and Enphase.

Any time you see a report made and paid for by a company beware.

The daily data we have from the logs shows a much different story.

I have a 10KW Enphase I installed and a 10KW Solar Edge installed. Both have zero shading and perfect south views at the same angle.
In the winter the Solar Edge makes 6 to 8kw a day more due to the panel power being so much higher in the cold and the Enphase flat tops.

In the heat of summer there near even. Down south it don't matter but up in the northern Midwest it matters ever winter day.

If a panel in the cold makes over 300 watts -- it's best to be able to use it.

Watching all these systems on line shows the real picture.

[sensij]

Sometime soon I'll share some comparison charts showing SE reported power vs revenue grade measured power, but I'm working on some other analysis at the moment first. Also, just to get in front of a possible objection... the CT's for my revenue grade meter are actually located inside the inverter, so termination or wire resistance shouldn't really be coming into play.

Actually, here is an example I can share, from 5/10 (a nearly clear day):

http://pvoutput.org/comparelive.jsp?sid=33758&sid1=35936&dt=20150510&gs=0&c=4&t=1

This is a gross consumption chart... the Tecolote Canyon system is using SE 5 min data (power, not energy), and the -Test system is using the EKM Meter (revenue grade) to report solar generation. Gross Consumption is calculated from generation and net consumption. Net consumption comes from an Eagle device, through Wattvision, and is the same for both systems (as you can see during the hours when the sun is down).

You can see early in the day, both systems show about the same gross consumption, and they maintain the same baseline established when the sun wasn't out. Later in the day, the SE generation has drifted slightly higher, pulling up the calculated gross consumption number to compensate, at a level that is higher than the baseline. The revenue grade system is continuing to maintain the same baseline, as might be expected.

The end result was that SE reported more energy than the revenue grade system.

EKM Meter: 20.85 kWh
SE sum of 5 or 15 min Power, multiplied by time: 21.00 kWh (0.7% above EKM)
SE Energy (from the SE portal): 21.63 kWh (3.0% above SE power, 3.7% above EKM)

I changed the parameters of the test after that day, so scrolling to days forward or backward may be comparing different things.

[sensij]

One of these days, one of us needs to shut off all of our breakers except the main and solar feed-in for a day. Then we can compare dashboard to poco meter...

It really isn't necessary. All you need is a predictable baseline load running during the day when no one is home, and you can see that the higher 3% energy numbers don't add up. I've added a revenue grade meter to my system to be sure, and am finding that even the sum of SolarEdge power numbers (3% less than the energy total), is also almost 1% over what the revenue grade meter shows. This is supported by the calculated baseline consumption numbers, when the math is done.

Sometime soon I'll share some comparison charts showing SE reported power vs revenue grade measured power, but I'm working on some other analysis at the moment first. Also, just to get in front of a possible objection... the CT's for my revenue grade meter are actually located inside the inverter, so termination or wire resistance shouldn't really be coming into play.

Really, SE is taking as much liberty with their numbers as they can, and while other inverters (Enphase, SMA) seem to do this a little bit, I think they tend to stick closer to the truth. It doesn't make SE a bad system... just, if accurate monitoring is your thing, SE gets an F for that.

[sensij]

http://enphase.com/global/files/PVEL...-SolarEdge.pdf


Maybe not....

That test was rigged to generate a particular outcome, as many are. There should be very little difference in the output between SE and Enphase.

[Alisobob]

This is a no brainer.

Solar Edge is by far the more efficient system.

http://enphase.com/global/files/PVEL_Study-on-EE-vs-SolarEdge.pdf

Maybe not....


solaredge.JPG

[sensij]

Yes, that's what I thought. But I can't explain the difference between the end-of-day production reported by the dashboard, and calculation by summing up all the 15-min wattages and divide by 4. The latter is always smaller by roughly 2-3%. I'm hoping you're right that the dashboard is the actual production, but the number seems a little too good comparing to Enphase (which on paper should have efficiency within 1% from SE).

If you look at the technical data API, you can see there are two data sources within the inverter. One gives apparent power that is exactly 3.0% greater than the other, at all times, in all irradiance environments. The higher one is used for the energy data, the lower one is used for the power data. The reason the relationship between power and energy drifts from the 3.0% in cloudy conditions is because the energy is averaged differently than the power, leading to some synchronization error that can stack up.

I've come up with no better explanation than SE is trying to make their system look better than the competition's by publishing high energy numbers, but still within the ±5% spec they need to meet to be compliant with CEC, among others. When I've pushed hard enough on the SE tech that I got passed to an engineer, they did not confirm or deny it, they just say that 5% is their spec and revenue grade monitoring is required if you want better.

[ericf1]

I'm hoping you're right that the dashboard is the actual production

One of these days, one of us needs to shut off all of our breakers except the main and solar feed-in for a day. Then we can compare dashboard to poco meter...

[thejq]

Really
45 dollars x 25 years 1200 bucks and where do you live where power is so cheap ?

Base here is .18

It's your money.

In San Diego (the land of expensive electricity), if you're on EV-TOU2 plan, the cost of clipping is currently @ $0.48/KWh. So it can get a little expensive depending where you're and the rate plan you're in.

[n9lri]

Seems like you are the snake oil man?

So you have never seen an inverter efficiency curve

here you go.

eff.jpg

[thejq]

FYI for you and I.

My Question:
I am curious, when I look at my portal, I see stuff such as "Yesterdays Energy: 44.34kWh" is this the DC produced energy from all my panels, or is this the DC->AC energy converted by the inverter? I suspect I am viewing the DC... Is there a way to see both metrics?

SE Response:
5/12/2015 2:59 PM | Tom Simpson
What you see on the dashboard is Export to AC what you see on the module level (layout) is DC/DC. The dashboard only shows actual production which is always in AC never DC. You can only export AC not DC. There are no changing the configuration without your own GUI. See sunpec errata if this is something you are interested in

Yes, that's what I thought. But I can't explain the difference between the end-of-day production reported by the dashboard, and calculation by summing up all the 15-min wattages and divide by 4. The latter is always smaller by roughly 2-3%. I'm hoping you're right that the dashboard is the actual production, but the number seems a little too good comparing to Enphase (which on paper should have efficiency within 1% from SE).

[n9lri]

Seems like you are the snake oil man?

Really
45 dollars x 25 years 1200 bucks and where do you live where power is so cheap ?

Base here is .18

It's your money.

[thejq]

To throw one more question into the mix. I originally had my system sized on last year's data of 12,484 and $2,746 based on Edison's 4 tier domestic rate I was on. However, since purchasing my Elec vehicle, I was able to switch to TOU-B and move most of my pool pump operation to night time. Also, my wife and I are both at work during the day so the TOU really works well for us. The end result is that this first month of TOU is shaping up to be about the same kWh as this period last year, but $100 less. If this extrapolate's out, I could be looking at around a $1k savings just by changing plans. With that in mind, should I stick with a 25 panel 7.6kW system? Or should I downsize a bit?

With TOU plans, your orientation also becomes important. I believe for SCE, SW or SWW facing work the best, because of on-peak time frame. If you can maximize generation and minimize usage during on-peak, there's no reason to size your system to cover > 90% of the usage. A much smaller system (eg. 70%-75%?) is probably all you need. But you have to go through the math yourself to be convinced. Another factor to consider is that all the CA utilities are considering moving towards having a minimum monthly charge, so it's better to pay something ($10-20) at the end of the month just so it's bigger than the minimum charge.

[AaronG]

I think it happens to everyone with SE. FYI, yesterday my SE shows 33.58KWh, but PVoutput registered 32.57KWh. The difference seems to be around 3% on a clear day and somewhat less on a cloudy day. A wild guess is maybe the inverter is calculating DC wattage, but reports the actual AC wattage in the API (which PVoutput reads), since SE claims 97% efficiency, hence 3% difference. Someone else also speculated that the 15-min update rate has something to do with it, but the symmetry of the bell curve should have taken care of the lagging effect. So for comparison with other brands, it's probably safer to use the PVoutputs data, or takes SE's data and divide 1.03.

FYI for you and I.

My Question:
I am curious, when I look at my portal, I see stuff such as "Yesterdays Energy: 44.34kWh" is this the DC produced energy from all my panels, or is this the DC->AC energy converted by the inverter? I suspect I am viewing the DC... Is there a way to see both metrics?

SE Response:
5/12/2015 2:59 PM | Tom Simpson
What you see on the dashboard is Export to AC what you see on the module level (layout) is DC/DC. The dashboard only shows actual production which is always in AC never DC. You can only export AC not DC. There are no changing the configuration without your own GUI. See sunpec errata if this is something you are interested in